WIP

tests/data/test_admin_models_v2.py

@@ -0,0 +1,524 @@
+"""
+Tests for Admin Models v2 - Batch Upload and Training Links.
+
+Tests for new SQLModel classes: BatchUpload, BatchUploadFile,
+TrainingDocumentLink, AnnotationHistory.
+"""
+
+import pytest
+from datetime import datetime
+from uuid import UUID, uuid4
+
+from src.data.admin_models import (
+    BatchUpload,
+    BatchUploadFile,
+    TrainingDocumentLink,
+    AnnotationHistory,
+    AdminDocument,
+    AdminAnnotation,
+    TrainingTask,
+    FIELD_CLASSES,
+    CSV_TO_CLASS_MAPPING,
+)
+
+
+class TestBatchUpload:
+    """Tests for BatchUpload model."""
+
+    def test_batch_upload_creation(self):
+        """Test basic batch upload creation."""
+        batch = BatchUpload(
+            admin_token="test-token",
+            filename="invoices.zip",
+            file_size=1024000,
+            upload_source="ui",
+        )
+
+        assert batch.batch_id is not None
+        assert isinstance(batch.batch_id, UUID)
+        assert batch.admin_token == "test-token"
+        assert batch.filename == "invoices.zip"
+        assert batch.file_size == 1024000
+        assert batch.upload_source == "ui"
+        assert batch.status == "processing"
+        assert batch.total_files == 0
+        assert batch.processed_files == 0
+        assert batch.successful_files == 0
+        assert batch.failed_files == 0
+        assert batch.error_message is None
+        assert batch.completed_at is None
+
+    def test_batch_upload_api_source(self):
+        """Test batch upload with API source."""
+        batch = BatchUpload(
+            admin_token="api-token",
+            filename="batch.zip",
+            file_size=2048000,
+            upload_source="api",
+        )
+
+        assert batch.upload_source == "api"
+
+    def test_batch_upload_with_progress(self):
+        """Test batch upload with progress tracking."""
+        batch = BatchUpload(
+            admin_token="test-token",
+            filename="large_batch.zip",
+            file_size=10240000,
+            total_files=100,
+            processed_files=50,
+            successful_files=48,
+            failed_files=2,
+            status="processing",
+        )
+
+        assert batch.total_files == 100
+        assert batch.processed_files == 50
+        assert batch.successful_files == 48
+        assert batch.failed_files == 2
+
+    def test_batch_upload_completed(self):
+        """Test completed batch upload."""
+        now = datetime.utcnow()
+        batch = BatchUpload(
+            admin_token="test-token",
+            filename="batch.zip",
+            file_size=1024000,
+            status="completed",
+            total_files=10,
+            processed_files=10,
+            successful_files=10,
+            failed_files=0,
+            completed_at=now,
+        )
+
+        assert batch.status == "completed"
+        assert batch.completed_at == now
+
+    def test_batch_upload_failed(self):
+        """Test failed batch upload."""
+        batch = BatchUpload(
+            admin_token="test-token",
+            filename="bad.zip",
+            file_size=1024,
+            status="failed",
+            error_message="Invalid ZIP file format",
+        )
+
+        assert batch.status == "failed"
+        assert batch.error_message == "Invalid ZIP file format"
+
+    def test_batch_upload_partial(self):
+        """Test partial batch upload with some failures."""
+        batch = BatchUpload(
+            admin_token="test-token",
+            filename="mixed.zip",
+            file_size=5120000,
+            status="partial",
+            total_files=20,
+            processed_files=20,
+            successful_files=15,
+            failed_files=5,
+        )
+
+        assert batch.status == "partial"
+        assert batch.failed_files == 5
+
+
+class TestBatchUploadFile:
+    """Tests for BatchUploadFile model."""
+
+    def test_batch_upload_file_creation(self):
+        """Test basic file record creation."""
+        batch_id = uuid4()
+        file_record = BatchUploadFile(
+            batch_id=batch_id,
+            filename="INV001.pdf",
+        )
+
+        assert file_record.file_id is not None
+        assert isinstance(file_record.file_id, UUID)
+        assert file_record.batch_id == batch_id
+        assert file_record.filename == "INV001.pdf"
+        assert file_record.status == "pending"
+        assert file_record.document_id is None
+        assert file_record.error_message is None
+        assert file_record.csv_row_data is None
+        assert file_record.processed_at is None
+
+    def test_batch_upload_file_with_document(self):
+        """Test file record linked to document."""
+        batch_id = uuid4()
+        document_id = uuid4()
+        file_record = BatchUploadFile(
+            batch_id=batch_id,
+            document_id=document_id,
+            filename="INV002.pdf",
+            status="completed",
+        )
+
+        assert file_record.document_id == document_id
+        assert file_record.status == "completed"
+
+    def test_batch_upload_file_with_csv_data(self):
+        """Test file record with CSV row data."""
+        batch_id = uuid4()
+        csv_data = {
+            "DocumentId": "INV003",
+            "InvoiceNumber": "F2024-003",
+            "Amount": "1500.00",
+            "OCR": "7350012345678",
+        }
+        file_record = BatchUploadFile(
+            batch_id=batch_id,
+            filename="INV003.pdf",
+            csv_row_data=csv_data,
+        )
+
+        assert file_record.csv_row_data == csv_data
+        assert file_record.csv_row_data["InvoiceNumber"] == "F2024-003"
+
+    def test_batch_upload_file_failed(self):
+        """Test failed file record."""
+        batch_id = uuid4()
+        file_record = BatchUploadFile(
+            batch_id=batch_id,
+            filename="corrupted.pdf",
+            status="failed",
+            error_message="Corrupted PDF file",
+        )
+
+        assert file_record.status == "failed"
+        assert file_record.error_message == "Corrupted PDF file"
+
+    def test_batch_upload_file_skipped(self):
+        """Test skipped file record."""
+        batch_id = uuid4()
+        file_record = BatchUploadFile(
+            batch_id=batch_id,
+            filename="not_a_pdf.txt",
+            status="skipped",
+            error_message="Not a PDF file",
+        )
+
+        assert file_record.status == "skipped"
+
+
+class TestTrainingDocumentLink:
+    """Tests for TrainingDocumentLink model."""
+
+    def test_training_document_link_creation(self):
+        """Test basic link creation."""
+        task_id = uuid4()
+        document_id = uuid4()
+        link = TrainingDocumentLink(
+            task_id=task_id,
+            document_id=document_id,
+        )
+
+        assert link.link_id is not None
+        assert isinstance(link.link_id, UUID)
+        assert link.task_id == task_id
+        assert link.document_id == document_id
+        assert link.annotation_snapshot is None
+
+    def test_training_document_link_with_snapshot(self):
+        """Test link with annotation snapshot."""
+        task_id = uuid4()
+        document_id = uuid4()
+        snapshot = {
+            "annotations": [
+                {
+                    "class_id": 0,
+                    "class_name": "invoice_number",
+                    "text_value": "F2024-001",
+                    "x_center": 0.5,
+                    "y_center": 0.3,
+                },
+                {
+                    "class_id": 6,
+                    "class_name": "amount",
+                    "text_value": "1500.00",
+                    "x_center": 0.7,
+                    "y_center": 0.6,
+                },
+            ],
+            "total_count": 2,
+            "snapshot_time": "2024-01-20T15:00:00",
+        }
+        link = TrainingDocumentLink(
+            task_id=task_id,
+            document_id=document_id,
+            annotation_snapshot=snapshot,
+        )
+
+        assert link.annotation_snapshot == snapshot
+        assert len(link.annotation_snapshot["annotations"]) == 2
+
+
+class TestAnnotationHistory:
+    """Tests for AnnotationHistory model."""
+
+    def test_annotation_history_created(self):
+        """Test history record for creation."""
+        annotation_id = uuid4()
+        new_value = {
+            "class_id": 0,
+            "class_name": "invoice_number",
+            "text_value": "F2024-001",
+            "bbox_x": 100,
+            "bbox_y": 200,
+            "bbox_width": 150,
+            "bbox_height": 30,
+            "source": "manual",
+        }
+        history = AnnotationHistory(
+            annotation_id=annotation_id,
+            action="created",
+            new_value=new_value,
+            changed_by="admin-token-123",
+        )
+
+        assert history.history_id is not None
+        assert history.annotation_id == annotation_id
+        assert history.action == "created"
+        assert history.previous_value is None
+        assert history.new_value == new_value
+        assert history.changed_by == "admin-token-123"
+
+    def test_annotation_history_updated(self):
+        """Test history record for update."""
+        annotation_id = uuid4()
+        previous_value = {
+            "text_value": "F2024-001",
+            "bbox_x": 100,
+        }
+        new_value = {
+            "text_value": "F2024-001-A",
+            "bbox_x": 110,
+        }
+        history = AnnotationHistory(
+            annotation_id=annotation_id,
+            action="updated",
+            previous_value=previous_value,
+            new_value=new_value,
+            changed_by="admin-token-123",
+            change_reason="Corrected OCR error",
+        )
+
+        assert history.action == "updated"
+        assert history.previous_value == previous_value
+        assert history.new_value == new_value
+        assert history.change_reason == "Corrected OCR error"
+
+    def test_annotation_history_override(self):
+        """Test history record for override."""
+        annotation_id = uuid4()
+        previous_value = {
+            "text_value": "F2024-001",
+            "source": "auto",
+            "confidence": 0.85,
+        }
+        new_value = {
+            "text_value": "F2024-001-CORRECTED",
+            "source": "manual",
+            "confidence": None,
+        }
+        history = AnnotationHistory(
+            annotation_id=annotation_id,
+            action="override",
+            previous_value=previous_value,
+            new_value=new_value,
+            changed_by="admin-token-123",
+            change_reason="Manual correction of auto-label",
+        )
+
+        assert history.action == "override"
+
+    def test_annotation_history_deleted(self):
+        """Test history record for deletion."""
+        annotation_id = uuid4()
+        previous_value = {
+            "class_id": 6,
+            "class_name": "amount",
+            "text_value": "1500.00",
+        }
+        history = AnnotationHistory(
+            annotation_id=annotation_id,
+            action="deleted",
+            previous_value=previous_value,
+            changed_by="admin-token-123",
+            change_reason="Incorrect annotation",
+        )
+
+        assert history.action == "deleted"
+        assert history.new_value is None
+
+
+class TestAdminDocumentExtensions:
+    """Tests for AdminDocument extensions."""
+
+    def test_document_with_upload_source(self):
+        """Test document with upload source field."""
+        doc = AdminDocument(
+            admin_token="test-token",
+            filename="invoice.pdf",
+            file_size=1024,
+            content_type="application/pdf",
+            file_path="/tmp/invoice.pdf",
+            upload_source="api",
+        )
+
+        assert doc.upload_source == "api"
+
+    def test_document_with_batch_id(self):
+        """Test document linked to batch upload."""
+        batch_id = uuid4()
+        doc = AdminDocument(
+            admin_token="test-token",
+            filename="invoice.pdf",
+            file_size=1024,
+            content_type="application/pdf",
+            file_path="/tmp/invoice.pdf",
+            batch_id=batch_id,
+        )
+
+        assert doc.batch_id == batch_id
+
+    def test_document_with_csv_field_values(self):
+        """Test document with CSV field values."""
+        csv_values = {
+            "InvoiceNumber": "F2024-001",
+            "Amount": "1500.00",
+            "OCR": "7350012345678",
+        }
+        doc = AdminDocument(
+            admin_token="test-token",
+            filename="invoice.pdf",
+            file_size=1024,
+            content_type="application/pdf",
+            file_path="/tmp/invoice.pdf",
+            csv_field_values=csv_values,
+        )
+
+        assert doc.csv_field_values == csv_values
+
+    def test_document_with_annotation_lock(self):
+        """Test document with annotation lock."""
+        lock_until = datetime.utcnow()
+        doc = AdminDocument(
+            admin_token="test-token",
+            filename="invoice.pdf",
+            file_size=1024,
+            content_type="application/pdf",
+            file_path="/tmp/invoice.pdf",
+            annotation_lock_until=lock_until,
+        )
+
+        assert doc.annotation_lock_until == lock_until
+
+
+class TestAdminAnnotationExtensions:
+    """Tests for AdminAnnotation extensions."""
+
+    def test_annotation_with_verification(self):
+        """Test annotation with verification fields."""
+        now = datetime.utcnow()
+        ann = AdminAnnotation(
+            document_id=uuid4(),
+            class_id=0,
+            class_name="invoice_number",
+            x_center=0.5,
+            y_center=0.3,
+            width=0.2,
+            height=0.05,
+            bbox_x=100,
+            bbox_y=200,
+            bbox_width=150,
+            bbox_height=30,
+            is_verified=True,
+            verified_at=now,
+            verified_by="admin-token-123",
+        )
+
+        assert ann.is_verified is True
+        assert ann.verified_at == now
+        assert ann.verified_by == "admin-token-123"
+
+    def test_annotation_with_override_info(self):
+        """Test annotation with override information."""
+        original_id = uuid4()
+        ann = AdminAnnotation(
+            document_id=uuid4(),
+            class_id=0,
+            class_name="invoice_number",
+            x_center=0.5,
+            y_center=0.3,
+            width=0.2,
+            height=0.05,
+            bbox_x=100,
+            bbox_y=200,
+            bbox_width=150,
+            bbox_height=30,
+            source="manual",
+            override_source="auto",
+            original_annotation_id=original_id,
+        )
+
+        assert ann.override_source == "auto"
+        assert ann.original_annotation_id == original_id
+
+
+class TestTrainingTaskExtensions:
+    """Tests for TrainingTask extensions."""
+
+    def test_training_task_with_document_count(self):
+        """Test training task with document count."""
+        task = TrainingTask(
+            admin_token="test-token",
+            name="Training Run 2024-01",
+            document_count=500,
+        )
+
+        assert task.document_count == 500
+
+    def test_training_task_with_metrics(self):
+        """Test training task with extracted metrics."""
+        task = TrainingTask(
+            admin_token="test-token",
+            name="Training Run 2024-01",
+            status="completed",
+            metrics_mAP=0.935,
+            metrics_precision=0.92,
+            metrics_recall=0.88,
+        )
+
+        assert task.metrics_mAP == 0.935
+        assert task.metrics_precision == 0.92
+        assert task.metrics_recall == 0.88
+
+
+class TestCSVToClassMapping:
+    """Tests for CSV column to class ID mapping."""
+
+    def test_csv_mapping_exists(self):
+        """Test that CSV mapping is defined."""
+        assert CSV_TO_CLASS_MAPPING is not None
+        assert len(CSV_TO_CLASS_MAPPING) > 0
+
+    def test_csv_mapping_values(self):
+        """Test specific CSV column mappings."""
+        assert CSV_TO_CLASS_MAPPING["InvoiceNumber"] == 0
+        assert CSV_TO_CLASS_MAPPING["InvoiceDate"] == 1
+        assert CSV_TO_CLASS_MAPPING["InvoiceDueDate"] == 2
+        assert CSV_TO_CLASS_MAPPING["OCR"] == 3
+        assert CSV_TO_CLASS_MAPPING["Bankgiro"] == 4
+        assert CSV_TO_CLASS_MAPPING["Plusgiro"] == 5
+        assert CSV_TO_CLASS_MAPPING["Amount"] == 6
+        assert CSV_TO_CLASS_MAPPING["supplier_organisation_number"] == 7
+        assert CSV_TO_CLASS_MAPPING["customer_number"] == 9
+
+    def test_csv_mapping_matches_field_classes(self):
+        """Test that CSV mapping is consistent with FIELD_CLASSES."""
+        for csv_name, class_id in CSV_TO_CLASS_MAPPING.items():
+            assert class_id in FIELD_CLASSES

tests/test_config.py

@@ -18,7 +18,7 @@ class TestDatabaseConfig:
     def test_config_loads_from_env(self):
         """Test that config loads successfully from .env file."""
         # Import config (should load .env automatically)
-        import config
+        from src import config
 
         # Verify database config is loaded
         assert config.DATABASE is not None
@@ -30,7 +30,7 @@ class TestDatabaseConfig:
 
     def test_database_password_loaded(self):
         """Test that database password is loaded from environment."""
-        import config
+        from src import config
 
         # Password should be loaded from .env
         assert config.DATABASE['password'] is not None
@@ -38,7 +38,7 @@ class TestDatabaseConfig:
 
     def test_database_connection_string(self):
         """Test database connection string generation."""
-        import config
+        from src import config
 
         conn_str = config.get_db_connection_string()
 
@@ -71,7 +71,7 @@ class TestPathsConfig:
 
     def test_paths_config_exists(self):
         """Test that PATHS configuration exists."""
-        import config
+        from src import config
 
         assert config.PATHS is not None
         assert 'csv_dir' in config.PATHS
@@ -85,7 +85,7 @@ class TestAutolabelConfig:
 
     def test_autolabel_config_exists(self):
         """Test that AUTOLABEL configuration exists."""
-        import config
+        from src import config
 
         assert config.AUTOLABEL is not None
         assert 'workers' in config.AUTOLABEL
@@ -95,7 +95,7 @@ class TestAutolabelConfig:
 
     def test_autolabel_ratios_sum_to_one(self):
         """Test that train/val/test ratios sum to 1.0."""
-        import config
+        from src import config
 
         total = (
             config.AUTOLABEL['train_ratio'] +

tests/web/__init__.py

@@ -0,0 +1 @@
+"""Tests for web API components."""

tests/web/conftest.py

@@ -0,0 +1,132 @@
+"""
+Test fixtures for web API tests.
+"""
+
+import tempfile
+from datetime import datetime, timedelta
+from pathlib import Path
+from unittest.mock import MagicMock, patch
+from uuid import UUID
+
+import pytest
+
+from src.data.async_request_db import ApiKeyConfig, AsyncRequestDB
+from src.data.models import AsyncRequest
+from src.web.workers.async_queue import AsyncTask, AsyncTaskQueue
+from src.web.services.async_processing import AsyncProcessingService
+from src.web.config import AsyncConfig, StorageConfig
+from src.web.core.rate_limiter import RateLimiter
+
+
+@pytest.fixture
+def mock_db():
+    """Create a mock AsyncRequestDB."""
+    db = MagicMock(spec=AsyncRequestDB)
+
+    # Default return values
+    db.is_valid_api_key.return_value = True
+    db.get_api_key_config.return_value = ApiKeyConfig(
+        api_key="test-api-key",
+        name="Test Key",
+        is_active=True,
+        requests_per_minute=10,
+        max_concurrent_jobs=3,
+        max_file_size_mb=50,
+    )
+    db.count_active_jobs.return_value = 0
+    db.get_queue_position.return_value = 1
+
+    return db
+
+
+@pytest.fixture
+def rate_limiter(mock_db):
+    """Create a RateLimiter with mock database."""
+    return RateLimiter(mock_db)
+
+
+@pytest.fixture
+def task_queue():
+    """Create an AsyncTaskQueue."""
+    return AsyncTaskQueue(max_size=10, worker_count=1)
+
+
+@pytest.fixture
+def async_config():
+    """Create an AsyncConfig for testing."""
+    with tempfile.TemporaryDirectory() as tmpdir:
+        yield AsyncConfig(
+            queue_max_size=10,
+            worker_count=1,
+            task_timeout_seconds=30,
+            result_retention_days=7,
+            temp_upload_dir=Path(tmpdir) / "async",
+            max_file_size_mb=10,
+        )
+
+
+@pytest.fixture
+def storage_config():
+    """Create a StorageConfig for testing."""
+    with tempfile.TemporaryDirectory() as tmpdir:
+        yield StorageConfig(
+            upload_dir=Path(tmpdir) / "uploads",
+            result_dir=Path(tmpdir) / "results",
+            max_file_size_mb=50,
+        )
+
+
+@pytest.fixture
+def mock_inference_service():
+    """Create a mock InferenceService."""
+    service = MagicMock()
+    service.is_initialized = True
+    service.gpu_available = False
+
+    # Mock process_pdf to return a successful result
+    mock_result = MagicMock()
+    mock_result.document_id = "test-doc"
+    mock_result.success = True
+    mock_result.document_type = "invoice"
+    mock_result.fields = {"InvoiceNumber": "12345", "Amount": "1000.00"}
+    mock_result.confidence = {"InvoiceNumber": 0.95, "Amount": 0.92}
+    mock_result.detections = []
+    mock_result.errors = []
+    mock_result.visualization_path = None
+
+    service.process_pdf.return_value = mock_result
+    service.process_image.return_value = mock_result
+
+    return service
+
+
+# Valid UUID for testing
+TEST_REQUEST_UUID = "550e8400-e29b-41d4-a716-446655440000"
+
+
+@pytest.fixture
+def sample_async_request():
+    """Create a sample AsyncRequest."""
+    return AsyncRequest(
+        request_id=UUID(TEST_REQUEST_UUID),
+        api_key="test-api-key",
+        status="pending",
+        filename="test.pdf",
+        file_size=1024,
+        content_type="application/pdf",
+        expires_at=datetime.utcnow() + timedelta(days=7),
+    )
+
+
+@pytest.fixture
+def sample_task():
+    """Create a sample AsyncTask."""
+    with tempfile.NamedTemporaryFile(suffix=".pdf", delete=False) as f:
+        f.write(b"fake pdf content")
+        return AsyncTask(
+            request_id=TEST_REQUEST_UUID,
+            api_key="test-api-key",
+            file_path=Path(f.name),
+            filename="test.pdf",
+            created_at=datetime.utcnow(),
+        )

tests/web/test_admin_annotations.py

@@ -0,0 +1,197 @@
+"""
+Tests for Admin Annotation Routes.
+"""
+
+import pytest
+from datetime import datetime
+from unittest.mock import MagicMock, patch
+from uuid import UUID
+
+from fastapi import HTTPException
+
+from src.data.admin_models import AdminAnnotation, AdminDocument, FIELD_CLASSES
+from src.web.api.v1.admin.annotations import _validate_uuid, create_annotation_router
+from src.web.schemas.admin import (
+    AnnotationCreate,
+    AnnotationUpdate,
+    AutoLabelRequest,
+    BoundingBox,
+)
+
+
+# Test UUIDs
+TEST_DOC_UUID = "550e8400-e29b-41d4-a716-446655440000"
+TEST_ANN_UUID = "660e8400-e29b-41d4-a716-446655440001"
+TEST_TOKEN = "test-admin-token-12345"
+
+
+class TestAnnotationRouterCreation:
+    """Tests for annotation router creation."""
+
+    def test_creates_router_with_endpoints(self):
+        """Test router is created with expected endpoints."""
+        router = create_annotation_router()
+
+        # Get route paths (includes prefix)
+        paths = [route.path for route in router.routes]
+
+        # Paths include the /admin/documents prefix
+        assert any("{document_id}/annotations" in p for p in paths)
+        assert any("{annotation_id}" in p for p in paths)
+        assert any("auto-label" in p for p in paths)
+        assert any("images" in p for p in paths)
+
+
+class TestAnnotationCreateSchema:
+    """Tests for AnnotationCreate schema."""
+
+    def test_valid_annotation(self):
+        """Test valid annotation creation."""
+        ann = AnnotationCreate(
+            page_number=1,
+            class_id=0,
+            bbox=BoundingBox(x=100, y=100, width=200, height=50),
+            text_value="12345",
+        )
+
+        assert ann.page_number == 1
+        assert ann.class_id == 0
+        assert ann.bbox.x == 100
+        assert ann.text_value == "12345"
+
+    def test_class_id_range(self):
+        """Test class_id must be 0-9."""
+        # Valid class IDs
+        for class_id in range(10):
+            ann = AnnotationCreate(
+                page_number=1,
+                class_id=class_id,
+                bbox=BoundingBox(x=0, y=0, width=100, height=50),
+            )
+            assert ann.class_id == class_id
+
+    def test_bbox_validation(self):
+        """Test bounding box validation."""
+        bbox = BoundingBox(x=0, y=0, width=100, height=50)
+        assert bbox.width >= 1
+        assert bbox.height >= 1
+
+
+class TestAnnotationUpdateSchema:
+    """Tests for AnnotationUpdate schema."""
+
+    def test_partial_update(self):
+        """Test partial update with only some fields."""
+        update = AnnotationUpdate(
+            text_value="new value",
+        )
+
+        assert update.text_value == "new value"
+        assert update.class_id is None
+        assert update.bbox is None
+
+    def test_bbox_update(self):
+        """Test bounding box update."""
+        update = AnnotationUpdate(
+            bbox=BoundingBox(x=50, y=50, width=150, height=75),
+        )
+
+        assert update.bbox.x == 50
+        assert update.bbox.width == 150
+
+
+class TestAutoLabelRequestSchema:
+    """Tests for AutoLabelRequest schema."""
+
+    def test_valid_request(self):
+        """Test valid auto-label request."""
+        request = AutoLabelRequest(
+            field_values={
+                "InvoiceNumber": "12345",
+                "Amount": "1000.00",
+            },
+            replace_existing=True,
+        )
+
+        assert len(request.field_values) == 2
+        assert request.field_values["InvoiceNumber"] == "12345"
+        assert request.replace_existing is True
+
+    def test_requires_field_values(self):
+        """Test that field_values is required."""
+        with pytest.raises(Exception):
+            AutoLabelRequest(replace_existing=True)
+
+
+class TestFieldClasses:
+    """Tests for field class mapping."""
+
+    def test_all_classes_defined(self):
+        """Test all 10 field classes are defined."""
+        assert len(FIELD_CLASSES) == 10
+
+    def test_class_ids_sequential(self):
+        """Test class IDs are 0-9."""
+        assert set(FIELD_CLASSES.keys()) == set(range(10))
+
+    def test_known_field_names(self):
+        """Test known field names are present."""
+        names = list(FIELD_CLASSES.values())
+
+        assert "invoice_number" in names
+        assert "invoice_date" in names
+        assert "amount" in names
+        assert "bankgiro" in names
+        assert "ocr_number" in names
+
+
+class TestAnnotationModel:
+    """Tests for AdminAnnotation model."""
+
+    def test_annotation_creation(self):
+        """Test annotation model creation."""
+        ann = AdminAnnotation(
+            document_id=UUID(TEST_DOC_UUID),
+            page_number=1,
+            class_id=0,
+            class_name="invoice_number",
+            x_center=0.5,
+            y_center=0.5,
+            width=0.2,
+            height=0.05,
+            bbox_x=100,
+            bbox_y=100,
+            bbox_width=200,
+            bbox_height=50,
+            text_value="12345",
+            confidence=0.95,
+            source="manual",
+        )
+
+        assert str(ann.document_id) == TEST_DOC_UUID
+        assert ann.class_id == 0
+        assert ann.x_center == 0.5
+        assert ann.source == "manual"
+
+    def test_normalized_coordinates(self):
+        """Test normalized coordinates are 0-1 range."""
+        # Valid normalized coords
+        ann = AdminAnnotation(
+            document_id=UUID(TEST_DOC_UUID),
+            page_number=1,
+            class_id=0,
+            class_name="test",
+            x_center=0.5,
+            y_center=0.5,
+            width=0.2,
+            height=0.05,
+            bbox_x=0,
+            bbox_y=0,
+            bbox_width=100,
+            bbox_height=50,
+        )
+
+        assert 0 <= ann.x_center <= 1
+        assert 0 <= ann.y_center <= 1
+        assert 0 <= ann.width <= 1
+        assert 0 <= ann.height <= 1

tests/web/test_admin_auth.py

@@ -0,0 +1,162 @@
+"""
+Tests for Admin Authentication.
+"""
+
+import pytest
+from datetime import datetime, timedelta
+from unittest.mock import MagicMock, patch
+
+from fastapi import HTTPException
+
+from src.data.admin_db import AdminDB
+from src.data.admin_models import AdminToken
+from src.web.core.auth import (
+    get_admin_db,
+    reset_admin_db,
+    validate_admin_token,
+)
+
+
+@pytest.fixture
+def mock_admin_db():
+    """Create a mock AdminDB."""
+    db = MagicMock(spec=AdminDB)
+    db.is_valid_admin_token.return_value = True
+    return db
+
+
+@pytest.fixture(autouse=True)
+def reset_db():
+    """Reset admin DB after each test."""
+    yield
+    reset_admin_db()
+
+
+class TestValidateAdminToken:
+    """Tests for validate_admin_token dependency."""
+
+    def test_missing_token_raises_401(self, mock_admin_db):
+        """Test that missing token raises 401."""
+        import asyncio
+
+        with pytest.raises(HTTPException) as exc_info:
+            asyncio.get_event_loop().run_until_complete(
+                validate_admin_token(None, mock_admin_db)
+            )
+
+        assert exc_info.value.status_code == 401
+        assert "Admin token required" in exc_info.value.detail
+
+    def test_invalid_token_raises_401(self, mock_admin_db):
+        """Test that invalid token raises 401."""
+        import asyncio
+
+        mock_admin_db.is_valid_admin_token.return_value = False
+
+        with pytest.raises(HTTPException) as exc_info:
+            asyncio.get_event_loop().run_until_complete(
+                validate_admin_token("invalid-token", mock_admin_db)
+            )
+
+        assert exc_info.value.status_code == 401
+        assert "Invalid or expired" in exc_info.value.detail
+
+    def test_valid_token_returns_token(self, mock_admin_db):
+        """Test that valid token is returned."""
+        import asyncio
+
+        token = "valid-test-token"
+        mock_admin_db.is_valid_admin_token.return_value = True
+
+        result = asyncio.get_event_loop().run_until_complete(
+            validate_admin_token(token, mock_admin_db)
+        )
+
+        assert result == token
+        mock_admin_db.update_admin_token_usage.assert_called_once_with(token)
+
+
+class TestAdminDB:
+    """Tests for AdminDB operations."""
+
+    def test_is_valid_admin_token_active(self):
+        """Test valid active token."""
+        with patch("src.data.admin_db.get_session_context") as mock_ctx:
+            mock_session = MagicMock()
+            mock_ctx.return_value.__enter__.return_value = mock_session
+
+            mock_token = AdminToken(
+                token="test-token",
+                name="Test",
+                is_active=True,
+                expires_at=None,
+            )
+            mock_session.get.return_value = mock_token
+
+            db = AdminDB()
+            assert db.is_valid_admin_token("test-token") is True
+
+    def test_is_valid_admin_token_inactive(self):
+        """Test inactive token."""
+        with patch("src.data.admin_db.get_session_context") as mock_ctx:
+            mock_session = MagicMock()
+            mock_ctx.return_value.__enter__.return_value = mock_session
+
+            mock_token = AdminToken(
+                token="test-token",
+                name="Test",
+                is_active=False,
+                expires_at=None,
+            )
+            mock_session.get.return_value = mock_token
+
+            db = AdminDB()
+            assert db.is_valid_admin_token("test-token") is False
+
+    def test_is_valid_admin_token_expired(self):
+        """Test expired token."""
+        with patch("src.data.admin_db.get_session_context") as mock_ctx:
+            mock_session = MagicMock()
+            mock_ctx.return_value.__enter__.return_value = mock_session
+
+            mock_token = AdminToken(
+                token="test-token",
+                name="Test",
+                is_active=True,
+                expires_at=datetime.utcnow() - timedelta(days=1),
+            )
+            mock_session.get.return_value = mock_token
+
+            db = AdminDB()
+            assert db.is_valid_admin_token("test-token") is False
+
+    def test_is_valid_admin_token_not_found(self):
+        """Test token not found."""
+        with patch("src.data.admin_db.get_session_context") as mock_ctx:
+            mock_session = MagicMock()
+            mock_ctx.return_value.__enter__.return_value = mock_session
+            mock_session.get.return_value = None
+
+            db = AdminDB()
+            assert db.is_valid_admin_token("nonexistent") is False
+
+
+class TestGetAdminDb:
+    """Tests for get_admin_db function."""
+
+    def test_returns_singleton(self):
+        """Test that get_admin_db returns singleton."""
+        reset_admin_db()
+
+        db1 = get_admin_db()
+        db2 = get_admin_db()
+
+        assert db1 is db2
+
+    def test_reset_clears_singleton(self):
+        """Test that reset clears singleton."""
+        db1 = get_admin_db()
+        reset_admin_db()
+        db2 = get_admin_db()
+
+        assert db1 is not db2

tests/web/test_admin_routes.py

@@ -0,0 +1,164 @@
+"""
+Tests for Admin Document Routes.
+"""
+
+import pytest
+from datetime import datetime
+from io import BytesIO
+from pathlib import Path
+from unittest.mock import MagicMock, patch
+from uuid import UUID
+
+from fastapi import HTTPException
+from fastapi.testclient import TestClient
+
+from src.data.admin_models import AdminDocument, AdminToken
+from src.web.api.v1.admin.documents import _validate_uuid, create_admin_router
+
+
+# Test UUID
+TEST_DOC_UUID = "550e8400-e29b-41d4-a716-446655440000"
+TEST_TOKEN = "test-admin-token-12345"
+
+
+class TestValidateUUID:
+    """Tests for UUID validation."""
+
+    def test_valid_uuid(self):
+        """Test valid UUID passes validation."""
+        _validate_uuid(TEST_DOC_UUID, "test")  # Should not raise
+
+    def test_invalid_uuid_raises_400(self):
+        """Test invalid UUID raises 400."""
+        with pytest.raises(HTTPException) as exc_info:
+            _validate_uuid("not-a-uuid", "document_id")
+
+        assert exc_info.value.status_code == 400
+        assert "Invalid document_id format" in exc_info.value.detail
+
+
+class TestAdminRouter:
+    """Tests for admin router creation."""
+
+    def test_creates_router_with_endpoints(self):
+        """Test router is created with expected endpoints."""
+        router = create_admin_router((".pdf", ".png", ".jpg"))
+
+        # Get route paths (include prefix from router)
+        paths = [route.path for route in router.routes]
+
+        # Paths include the /admin prefix
+        assert any("/auth/token" in p for p in paths)
+        assert any("/documents" in p for p in paths)
+        assert any("/documents/stats" in p for p in paths)
+        assert any("{document_id}" in p for p in paths)
+
+
+class TestCreateTokenEndpoint:
+    """Tests for POST /admin/auth/token endpoint."""
+
+    @pytest.fixture
+    def mock_db(self):
+        """Create mock AdminDB."""
+        db = MagicMock()
+        db.is_valid_admin_token.return_value = True
+        return db
+
+    def test_create_token_success(self, mock_db):
+        """Test successful token creation."""
+        from src.web.schemas.admin import AdminTokenCreate
+
+        request = AdminTokenCreate(name="Test Token", expires_in_days=30)
+
+        # The actual endpoint would generate a token
+        # This tests the schema validation
+        assert request.name == "Test Token"
+        assert request.expires_in_days == 30
+
+
+class TestDocumentUploadEndpoint:
+    """Tests for POST /admin/documents endpoint."""
+
+    @pytest.fixture
+    def sample_pdf_bytes(self):
+        """Create sample PDF-like bytes."""
+        # Minimal PDF header
+        return b"%PDF-1.4\n%\xe2\xe3\xcf\xd3\n"
+
+    @pytest.fixture
+    def mock_admin_db(self):
+        """Create mock AdminDB."""
+        db = MagicMock()
+        db.is_valid_admin_token.return_value = True
+        db.create_document.return_value = TEST_DOC_UUID
+        return db
+
+    def test_rejects_invalid_extension(self):
+        """Test that invalid file extensions are rejected."""
+        # Schema validation would happen at the route level
+        allowed = (".pdf", ".png", ".jpg")
+        file_ext = ".exe"
+
+        assert file_ext not in allowed
+
+
+class TestDocumentListEndpoint:
+    """Tests for GET /admin/documents endpoint."""
+
+    @pytest.fixture
+    def sample_documents(self):
+        """Create sample documents."""
+        return [
+            AdminDocument(
+                document_id=UUID(TEST_DOC_UUID),
+                admin_token=TEST_TOKEN,
+                filename="test.pdf",
+                file_size=1024,
+                content_type="application/pdf",
+                file_path="/tmp/test.pdf",
+                page_count=1,
+                status="pending",
+            ),
+        ]
+
+    def test_validates_status_filter(self):
+        """Test that invalid status filter is rejected."""
+        valid_statuses = ("pending", "auto_labeling", "labeled", "exported")
+
+        assert "invalid_status" not in valid_statuses
+        assert "pending" in valid_statuses
+
+
+class TestDocumentDetailEndpoint:
+    """Tests for GET /admin/documents/{document_id} endpoint."""
+
+    def test_requires_valid_uuid(self):
+        """Test that invalid UUID is rejected."""
+        with pytest.raises(HTTPException) as exc_info:
+            _validate_uuid("invalid", "document_id")
+
+        assert exc_info.value.status_code == 400
+
+
+class TestDocumentDeleteEndpoint:
+    """Tests for DELETE /admin/documents/{document_id} endpoint."""
+
+    def test_validates_document_id(self):
+        """Test that document_id is validated."""
+        # Valid UUID should not raise
+        _validate_uuid(TEST_DOC_UUID, "document_id")
+
+        # Invalid should raise
+        with pytest.raises(HTTPException):
+            _validate_uuid("bad-id", "document_id")
+
+
+class TestDocumentStatusUpdateEndpoint:
+    """Tests for PATCH /admin/documents/{document_id}/status endpoint."""
+
+    def test_validates_status_values(self):
+        """Test that only valid statuses are accepted."""
+        valid_statuses = ("pending", "labeled", "exported")
+
+        assert "pending" in valid_statuses
+        assert "invalid" not in valid_statuses

tests/web/test_admin_routes_enhanced.py

@@ -0,0 +1,351 @@
+"""
+Tests for Enhanced Admin Document Routes (Phase 3).
+"""
+
+import pytest
+from datetime import datetime
+from uuid import uuid4
+
+from fastapi import FastAPI
+from fastapi.testclient import TestClient
+
+from src.web.api.v1.admin.documents import create_admin_router
+from src.web.core.auth import validate_admin_token, get_admin_db
+
+
+class MockAdminDocument:
+    """Mock AdminDocument for testing."""
+
+    def __init__(self, **kwargs):
+        self.document_id = kwargs.get('document_id', uuid4())
+        self.admin_token = kwargs.get('admin_token', 'test-token')
+        self.filename = kwargs.get('filename', 'test.pdf')
+        self.file_size = kwargs.get('file_size', 100000)
+        self.content_type = kwargs.get('content_type', 'application/pdf')
+        self.page_count = kwargs.get('page_count', 1)
+        self.status = kwargs.get('status', 'pending')
+        self.auto_label_status = kwargs.get('auto_label_status', None)
+        self.auto_label_error = kwargs.get('auto_label_error', None)
+        self.upload_source = kwargs.get('upload_source', 'ui')
+        self.batch_id = kwargs.get('batch_id', None)
+        self.csv_field_values = kwargs.get('csv_field_values', None)
+        self.annotation_lock_until = kwargs.get('annotation_lock_until', None)
+        self.created_at = kwargs.get('created_at', datetime.utcnow())
+        self.updated_at = kwargs.get('updated_at', datetime.utcnow())
+
+
+class MockAnnotation:
+    """Mock AdminAnnotation for testing."""
+
+    def __init__(self, **kwargs):
+        self.annotation_id = kwargs.get('annotation_id', uuid4())
+        self.document_id = kwargs.get('document_id')
+        self.page_number = kwargs.get('page_number', 1)
+        self.class_id = kwargs.get('class_id', 0)
+        self.class_name = kwargs.get('class_name', 'invoice_number')
+        self.bbox_x = kwargs.get('bbox_x', 100.0)
+        self.bbox_y = kwargs.get('bbox_y', 100.0)
+        self.bbox_width = kwargs.get('bbox_width', 200.0)
+        self.bbox_height = kwargs.get('bbox_height', 50.0)
+        self.x_center = kwargs.get('x_center', 0.5)
+        self.y_center = kwargs.get('y_center', 0.5)
+        self.width = kwargs.get('width', 0.3)
+        self.height = kwargs.get('height', 0.1)
+        self.text_value = kwargs.get('text_value', 'INV-001')
+        self.confidence = kwargs.get('confidence', 0.95)
+        self.source = kwargs.get('source', 'manual')
+        self.created_at = kwargs.get('created_at', datetime.utcnow())
+
+
+class MockAdminDB:
+    """Mock AdminDB for testing enhanced features."""
+
+    def __init__(self):
+        self.documents = {}
+        self.annotations = {}
+
+    def get_documents_by_token(
+        self,
+        admin_token,
+        status=None,
+        upload_source=None,
+        has_annotations=None,
+        auto_label_status=None,
+        batch_id=None,
+        limit=20,
+        offset=0
+    ):
+        """Get filtered documents."""
+        docs = list(self.documents.values())
+
+        # Apply filters
+        if status:
+            docs = [d for d in docs if d.status == status]
+        if upload_source:
+            docs = [d for d in docs if d.upload_source == upload_source]
+        if has_annotations is not None:
+            for d in docs[:]:
+                ann_count = len(self.annotations.get(str(d.document_id), []))
+                if has_annotations and ann_count == 0:
+                    docs.remove(d)
+                elif not has_annotations and ann_count > 0:
+                    docs.remove(d)
+        if auto_label_status:
+            docs = [d for d in docs if d.auto_label_status == auto_label_status]
+        if batch_id:
+            docs = [d for d in docs if str(d.batch_id) == str(batch_id)]
+
+        total = len(docs)
+        return docs[offset:offset+limit], total
+
+    def get_annotations_for_document(self, document_id):
+        """Get annotations for document."""
+        return self.annotations.get(str(document_id), [])
+
+    def count_documents_by_status(self, admin_token):
+        """Count documents by status."""
+        counts = {}
+        for doc in self.documents.values():
+            if doc.admin_token == admin_token:
+                counts[doc.status] = counts.get(doc.status, 0) + 1
+        return counts
+
+    def get_document_by_token(self, document_id, admin_token):
+        """Get single document by ID and token."""
+        doc = self.documents.get(document_id)
+        if doc and doc.admin_token == admin_token:
+            return doc
+        return None
+
+    def get_document_training_tasks(self, document_id):
+        """Get training tasks that used this document."""
+        return []  # No training history in this test
+
+    def get_training_task(self, task_id):
+        """Get training task by ID."""
+        return None  # No training tasks in this test
+
+
+@pytest.fixture
+def app():
+    """Create test FastAPI app."""
+    app = FastAPI()
+
+    # Create mock DB
+    mock_db = MockAdminDB()
+
+    # Add test documents
+    doc1 = MockAdminDocument(
+        filename="INV001.pdf",
+        status="labeled",
+        upload_source="ui",
+        auto_label_status=None,
+        batch_id=None
+    )
+    doc2 = MockAdminDocument(
+        filename="INV002.pdf",
+        status="labeled",
+        upload_source="api",
+        auto_label_status="completed",
+        batch_id=uuid4()
+    )
+    doc3 = MockAdminDocument(
+        filename="INV003.pdf",
+        status="pending",
+        upload_source="ui",
+        auto_label_status=None,  # Not auto-labeled yet
+        batch_id=None
+    )
+
+    mock_db.documents[str(doc1.document_id)] = doc1
+    mock_db.documents[str(doc2.document_id)] = doc2
+    mock_db.documents[str(doc3.document_id)] = doc3
+
+    # Add annotations to doc1 and doc2
+    mock_db.annotations[str(doc1.document_id)] = [
+        MockAnnotation(
+            document_id=doc1.document_id,
+            class_name="invoice_number",
+            text_value="INV-001"
+        )
+    ]
+    mock_db.annotations[str(doc2.document_id)] = [
+        MockAnnotation(
+            document_id=doc2.document_id,
+            class_id=6,
+            class_name="amount",
+            text_value="1500.00"
+        ),
+        MockAnnotation(
+            document_id=doc2.document_id,
+            class_id=1,
+            class_name="invoice_date",
+            text_value="2024-01-15"
+        )
+    ]
+
+    # Override dependencies
+    app.dependency_overrides[validate_admin_token] = lambda: "test-token"
+    app.dependency_overrides[get_admin_db] = lambda: mock_db
+
+    # Include router
+    router = create_admin_router((".pdf", ".png", ".jpg"))
+    app.include_router(router)
+
+    return app
+
+
+@pytest.fixture
+def client(app):
+    """Create test client."""
+    return TestClient(app)
+
+
+class TestEnhancedDocumentList:
+    """Tests for enhanced document list endpoint."""
+
+    def test_list_documents_filter_by_upload_source_ui(self, client):
+        """Test filtering documents by upload_source=ui."""
+        response = client.get("/admin/documents?upload_source=ui")
+
+        assert response.status_code == 200
+        data = response.json()
+        assert data["total"] == 2
+        assert all(doc["filename"].startswith("INV") for doc in data["documents"])
+
+    def test_list_documents_filter_by_upload_source_api(self, client):
+        """Test filtering documents by upload_source=api."""
+        response = client.get("/admin/documents?upload_source=api")
+
+        assert response.status_code == 200
+        data = response.json()
+        assert data["total"] == 1
+        assert data["documents"][0]["filename"] == "INV002.pdf"
+
+    def test_list_documents_filter_by_has_annotations_true(self, client):
+        """Test filtering documents with annotations."""
+        response = client.get("/admin/documents?has_annotations=true")
+
+        assert response.status_code == 200
+        data = response.json()
+        assert data["total"] == 2
+
+    def test_list_documents_filter_by_has_annotations_false(self, client):
+        """Test filtering documents without annotations."""
+        response = client.get("/admin/documents?has_annotations=false")
+
+        assert response.status_code == 200
+        data = response.json()
+        assert data["total"] == 1
+
+    def test_list_documents_filter_by_auto_label_status(self, client):
+        """Test filtering by auto_label_status."""
+        response = client.get("/admin/documents?auto_label_status=completed")
+
+        assert response.status_code == 200
+        data = response.json()
+        assert data["total"] == 1
+        assert data["documents"][0]["filename"] == "INV002.pdf"
+
+    def test_list_documents_filter_by_batch_id(self, client):
+        """Test filtering by batch_id."""
+        # Get a batch_id from the test data
+        response_all = client.get("/admin/documents?upload_source=api")
+        batch_id = response_all.json()["documents"][0]["batch_id"]
+
+        response = client.get(f"/admin/documents?batch_id={batch_id}")
+
+        assert response.status_code == 200
+        data = response.json()
+        assert data["total"] == 1
+
+    def test_list_documents_combined_filters(self, client):
+        """Test combining multiple filters."""
+        response = client.get(
+            "/admin/documents?status=labeled&upload_source=api"
+        )
+
+        assert response.status_code == 200
+        data = response.json()
+        assert data["total"] == 1
+        assert data["documents"][0]["filename"] == "INV002.pdf"
+
+    def test_document_item_includes_new_fields(self, client):
+        """Test DocumentItem includes new Phase 2/3 fields."""
+        response = client.get("/admin/documents?upload_source=api")
+
+        assert response.status_code == 200
+        data = response.json()
+        doc = data["documents"][0]
+
+        # Check new fields exist
+        assert "upload_source" in doc
+        assert doc["upload_source"] == "api"
+        assert "batch_id" in doc
+        assert doc["batch_id"] is not None
+        assert "can_annotate" in doc
+        assert isinstance(doc["can_annotate"], bool)
+
+
+class TestEnhancedDocumentDetail:
+    """Tests for enhanced document detail endpoint."""
+
+    def test_document_detail_includes_new_fields(self, client, app):
+        """Test DocumentDetailResponse includes new Phase 2/3 fields."""
+        # Get a document ID from list
+        response = client.get("/admin/documents?upload_source=api")
+        assert response.status_code == 200
+        doc_list = response.json()
+        document_id = doc_list["documents"][0]["document_id"]
+
+        # Get document detail
+        response = client.get(f"/admin/documents/{document_id}")
+        assert response.status_code == 200
+        doc = response.json()
+
+        # Check new fields exist
+        assert "upload_source" in doc
+        assert doc["upload_source"] == "api"
+        assert "batch_id" in doc
+        assert doc["batch_id"] is not None
+        assert "can_annotate" in doc
+        assert isinstance(doc["can_annotate"], bool)
+        assert "csv_field_values" in doc
+        assert "annotation_lock_until" in doc
+
+    def test_document_detail_ui_upload_defaults(self, client, app):
+        """Test UI-uploaded document has correct defaults."""
+        # Get a UI-uploaded document
+        response = client.get("/admin/documents?upload_source=ui")
+        assert response.status_code == 200
+        doc_list = response.json()
+        document_id = doc_list["documents"][0]["document_id"]
+
+        # Get document detail
+        response = client.get(f"/admin/documents/{document_id}")
+        assert response.status_code == 200
+        doc = response.json()
+
+        # UI uploads should have these defaults
+        assert doc["upload_source"] == "ui"
+        assert doc["batch_id"] is None
+        assert doc["csv_field_values"] is None
+        assert doc["can_annotate"] is True
+        assert doc["annotation_lock_until"] is None
+
+    def test_document_detail_with_annotations(self, client, app):
+        """Test document detail includes annotations."""
+        # Get a document with annotations
+        response = client.get("/admin/documents?has_annotations=true")
+        assert response.status_code == 200
+        doc_list = response.json()
+        document_id = doc_list["documents"][0]["document_id"]
+
+        # Get document detail
+        response = client.get(f"/admin/documents/{document_id}")
+        assert response.status_code == 200
+        doc = response.json()
+
+        # Should have annotations
+        assert "annotations" in doc
+        assert len(doc["annotations"]) > 0

tests/web/test_admin_training.py

@@ -0,0 +1,247 @@
+"""
+Tests for Admin Training Routes and Scheduler.
+"""
+
+import pytest
+from datetime import datetime, timedelta
+from unittest.mock import MagicMock, patch
+from uuid import UUID
+
+from src.data.admin_models import TrainingTask, TrainingLog
+from src.web.api.v1.admin.training import _validate_uuid, create_training_router
+from src.web.core.scheduler import (
+    TrainingScheduler,
+    get_training_scheduler,
+    start_scheduler,
+    stop_scheduler,
+)
+from src.web.schemas.admin import (
+    TrainingConfig,
+    TrainingStatus,
+    TrainingTaskCreate,
+    TrainingType,
+)
+
+
+# Test UUIDs
+TEST_TASK_UUID = "770e8400-e29b-41d4-a716-446655440002"
+TEST_TOKEN = "test-admin-token-12345"
+
+
+class TestTrainingRouterCreation:
+    """Tests for training router creation."""
+
+    def test_creates_router_with_endpoints(self):
+        """Test router is created with expected endpoints."""
+        router = create_training_router()
+
+        # Get route paths (include prefix)
+        paths = [route.path for route in router.routes]
+
+        # Paths include the /admin/training prefix
+        assert any("/tasks" in p for p in paths)
+        assert any("{task_id}" in p for p in paths)
+        assert any("cancel" in p for p in paths)
+        assert any("logs" in p for p in paths)
+        assert any("export" in p for p in paths)
+
+
+class TestTrainingConfigSchema:
+    """Tests for TrainingConfig schema."""
+
+    def test_default_config(self):
+        """Test default training configuration."""
+        config = TrainingConfig()
+
+        assert config.model_name == "yolo11n.pt"
+        assert config.epochs == 100
+        assert config.batch_size == 16
+        assert config.image_size == 640
+        assert config.learning_rate == 0.01
+        assert config.device == "0"
+
+    def test_custom_config(self):
+        """Test custom training configuration."""
+        config = TrainingConfig(
+            model_name="yolo11s.pt",
+            epochs=50,
+            batch_size=8,
+            image_size=416,
+            learning_rate=0.001,
+            device="cpu",
+        )
+
+        assert config.model_name == "yolo11s.pt"
+        assert config.epochs == 50
+        assert config.batch_size == 8
+
+    def test_config_validation(self):
+        """Test config validation constraints."""
+        # Epochs must be 1-1000
+        config = TrainingConfig(epochs=1)
+        assert config.epochs == 1
+
+        config = TrainingConfig(epochs=1000)
+        assert config.epochs == 1000
+
+
+class TestTrainingTaskCreateSchema:
+    """Tests for TrainingTaskCreate schema."""
+
+    def test_minimal_task(self):
+        """Test minimal task creation."""
+        task = TrainingTaskCreate(name="Test Training")
+
+        assert task.name == "Test Training"
+        assert task.task_type == TrainingType.TRAIN
+        assert task.description is None
+        assert task.scheduled_at is None
+
+    def test_scheduled_task(self):
+        """Test scheduled task creation."""
+        scheduled_time = datetime.utcnow() + timedelta(hours=1)
+        task = TrainingTaskCreate(
+            name="Scheduled Training",
+            scheduled_at=scheduled_time,
+        )
+
+        assert task.scheduled_at == scheduled_time
+
+    def test_recurring_task(self):
+        """Test recurring task with cron expression."""
+        task = TrainingTaskCreate(
+            name="Recurring Training",
+            cron_expression="0 0 * * 0",  # Every Sunday at midnight
+        )
+
+        assert task.cron_expression == "0 0 * * 0"
+
+
+class TestTrainingTaskModel:
+    """Tests for TrainingTask model."""
+
+    def test_task_creation(self):
+        """Test training task model creation."""
+        task = TrainingTask(
+            admin_token=TEST_TOKEN,
+            name="Test Task",
+            task_type="train",
+            status="pending",
+        )
+
+        assert task.name == "Test Task"
+        assert task.task_type == "train"
+        assert task.status == "pending"
+
+    def test_task_with_config(self):
+        """Test task with configuration."""
+        config = {
+            "model_name": "yolo11n.pt",
+            "epochs": 100,
+        }
+        task = TrainingTask(
+            admin_token=TEST_TOKEN,
+            name="Configured Task",
+            task_type="train",
+            config=config,
+        )
+
+        assert task.config == config
+        assert task.config["epochs"] == 100
+
+
+class TestTrainingLogModel:
+    """Tests for TrainingLog model."""
+
+    def test_log_creation(self):
+        """Test training log creation."""
+        log = TrainingLog(
+            task_id=UUID(TEST_TASK_UUID),
+            level="INFO",
+            message="Training started",
+        )
+
+        assert str(log.task_id) == TEST_TASK_UUID
+        assert log.level == "INFO"
+        assert log.message == "Training started"
+
+    def test_log_with_details(self):
+        """Test log with additional details."""
+        details = {
+            "epoch": 10,
+            "loss": 0.5,
+            "mAP": 0.85,
+        }
+        log = TrainingLog(
+            task_id=UUID(TEST_TASK_UUID),
+            level="INFO",
+            message="Epoch completed",
+            details=details,
+        )
+
+        assert log.details == details
+        assert log.details["epoch"] == 10
+
+
+class TestTrainingScheduler:
+    """Tests for TrainingScheduler."""
+
+    @pytest.fixture
+    def scheduler(self):
+        """Create a scheduler for testing."""
+        return TrainingScheduler(check_interval_seconds=1)
+
+    def test_scheduler_creation(self, scheduler):
+        """Test scheduler creation."""
+        assert scheduler._check_interval == 1
+        assert scheduler._running is False
+        assert scheduler._thread is None
+
+    def test_scheduler_start_stop(self, scheduler):
+        """Test scheduler start and stop."""
+        with patch.object(scheduler, "_check_pending_tasks"):
+            scheduler.start()
+            assert scheduler._running is True
+            assert scheduler._thread is not None
+
+            scheduler.stop()
+            assert scheduler._running is False
+
+    def test_scheduler_singleton(self):
+        """Test get_training_scheduler returns singleton."""
+        # Reset any existing scheduler
+        stop_scheduler()
+
+        s1 = get_training_scheduler()
+        s2 = get_training_scheduler()
+
+        assert s1 is s2
+
+        # Cleanup
+        stop_scheduler()
+
+
+class TestTrainingStatusEnum:
+    """Tests for TrainingStatus enum."""
+
+    def test_all_statuses(self):
+        """Test all training statuses are defined."""
+        statuses = [s.value for s in TrainingStatus]
+
+        assert "pending" in statuses
+        assert "scheduled" in statuses
+        assert "running" in statuses
+        assert "completed" in statuses
+        assert "failed" in statuses
+        assert "cancelled" in statuses
+
+
+class TestTrainingTypeEnum:
+    """Tests for TrainingType enum."""
+
+    def test_all_types(self):
+        """Test all training types are defined."""
+        types = [t.value for t in TrainingType]
+
+        assert "train" in types
+        assert "finetune" in types

tests/web/test_annotation_locks.py

@@ -0,0 +1,276 @@
+"""
+Tests for Annotation Lock Mechanism (Phase 3.3).
+"""
+
+import pytest
+from datetime import datetime, timedelta, timezone
+from uuid import uuid4
+
+from fastapi import FastAPI
+from fastapi.testclient import TestClient
+
+from src.web.api.v1.admin.documents import create_admin_router
+from src.web.core.auth import validate_admin_token, get_admin_db
+
+
+class MockAdminDocument:
+    """Mock AdminDocument for testing."""
+
+    def __init__(self, **kwargs):
+        self.document_id = kwargs.get('document_id', uuid4())
+        self.admin_token = kwargs.get('admin_token', 'test-token')
+        self.filename = kwargs.get('filename', 'test.pdf')
+        self.file_size = kwargs.get('file_size', 100000)
+        self.content_type = kwargs.get('content_type', 'application/pdf')
+        self.page_count = kwargs.get('page_count', 1)
+        self.status = kwargs.get('status', 'pending')
+        self.auto_label_status = kwargs.get('auto_label_status', None)
+        self.auto_label_error = kwargs.get('auto_label_error', None)
+        self.upload_source = kwargs.get('upload_source', 'ui')
+        self.batch_id = kwargs.get('batch_id', None)
+        self.csv_field_values = kwargs.get('csv_field_values', None)
+        self.annotation_lock_until = kwargs.get('annotation_lock_until', None)
+        self.created_at = kwargs.get('created_at', datetime.utcnow())
+        self.updated_at = kwargs.get('updated_at', datetime.utcnow())
+
+
+class MockAdminDB:
+    """Mock AdminDB for testing annotation locks."""
+
+    def __init__(self):
+        self.documents = {}
+
+    def get_document_by_token(self, document_id, admin_token):
+        """Get single document by ID and token."""
+        doc = self.documents.get(document_id)
+        if doc and doc.admin_token == admin_token:
+            return doc
+        return None
+
+    def acquire_annotation_lock(self, document_id, admin_token, duration_seconds=300):
+        """Acquire annotation lock for a document."""
+        doc = self.documents.get(document_id)
+        if not doc or doc.admin_token != admin_token:
+            return None
+
+        # Check if already locked
+        now = datetime.now(timezone.utc)
+        if doc.annotation_lock_until and doc.annotation_lock_until > now:
+            return None
+
+        # Acquire lock
+        doc.annotation_lock_until = now + timedelta(seconds=duration_seconds)
+        return doc
+
+    def release_annotation_lock(self, document_id, admin_token, force=False):
+        """Release annotation lock for a document."""
+        doc = self.documents.get(document_id)
+        if not doc or doc.admin_token != admin_token:
+            return None
+
+        # Release lock
+        doc.annotation_lock_until = None
+        return doc
+
+    def extend_annotation_lock(self, document_id, admin_token, additional_seconds=300):
+        """Extend an existing annotation lock."""
+        doc = self.documents.get(document_id)
+        if not doc or doc.admin_token != admin_token:
+            return None
+
+        # Check if lock exists and is still valid
+        now = datetime.now(timezone.utc)
+        if not doc.annotation_lock_until or doc.annotation_lock_until <= now:
+            return None
+
+        # Extend lock
+        doc.annotation_lock_until = doc.annotation_lock_until + timedelta(seconds=additional_seconds)
+        return doc
+
+
+@pytest.fixture
+def app():
+    """Create test FastAPI app."""
+    app = FastAPI()
+
+    # Create mock DB
+    mock_db = MockAdminDB()
+
+    # Add test document
+    doc1 = MockAdminDocument(
+        filename="INV001.pdf",
+        status="pending",
+        upload_source="ui",
+    )
+
+    mock_db.documents[str(doc1.document_id)] = doc1
+
+    # Override dependencies
+    app.dependency_overrides[validate_admin_token] = lambda: "test-token"
+    app.dependency_overrides[get_admin_db] = lambda: mock_db
+
+    # Include router
+    router = create_admin_router((".pdf", ".png", ".jpg"))
+    app.include_router(router)
+
+    return app
+
+
+@pytest.fixture
+def client(app):
+    """Create test client."""
+    return TestClient(app)
+
+
+@pytest.fixture
+def document_id(app):
+    """Get document ID from the mock DB."""
+    mock_db = app.dependency_overrides[get_admin_db]()
+    return str(list(mock_db.documents.keys())[0])
+
+
+class TestAnnotationLocks:
+    """Tests for annotation lock endpoints."""
+
+    def test_acquire_lock_success(self, client, document_id):
+        """Test successfully acquiring an annotation lock."""
+        response = client.post(
+            f"/admin/documents/{document_id}/lock",
+            json={"duration_seconds": 300}
+        )
+
+        assert response.status_code == 200
+        data = response.json()
+        assert data["document_id"] == document_id
+        assert data["locked"] is True
+        assert data["lock_expires_at"] is not None
+        assert "Lock acquired for 300 seconds" in data["message"]
+
+    def test_acquire_lock_already_locked(self, client, document_id):
+        """Test acquiring lock on already locked document."""
+        # First lock
+        response1 = client.post(
+            f"/admin/documents/{document_id}/lock",
+            json={"duration_seconds": 300}
+        )
+        assert response1.status_code == 200
+
+        # Try to lock again
+        response2 = client.post(
+            f"/admin/documents/{document_id}/lock",
+            json={"duration_seconds": 300}
+        )
+        assert response2.status_code == 409
+        assert "already locked" in response2.json()["detail"]
+
+    def test_release_lock_success(self, client, document_id):
+        """Test successfully releasing an annotation lock."""
+        # First acquire lock
+        client.post(
+            f"/admin/documents/{document_id}/lock",
+            json={"duration_seconds": 300}
+        )
+
+        # Then release it
+        response = client.delete(f"/admin/documents/{document_id}/lock")
+
+        assert response.status_code == 200
+        data = response.json()
+        assert data["document_id"] == document_id
+        assert data["locked"] is False
+        assert data["lock_expires_at"] is None
+        assert "released successfully" in data["message"]
+
+    def test_release_lock_not_locked(self, client, document_id):
+        """Test releasing lock on unlocked document."""
+        response = client.delete(f"/admin/documents/{document_id}/lock")
+
+        # Should succeed even if not locked
+        assert response.status_code == 200
+        data = response.json()
+        assert data["locked"] is False
+
+    def test_extend_lock_success(self, client, document_id):
+        """Test successfully extending an annotation lock."""
+        # First acquire lock
+        response1 = client.post(
+            f"/admin/documents/{document_id}/lock",
+            json={"duration_seconds": 300}
+        )
+        original_expiry = response1.json()["lock_expires_at"]
+
+        # Extend lock
+        response2 = client.patch(
+            f"/admin/documents/{document_id}/lock",
+            json={"duration_seconds": 300}
+        )
+
+        assert response2.status_code == 200
+        data = response2.json()
+        assert data["document_id"] == document_id
+        assert data["locked"] is True
+        assert data["lock_expires_at"] != original_expiry
+        assert "extended by 300 seconds" in data["message"]
+
+    def test_extend_lock_not_locked(self, client, document_id):
+        """Test extending lock on unlocked document."""
+        response = client.patch(
+            f"/admin/documents/{document_id}/lock",
+            json={"duration_seconds": 300}
+        )
+
+        assert response.status_code == 409
+        assert "doesn't exist or has expired" in response.json()["detail"]
+
+    def test_acquire_lock_custom_duration(self, client, document_id):
+        """Test acquiring lock with custom duration."""
+        response = client.post(
+            f"/admin/documents/{document_id}/lock",
+            json={"duration_seconds": 600}
+        )
+
+        assert response.status_code == 200
+        data = response.json()
+        assert "Lock acquired for 600 seconds" in data["message"]
+
+    def test_acquire_lock_invalid_document(self, client):
+        """Test acquiring lock on non-existent document."""
+        fake_id = str(uuid4())
+        response = client.post(
+            f"/admin/documents/{fake_id}/lock",
+            json={"duration_seconds": 300}
+        )
+
+        assert response.status_code == 404
+        assert "not found" in response.json()["detail"]
+
+    def test_lock_lifecycle(self, client, document_id):
+        """Test complete lock lifecycle: acquire -> extend -> release."""
+        # Acquire
+        response1 = client.post(
+            f"/admin/documents/{document_id}/lock",
+            json={"duration_seconds": 300}
+        )
+        assert response1.status_code == 200
+        assert response1.json()["locked"] is True
+
+        # Extend
+        response2 = client.patch(
+            f"/admin/documents/{document_id}/lock",
+            json={"duration_seconds": 300}
+        )
+        assert response2.status_code == 200
+        assert response2.json()["locked"] is True
+
+        # Release
+        response3 = client.delete(f"/admin/documents/{document_id}/lock")
+        assert response3.status_code == 200
+        assert response3.json()["locked"] is False
+
+        # Verify can acquire again after release
+        response4 = client.post(
+            f"/admin/documents/{document_id}/lock",
+            json={"duration_seconds": 300}
+        )
+        assert response4.status_code == 200
+        assert response4.json()["locked"] is True

tests/web/test_annotation_phase5.py

@@ -0,0 +1,420 @@
+"""
+Tests for Phase 5: Annotation Enhancement (Verification and Override)
+"""
+
+import pytest
+from datetime import datetime
+from uuid import uuid4
+
+from fastapi import FastAPI
+from fastapi.testclient import TestClient
+
+from src.web.api.v1.admin.annotations import create_annotation_router
+from src.web.core.auth import validate_admin_token, get_admin_db
+
+
+class MockAdminDocument:
+    """Mock AdminDocument for testing."""
+
+    def __init__(self, **kwargs):
+        self.document_id = kwargs.get('document_id', uuid4())
+        self.admin_token = kwargs.get('admin_token', 'test-token')
+        self.filename = kwargs.get('filename', 'test.pdf')
+        self.file_size = kwargs.get('file_size', 100000)
+        self.content_type = kwargs.get('content_type', 'application/pdf')
+        self.page_count = kwargs.get('page_count', 1)
+        self.status = kwargs.get('status', 'labeled')
+        self.auto_label_status = kwargs.get('auto_label_status', None)
+        self.created_at = kwargs.get('created_at', datetime.utcnow())
+        self.updated_at = kwargs.get('updated_at', datetime.utcnow())
+
+
+class MockAnnotation:
+    """Mock AdminAnnotation for testing."""
+
+    def __init__(self, **kwargs):
+        self.annotation_id = kwargs.get('annotation_id', uuid4())
+        self.document_id = kwargs.get('document_id')
+        self.page_number = kwargs.get('page_number', 1)
+        self.class_id = kwargs.get('class_id', 0)
+        self.class_name = kwargs.get('class_name', 'invoice_number')
+        self.bbox_x = kwargs.get('bbox_x', 100)
+        self.bbox_y = kwargs.get('bbox_y', 100)
+        self.bbox_width = kwargs.get('bbox_width', 200)
+        self.bbox_height = kwargs.get('bbox_height', 50)
+        self.x_center = kwargs.get('x_center', 0.5)
+        self.y_center = kwargs.get('y_center', 0.5)
+        self.width = kwargs.get('width', 0.3)
+        self.height = kwargs.get('height', 0.1)
+        self.text_value = kwargs.get('text_value', 'INV-001')
+        self.confidence = kwargs.get('confidence', 0.95)
+        self.source = kwargs.get('source', 'auto')
+        self.is_verified = kwargs.get('is_verified', False)
+        self.verified_at = kwargs.get('verified_at', None)
+        self.verified_by = kwargs.get('verified_by', None)
+        self.override_source = kwargs.get('override_source', None)
+        self.original_annotation_id = kwargs.get('original_annotation_id', None)
+        self.created_at = kwargs.get('created_at', datetime.utcnow())
+        self.updated_at = kwargs.get('updated_at', datetime.utcnow())
+
+
+class MockAnnotationHistory:
+    """Mock AnnotationHistory for testing."""
+
+    def __init__(self, **kwargs):
+        self.history_id = kwargs.get('history_id', uuid4())
+        self.annotation_id = kwargs.get('annotation_id')
+        self.document_id = kwargs.get('document_id')
+        self.action = kwargs.get('action', 'override')
+        self.previous_value = kwargs.get('previous_value', {})
+        self.new_value = kwargs.get('new_value', {})
+        self.changed_by = kwargs.get('changed_by', 'test-token')
+        self.change_reason = kwargs.get('change_reason', None)
+        self.created_at = kwargs.get('created_at', datetime.utcnow())
+
+
+class MockAdminDB:
+    """Mock AdminDB for testing Phase 5."""
+
+    def __init__(self):
+        self.documents = {}
+        self.annotations = {}
+        self.annotation_history = {}
+
+    def get_document_by_token(self, document_id, admin_token):
+        """Get document by ID and token."""
+        doc = self.documents.get(str(document_id))
+        if doc and doc.admin_token == admin_token:
+            return doc
+        return None
+
+    def verify_annotation(self, annotation_id, admin_token):
+        """Mark annotation as verified."""
+        annotation = self.annotations.get(str(annotation_id))
+        if annotation:
+            annotation.is_verified = True
+            annotation.verified_at = datetime.utcnow()
+            annotation.verified_by = admin_token
+            return annotation
+        return None
+
+    def override_annotation(
+        self,
+        annotation_id,
+        admin_token,
+        change_reason=None,
+        **updates,
+    ):
+        """Override an annotation."""
+        annotation = self.annotations.get(str(annotation_id))
+        if annotation:
+            # Apply updates
+            for key, value in updates.items():
+                if hasattr(annotation, key):
+                    setattr(annotation, key, value)
+
+            # Mark as overridden if was auto-generated
+            if annotation.source == "auto":
+                annotation.override_source = "auto"
+                annotation.source = "manual"
+
+            # Create history record
+            history = MockAnnotationHistory(
+                annotation_id=uuid4().hex if isinstance(annotation_id, str) else annotation_id,
+                document_id=annotation.document_id,
+                action="override",
+                changed_by=admin_token,
+                change_reason=change_reason,
+            )
+            self.annotation_history[str(annotation.annotation_id)] = [history]
+
+            return annotation
+        return None
+
+    def get_annotation_history(self, annotation_id):
+        """Get annotation history."""
+        return self.annotation_history.get(str(annotation_id), [])
+
+
+@pytest.fixture
+def app():
+    """Create test FastAPI app."""
+    app = FastAPI()
+
+    # Create mock DB
+    mock_db = MockAdminDB()
+
+    # Add test document
+    doc1 = MockAdminDocument(
+        filename="TEST001.pdf",
+        status="labeled",
+    )
+    mock_db.documents[str(doc1.document_id)] = doc1
+
+    # Add test annotations
+    ann1 = MockAnnotation(
+        document_id=doc1.document_id,
+        class_id=0,
+        class_name="invoice_number",
+        text_value="INV-001",
+        source="auto",
+        confidence=0.95,
+    )
+    ann2 = MockAnnotation(
+        document_id=doc1.document_id,
+        class_id=6,
+        class_name="amount",
+        text_value="1500.00",
+        source="auto",
+        confidence=0.98,
+    )
+
+    mock_db.annotations[str(ann1.annotation_id)] = ann1
+    mock_db.annotations[str(ann2.annotation_id)] = ann2
+
+    # Store document ID and annotation IDs for tests
+    app.state.document_id = str(doc1.document_id)
+    app.state.annotation_id_1 = str(ann1.annotation_id)
+    app.state.annotation_id_2 = str(ann2.annotation_id)
+
+    # Override dependencies
+    app.dependency_overrides[validate_admin_token] = lambda: "test-token"
+    app.dependency_overrides[get_admin_db] = lambda: mock_db
+
+    # Include router
+    router = create_annotation_router()
+    app.include_router(router)
+
+    return app
+
+
+@pytest.fixture
+def client(app):
+    """Create test client."""
+    return TestClient(app)
+
+
+class TestAnnotationVerification:
+    """Tests for POST /admin/documents/{document_id}/annotations/{annotation_id}/verify endpoint."""
+
+    def test_verify_annotation_success(self, client, app):
+        """Test successfully verifying an annotation."""
+        document_id = app.state.document_id
+        annotation_id = app.state.annotation_id_1
+
+        response = client.post(
+            f"/admin/documents/{document_id}/annotations/{annotation_id}/verify"
+        )
+
+        assert response.status_code == 200
+        data = response.json()
+        assert data["annotation_id"] == annotation_id
+        assert data["is_verified"] is True
+        assert data["verified_at"] is not None
+        assert data["verified_by"] == "test-token"
+        assert "verified successfully" in data["message"].lower()
+
+    def test_verify_annotation_not_found(self, client, app):
+        """Test verifying non-existent annotation."""
+        document_id = app.state.document_id
+        fake_annotation_id = str(uuid4())
+
+        response = client.post(
+            f"/admin/documents/{document_id}/annotations/{fake_annotation_id}/verify"
+        )
+
+        assert response.status_code == 404
+        assert "not found" in response.json()["detail"].lower()
+
+    def test_verify_annotation_document_not_found(self, client):
+        """Test verifying annotation with non-existent document."""
+        fake_document_id = str(uuid4())
+        fake_annotation_id = str(uuid4())
+
+        response = client.post(
+            f"/admin/documents/{fake_document_id}/annotations/{fake_annotation_id}/verify"
+        )
+
+        assert response.status_code == 404
+        assert "not found" in response.json()["detail"].lower()
+
+    def test_verify_annotation_invalid_uuid(self, client, app):
+        """Test verifying annotation with invalid UUID format."""
+        document_id = app.state.document_id
+
+        response = client.post(
+            f"/admin/documents/{document_id}/annotations/invalid-uuid/verify"
+        )
+
+        assert response.status_code == 400
+        assert "invalid" in response.json()["detail"].lower()
+
+
+class TestAnnotationOverride:
+    """Tests for PATCH /admin/documents/{document_id}/annotations/{annotation_id}/override endpoint."""
+
+    def test_override_annotation_text_value(self, client, app):
+        """Test overriding annotation text value."""
+        document_id = app.state.document_id
+        annotation_id = app.state.annotation_id_1
+
+        response = client.patch(
+            f"/admin/documents/{document_id}/annotations/{annotation_id}/override",
+            json={
+                "text_value": "INV-001-CORRECTED",
+                "reason": "OCR error correction"
+            }
+        )
+
+        assert response.status_code == 200
+        data = response.json()
+        assert data["annotation_id"] == annotation_id
+        assert data["source"] == "manual"
+        assert data["override_source"] == "auto"
+        assert "successfully" in data["message"].lower()
+        assert "history_id" in data
+
+    def test_override_annotation_bbox(self, client, app):
+        """Test overriding annotation bounding box."""
+        document_id = app.state.document_id
+        annotation_id = app.state.annotation_id_1
+
+        response = client.patch(
+            f"/admin/documents/{document_id}/annotations/{annotation_id}/override",
+            json={
+                "bbox": {
+                    "x": 110,
+                    "y": 205,
+                    "width": 195,
+                    "height": 48
+                },
+                "reason": "Bbox adjustment"
+            }
+        )
+
+        assert response.status_code == 200
+        data = response.json()
+        assert data["annotation_id"] == annotation_id
+        assert data["source"] == "manual"
+
+    def test_override_annotation_class(self, client, app):
+        """Test overriding annotation class."""
+        document_id = app.state.document_id
+        annotation_id = app.state.annotation_id_1
+
+        response = client.patch(
+            f"/admin/documents/{document_id}/annotations/{annotation_id}/override",
+            json={
+                "class_id": 1,
+                "class_name": "invoice_date",
+                "reason": "Wrong field classification"
+            }
+        )
+
+        assert response.status_code == 200
+        data = response.json()
+        assert data["annotation_id"] == annotation_id
+
+    def test_override_annotation_multiple_fields(self, client, app):
+        """Test overriding multiple annotation fields at once."""
+        document_id = app.state.document_id
+        annotation_id = app.state.annotation_id_2
+
+        response = client.patch(
+            f"/admin/documents/{document_id}/annotations/{annotation_id}/override",
+            json={
+                "text_value": "1550.00",
+                "bbox": {
+                    "x": 120,
+                    "y": 210,
+                    "width": 180,
+                    "height": 45
+                },
+                "reason": "Multiple corrections"
+            }
+        )
+
+        assert response.status_code == 200
+        data = response.json()
+        assert data["annotation_id"] == annotation_id
+
+    def test_override_annotation_no_updates(self, client, app):
+        """Test overriding annotation without providing any updates."""
+        document_id = app.state.document_id
+        annotation_id = app.state.annotation_id_1
+
+        response = client.patch(
+            f"/admin/documents/{document_id}/annotations/{annotation_id}/override",
+            json={}
+        )
+
+        assert response.status_code == 400
+        assert "no updates" in response.json()["detail"].lower()
+
+    def test_override_annotation_not_found(self, client, app):
+        """Test overriding non-existent annotation."""
+        document_id = app.state.document_id
+        fake_annotation_id = str(uuid4())
+
+        response = client.patch(
+            f"/admin/documents/{document_id}/annotations/{fake_annotation_id}/override",
+            json={
+                "text_value": "TEST"
+            }
+        )
+
+        assert response.status_code == 404
+        assert "not found" in response.json()["detail"].lower()
+
+    def test_override_annotation_document_not_found(self, client):
+        """Test overriding annotation with non-existent document."""
+        fake_document_id = str(uuid4())
+        fake_annotation_id = str(uuid4())
+
+        response = client.patch(
+            f"/admin/documents/{fake_document_id}/annotations/{fake_annotation_id}/override",
+            json={
+                "text_value": "TEST"
+            }
+        )
+
+        assert response.status_code == 404
+        assert "not found" in response.json()["detail"].lower()
+
+    def test_override_annotation_creates_history(self, client, app):
+        """Test that overriding annotation creates history record."""
+        document_id = app.state.document_id
+        annotation_id = app.state.annotation_id_1
+
+        response = client.patch(
+            f"/admin/documents/{document_id}/annotations/{annotation_id}/override",
+            json={
+                "text_value": "INV-CORRECTED",
+                "reason": "Test history creation"
+            }
+        )
+
+        assert response.status_code == 200
+        data = response.json()
+        # History ID should be present and valid
+        assert "history_id" in data
+        assert data["history_id"] != ""
+
+    def test_override_annotation_with_reason(self, client, app):
+        """Test overriding annotation with change reason."""
+        document_id = app.state.document_id
+        annotation_id = app.state.annotation_id_1
+
+        change_reason = "Correcting OCR misread"
+        response = client.patch(
+            f"/admin/documents/{document_id}/annotations/{annotation_id}/override",
+            json={
+                "text_value": "INV-002",
+                "reason": change_reason
+            }
+        )
+
+        assert response.status_code == 200
+        # Reason is stored in history, not returned in response
+        data = response.json()
+        assert data["annotation_id"] == annotation_id

tests/web/test_async_queue.py

@@ -0,0 +1,217 @@
+"""
+Tests for the AsyncTaskQueue class.
+"""
+
+import tempfile
+import time
+from datetime import datetime
+from pathlib import Path
+from threading import Event
+from unittest.mock import MagicMock
+
+import pytest
+
+from src.web.workers.async_queue import AsyncTask, AsyncTaskQueue
+
+
+class TestAsyncTask:
+    """Tests for AsyncTask dataclass."""
+
+    def test_create_task(self):
+        """Test creating an AsyncTask."""
+        task = AsyncTask(
+            request_id="test-id",
+            api_key="test-key",
+            file_path=Path("/tmp/test.pdf"),
+            filename="test.pdf",
+        )
+
+        assert task.request_id == "test-id"
+        assert task.api_key == "test-key"
+        assert task.filename == "test.pdf"
+        assert task.priority == 0
+        assert task.created_at is not None
+
+
+class TestAsyncTaskQueue:
+    """Tests for AsyncTaskQueue."""
+
+    def test_init(self):
+        """Test queue initialization."""
+        queue = AsyncTaskQueue(max_size=50, worker_count=2)
+
+        assert queue._worker_count == 2
+        assert queue._queue.maxsize == 50
+        assert not queue._started
+
+    def test_submit_task(self, task_queue, sample_task):
+        """Test submitting a task to the queue."""
+        success = task_queue.submit(sample_task)
+
+        assert success is True
+        assert task_queue.get_queue_depth() == 1
+
+    def test_submit_when_full(self, sample_task):
+        """Test submitting to a full queue."""
+        queue = AsyncTaskQueue(max_size=1, worker_count=1)
+
+        # Submit first task
+        queue.submit(sample_task)
+
+        # Create second task
+        task2 = AsyncTask(
+            request_id="test-2",
+            api_key="test-key",
+            file_path=sample_task.file_path,
+            filename="test2.pdf",
+        )
+
+        # Queue should be full
+        success = queue.submit(task2)
+        assert success is False
+
+    def test_get_queue_depth(self, task_queue, sample_task):
+        """Test getting queue depth."""
+        assert task_queue.get_queue_depth() == 0
+
+        task_queue.submit(sample_task)
+        assert task_queue.get_queue_depth() == 1
+
+    def test_start_and_stop(self, task_queue):
+        """Test starting and stopping the queue."""
+        handler = MagicMock()
+
+        task_queue.start(handler)
+        assert task_queue._started is True
+        assert task_queue.is_running is True
+        assert len(task_queue._workers) == 1
+
+        task_queue.stop(timeout=5.0)
+        assert task_queue._started is False
+        assert task_queue.is_running is False
+        assert len(task_queue._workers) == 0
+
+    def test_worker_processes_task(self, sample_task):
+        """Test that worker thread processes tasks."""
+        queue = AsyncTaskQueue(max_size=10, worker_count=1)
+        processed = Event()
+
+        def handler(task):
+            processed.set()
+
+        queue.start(handler)
+        queue.submit(sample_task)
+
+        # Wait for processing
+        assert processed.wait(timeout=5.0)
+
+        queue.stop()
+
+    def test_worker_handles_errors(self, sample_task):
+        """Test that worker handles errors gracefully."""
+        queue = AsyncTaskQueue(max_size=10, worker_count=1)
+        error_handled = Event()
+
+        def failing_handler(task):
+            error_handled.set()
+            raise ValueError("Test error")
+
+        queue.start(failing_handler)
+        queue.submit(sample_task)
+
+        # Should not crash
+        assert error_handled.wait(timeout=5.0)
+        time.sleep(0.5)  # Give time for error handling
+
+        assert queue.is_running
+
+        queue.stop()
+
+    def test_processing_tracking(self, task_queue, sample_task):
+        """Test tracking of processing tasks."""
+        processed = Event()
+
+        def slow_handler(task):
+            processed.set()
+            time.sleep(0.5)
+
+        task_queue.start(slow_handler)
+        task_queue.submit(sample_task)
+
+        # Wait for processing to start
+        assert processed.wait(timeout=5.0)
+
+        # Task should be in processing set
+        assert task_queue.get_processing_count() == 1
+        assert task_queue.is_processing(sample_task.request_id)
+
+        # Wait for completion
+        time.sleep(1.0)
+
+        assert task_queue.get_processing_count() == 0
+        assert not task_queue.is_processing(sample_task.request_id)
+
+        task_queue.stop()
+
+    def test_multiple_workers(self, sample_task):
+        """Test queue with multiple workers."""
+        queue = AsyncTaskQueue(max_size=10, worker_count=3)
+        processed_count = []
+
+        def handler(task):
+            processed_count.append(task.request_id)
+            time.sleep(0.1)
+
+        queue.start(handler)
+
+        # Submit multiple tasks
+        for i in range(5):
+            task = AsyncTask(
+                request_id=f"task-{i}",
+                api_key="test-key",
+                file_path=sample_task.file_path,
+                filename=f"test-{i}.pdf",
+            )
+            queue.submit(task)
+
+        # Wait for all tasks
+        time.sleep(2.0)
+
+        assert len(processed_count) == 5
+
+        queue.stop()
+
+    def test_graceful_shutdown(self, sample_task):
+        """Test graceful shutdown waits for current task."""
+        queue = AsyncTaskQueue(max_size=10, worker_count=1)
+        started = Event()
+        finished = Event()
+
+        def slow_handler(task):
+            started.set()
+            time.sleep(0.5)
+            finished.set()
+
+        queue.start(slow_handler)
+        queue.submit(sample_task)
+
+        # Wait for processing to start
+        assert started.wait(timeout=5.0)
+
+        # Stop should wait for task to finish
+        queue.stop(timeout=5.0)
+
+        assert finished.is_set()
+
+    def test_double_start(self, task_queue):
+        """Test that starting twice doesn't create duplicate workers."""
+        handler = MagicMock()
+
+        task_queue.start(handler)
+        assert len(task_queue._workers) == 1
+
+        # Starting again should not add more workers
+        task_queue.start(handler)
+        assert len(task_queue._workers) == 1
+
+        task_queue.stop()

tests/web/test_async_routes.py

@@ -0,0 +1,409 @@
+"""
+Tests for the async API routes.
+"""
+
+import tempfile
+from datetime import datetime, timedelta
+from pathlib import Path
+from unittest.mock import MagicMock, patch
+
+import pytest
+from fastapi import FastAPI
+from fastapi.testclient import TestClient
+
+from src.data.async_request_db import ApiKeyConfig, AsyncRequest, AsyncRequestDB
+from src.web.api.v1.async_api.routes import create_async_router, set_async_service
+from src.web.services.async_processing import AsyncSubmitResult
+from src.web.dependencies import init_dependencies
+from src.web.rate_limiter import RateLimiter, RateLimitStatus
+from src.web.schemas.inference import AsyncStatus
+
+# Valid UUID for testing
+TEST_REQUEST_UUID = "550e8400-e29b-41d4-a716-446655440000"
+INVALID_UUID = "nonexistent-id"
+
+
+@pytest.fixture
+def mock_async_service():
+    """Create a mock AsyncProcessingService."""
+    service = MagicMock()
+
+    # Mock config
+    mock_config = MagicMock()
+    mock_config.max_file_size_mb = 50
+    service._async_config = mock_config
+
+    # Default submit result
+    service.submit_request.return_value = AsyncSubmitResult(
+        success=True,
+        request_id="test-request-id",
+        estimated_wait_seconds=30,
+    )
+
+    return service
+
+
+@pytest.fixture
+def mock_rate_limiter(mock_db):
+    """Create a mock RateLimiter."""
+    limiter = MagicMock(spec=RateLimiter)
+
+    # Default: allow all requests
+    limiter.check_submit_limit.return_value = RateLimitStatus(
+        allowed=True,
+        remaining_requests=9,
+        reset_at=datetime.utcnow() + timedelta(seconds=60),
+    )
+    limiter.check_poll_limit.return_value = RateLimitStatus(
+        allowed=True,
+        remaining_requests=999,
+        reset_at=datetime.utcnow(),
+    )
+    limiter.get_rate_limit_headers.return_value = {}
+
+    return limiter
+
+
+@pytest.fixture
+def app(mock_db, mock_rate_limiter, mock_async_service):
+    """Create a test FastAPI app with async routes."""
+    app = FastAPI()
+
+    # Initialize dependencies
+    init_dependencies(mock_db, mock_rate_limiter)
+    set_async_service(mock_async_service)
+
+    # Add routes
+    router = create_async_router(allowed_extensions=(".pdf", ".png", ".jpg", ".jpeg"))
+    app.include_router(router, prefix="/api/v1")
+
+    return app
+
+
+@pytest.fixture
+def client(app):
+    """Create a test client."""
+    return TestClient(app)
+
+
+class TestAsyncSubmitEndpoint:
+    """Tests for POST /api/v1/async/submit."""
+
+    def test_submit_success(self, client, mock_async_service):
+        """Test successful submission."""
+        with tempfile.NamedTemporaryFile(suffix=".pdf", delete=False) as f:
+            f.write(b"fake pdf content")
+            f.seek(0)
+
+            response = client.post(
+                "/api/v1/async/submit",
+                files={"file": ("test.pdf", f, "application/pdf")},
+                headers={"X-API-Key": "test-api-key"},
+            )
+
+        assert response.status_code == 200
+        data = response.json()
+        assert data["status"] == "accepted"
+        assert data["request_id"] == "test-request-id"
+        assert "poll_url" in data
+
+    def test_submit_missing_api_key(self, client):
+        """Test submission without API key."""
+        with tempfile.NamedTemporaryFile(suffix=".pdf", delete=False) as f:
+            f.write(b"fake pdf content")
+            f.seek(0)
+
+            response = client.post(
+                "/api/v1/async/submit",
+                files={"file": ("test.pdf", f, "application/pdf")},
+            )
+
+        assert response.status_code == 401
+        assert "X-API-Key" in response.json()["detail"]
+
+    def test_submit_invalid_api_key(self, client, mock_db):
+        """Test submission with invalid API key."""
+        mock_db.is_valid_api_key.return_value = False
+
+        with tempfile.NamedTemporaryFile(suffix=".pdf", delete=False) as f:
+            f.write(b"fake pdf content")
+            f.seek(0)
+
+            response = client.post(
+                "/api/v1/async/submit",
+                files={"file": ("test.pdf", f, "application/pdf")},
+                headers={"X-API-Key": "invalid-key"},
+            )
+
+        assert response.status_code == 401
+
+    def test_submit_unsupported_file_type(self, client):
+        """Test submission with unsupported file type."""
+        with tempfile.NamedTemporaryFile(suffix=".txt", delete=False) as f:
+            f.write(b"text content")
+            f.seek(0)
+
+            response = client.post(
+                "/api/v1/async/submit",
+                files={"file": ("test.txt", f, "text/plain")},
+                headers={"X-API-Key": "test-api-key"},
+            )
+
+        assert response.status_code == 400
+        assert "Unsupported file type" in response.json()["detail"]
+
+    def test_submit_rate_limited(self, client, mock_rate_limiter):
+        """Test submission when rate limited."""
+        mock_rate_limiter.check_submit_limit.return_value = RateLimitStatus(
+            allowed=False,
+            remaining_requests=0,
+            reset_at=datetime.utcnow() + timedelta(seconds=30),
+            retry_after_seconds=30,
+            reason="Rate limit exceeded",
+        )
+        mock_rate_limiter.get_rate_limit_headers.return_value = {"Retry-After": "30"}
+
+        with tempfile.NamedTemporaryFile(suffix=".pdf", delete=False) as f:
+            f.write(b"fake pdf content")
+            f.seek(0)
+
+            response = client.post(
+                "/api/v1/async/submit",
+                files={"file": ("test.pdf", f, "application/pdf")},
+                headers={"X-API-Key": "test-api-key"},
+            )
+
+        assert response.status_code == 429
+        assert "Retry-After" in response.headers
+
+    def test_submit_queue_full(self, client, mock_async_service):
+        """Test submission when queue is full."""
+        mock_async_service.submit_request.return_value = AsyncSubmitResult(
+            success=False,
+            request_id="test-id",
+            error="Processing queue is full",
+        )
+
+        with tempfile.NamedTemporaryFile(suffix=".pdf", delete=False) as f:
+            f.write(b"fake pdf content")
+            f.seek(0)
+
+            response = client.post(
+                "/api/v1/async/submit",
+                files={"file": ("test.pdf", f, "application/pdf")},
+                headers={"X-API-Key": "test-api-key"},
+            )
+
+        assert response.status_code == 503
+
+
+class TestAsyncStatusEndpoint:
+    """Tests for GET /api/v1/async/status/{request_id}."""
+
+    def test_get_status_pending(self, client, mock_db, sample_async_request):
+        """Test getting status of pending request."""
+        mock_db.get_request_by_api_key.return_value = sample_async_request
+        mock_db.get_queue_position.return_value = 3
+
+        response = client.get(
+            "/api/v1/async/status/550e8400-e29b-41d4-a716-446655440000",
+            headers={"X-API-Key": "test-api-key"},
+        )
+
+        assert response.status_code == 200
+        data = response.json()
+        assert data["status"] == "pending"
+        assert data["position_in_queue"] == 3
+        assert data["result_url"] is None
+
+    def test_get_status_completed(self, client, mock_db, sample_async_request):
+        """Test getting status of completed request."""
+        sample_async_request.status = "completed"
+        sample_async_request.completed_at = datetime.utcnow()
+        mock_db.get_request_by_api_key.return_value = sample_async_request
+
+        response = client.get(
+            "/api/v1/async/status/550e8400-e29b-41d4-a716-446655440000",
+            headers={"X-API-Key": "test-api-key"},
+        )
+
+        assert response.status_code == 200
+        data = response.json()
+        assert data["status"] == "completed"
+        assert data["result_url"] is not None
+
+    def test_get_status_not_found(self, client, mock_db):
+        """Test getting status of non-existent request."""
+        mock_db.get_request_by_api_key.return_value = None
+
+        response = client.get(
+            "/api/v1/async/status/00000000-0000-0000-0000-000000000000",
+            headers={"X-API-Key": "test-api-key"},
+        )
+
+        assert response.status_code == 404
+
+    def test_get_status_wrong_api_key(self, client, mock_db, sample_async_request):
+        """Test that requests are isolated by API key."""
+        # Request belongs to different API key
+        mock_db.get_request_by_api_key.return_value = None
+
+        response = client.get(
+            "/api/v1/async/status/550e8400-e29b-41d4-a716-446655440000",
+            headers={"X-API-Key": "different-api-key"},
+        )
+
+        assert response.status_code == 404
+
+
+class TestAsyncResultEndpoint:
+    """Tests for GET /api/v1/async/result/{request_id}."""
+
+    def test_get_result_completed(self, client, mock_db, sample_async_request):
+        """Test getting result of completed request."""
+        sample_async_request.status = "completed"
+        sample_async_request.completed_at = datetime.utcnow()
+        sample_async_request.processing_time_ms = 1234.5
+        sample_async_request.result = {
+            "document_id": "test-doc",
+            "success": True,
+            "document_type": "invoice",
+            "fields": {"InvoiceNumber": "12345"},
+            "confidence": {"InvoiceNumber": 0.95},
+            "detections": [],
+            "errors": [],
+        }
+        mock_db.get_request_by_api_key.return_value = sample_async_request
+
+        response = client.get(
+            "/api/v1/async/result/550e8400-e29b-41d4-a716-446655440000",
+            headers={"X-API-Key": "test-api-key"},
+        )
+
+        assert response.status_code == 200
+        data = response.json()
+        assert data["status"] == "completed"
+        assert data["result"] is not None
+        assert data["result"]["fields"]["InvoiceNumber"] == "12345"
+
+    def test_get_result_not_completed(self, client, mock_db, sample_async_request):
+        """Test getting result of pending request."""
+        mock_db.get_request_by_api_key.return_value = sample_async_request
+
+        response = client.get(
+            "/api/v1/async/result/550e8400-e29b-41d4-a716-446655440000",
+            headers={"X-API-Key": "test-api-key"},
+        )
+
+        assert response.status_code == 409
+        assert "not yet completed" in response.json()["detail"]
+
+    def test_get_result_failed(self, client, mock_db, sample_async_request):
+        """Test getting result of failed request."""
+        sample_async_request.status = "failed"
+        sample_async_request.error_message = "Processing failed"
+        sample_async_request.processing_time_ms = 500.0
+        mock_db.get_request_by_api_key.return_value = sample_async_request
+
+        response = client.get(
+            "/api/v1/async/result/550e8400-e29b-41d4-a716-446655440000",
+            headers={"X-API-Key": "test-api-key"},
+        )
+
+        assert response.status_code == 200
+        data = response.json()
+        assert data["status"] == "failed"
+
+
+class TestAsyncListEndpoint:
+    """Tests for GET /api/v1/async/requests."""
+
+    def test_list_requests(self, client, mock_db, sample_async_request):
+        """Test listing requests."""
+        mock_db.get_requests_by_api_key.return_value = ([sample_async_request], 1)
+
+        response = client.get(
+            "/api/v1/async/requests",
+            headers={"X-API-Key": "test-api-key"},
+        )
+
+        assert response.status_code == 200
+        data = response.json()
+        assert data["total"] == 1
+        assert len(data["requests"]) == 1
+
+    def test_list_requests_with_status_filter(self, client, mock_db):
+        """Test listing requests with status filter."""
+        mock_db.get_requests_by_api_key.return_value = ([], 0)
+
+        response = client.get(
+            "/api/v1/async/requests?status=completed",
+            headers={"X-API-Key": "test-api-key"},
+        )
+
+        assert response.status_code == 200
+        mock_db.get_requests_by_api_key.assert_called_once()
+        call_kwargs = mock_db.get_requests_by_api_key.call_args[1]
+        assert call_kwargs["status"] == "completed"
+
+    def test_list_requests_pagination(self, client, mock_db):
+        """Test listing requests with pagination."""
+        mock_db.get_requests_by_api_key.return_value = ([], 0)
+
+        response = client.get(
+            "/api/v1/async/requests?limit=50&offset=10",
+            headers={"X-API-Key": "test-api-key"},
+        )
+
+        assert response.status_code == 200
+        call_kwargs = mock_db.get_requests_by_api_key.call_args[1]
+        assert call_kwargs["limit"] == 50
+        assert call_kwargs["offset"] == 10
+
+    def test_list_requests_invalid_status(self, client, mock_db):
+        """Test listing with invalid status filter."""
+        response = client.get(
+            "/api/v1/async/requests?status=invalid",
+            headers={"X-API-Key": "test-api-key"},
+        )
+
+        assert response.status_code == 400
+
+
+class TestAsyncDeleteEndpoint:
+    """Tests for DELETE /api/v1/async/requests/{request_id}."""
+
+    def test_delete_pending_request(self, client, mock_db, sample_async_request):
+        """Test deleting a pending request."""
+        mock_db.get_request_by_api_key.return_value = sample_async_request
+
+        response = client.delete(
+            "/api/v1/async/requests/550e8400-e29b-41d4-a716-446655440000",
+            headers={"X-API-Key": "test-api-key"},
+        )
+
+        assert response.status_code == 200
+        assert response.json()["status"] == "deleted"
+
+    def test_delete_processing_request(self, client, mock_db, sample_async_request):
+        """Test that processing requests cannot be deleted."""
+        sample_async_request.status = "processing"
+        mock_db.get_request_by_api_key.return_value = sample_async_request
+
+        response = client.delete(
+            "/api/v1/async/requests/550e8400-e29b-41d4-a716-446655440000",
+            headers={"X-API-Key": "test-api-key"},
+        )
+
+        assert response.status_code == 409
+
+    def test_delete_not_found(self, client, mock_db):
+        """Test deleting non-existent request."""
+        mock_db.get_request_by_api_key.return_value = None
+
+        response = client.delete(
+            "/api/v1/async/requests/00000000-0000-0000-0000-000000000000",
+            headers={"X-API-Key": "test-api-key"},
+        )
+
+        assert response.status_code == 404

tests/web/test_async_service.py

@@ -0,0 +1,266 @@
+"""
+Tests for the AsyncProcessingService class.
+"""
+
+import tempfile
+import time
+from datetime import datetime, timedelta
+from pathlib import Path
+from unittest.mock import MagicMock, patch
+
+import pytest
+
+from src.data.async_request_db import AsyncRequest
+from src.web.workers.async_queue import AsyncTask, AsyncTaskQueue
+from src.web.services.async_processing import AsyncProcessingService, AsyncSubmitResult
+from src.web.config import AsyncConfig, StorageConfig
+from src.web.rate_limiter import RateLimiter
+
+
+@pytest.fixture
+def async_service(mock_db, mock_inference_service, rate_limiter, storage_config):
+    """Create an AsyncProcessingService for testing."""
+    with tempfile.TemporaryDirectory() as tmpdir:
+        async_config = AsyncConfig(
+            queue_max_size=10,
+            worker_count=1,
+            task_timeout_seconds=30,
+            result_retention_days=7,
+            temp_upload_dir=Path(tmpdir) / "async",
+            max_file_size_mb=10,
+        )
+
+        queue = AsyncTaskQueue(max_size=10, worker_count=1)
+
+        service = AsyncProcessingService(
+            inference_service=mock_inference_service,
+            db=mock_db,
+            queue=queue,
+            rate_limiter=rate_limiter,
+            async_config=async_config,
+            storage_config=storage_config,
+        )
+
+        yield service
+
+        # Cleanup
+        if service._queue._started:
+            service.stop()
+
+
+class TestAsyncProcessingService:
+    """Tests for AsyncProcessingService."""
+
+    def test_submit_request_success(self, async_service, mock_db):
+        """Test successful request submission."""
+        mock_db.create_request.return_value = "test-request-id"
+
+        result = async_service.submit_request(
+            api_key="test-api-key",
+            file_content=b"fake pdf content",
+            filename="test.pdf",
+            content_type="application/pdf",
+        )
+
+        assert result.success is True
+        assert result.request_id is not None
+        assert result.estimated_wait_seconds >= 0
+        assert result.error is None
+
+    def test_submit_request_creates_db_record(self, async_service, mock_db):
+        """Test that submission creates database record."""
+        async_service.submit_request(
+            api_key="test-api-key",
+            file_content=b"fake pdf content",
+            filename="test.pdf",
+            content_type="application/pdf",
+        )
+
+        mock_db.create_request.assert_called_once()
+        call_kwargs = mock_db.create_request.call_args[1]
+        assert call_kwargs["api_key"] == "test-api-key"
+        assert call_kwargs["filename"] == "test.pdf"
+        assert call_kwargs["content_type"] == "application/pdf"
+
+    def test_submit_request_saves_file(self, async_service, mock_db):
+        """Test that submission saves file to temp directory."""
+        content = b"fake pdf content"
+
+        result = async_service.submit_request(
+            api_key="test-api-key",
+            file_content=content,
+            filename="test.pdf",
+            content_type="application/pdf",
+        )
+
+        # File should exist in temp directory
+        temp_dir = async_service._async_config.temp_upload_dir
+        files = list(temp_dir.iterdir())
+
+        # Note: file may be cleaned up quickly if queue processes it
+        # So we just check that the operation succeeded
+        assert result.success is True
+
+    def test_submit_request_records_rate_limit(self, async_service, mock_db, rate_limiter):
+        """Test that submission records rate limit event."""
+        async_service.submit_request(
+            api_key="test-api-key",
+            file_content=b"fake pdf content",
+            filename="test.pdf",
+            content_type="application/pdf",
+        )
+
+        # Rate limiter should have recorded the request
+        mock_db.record_rate_limit_event.assert_called()
+
+    def test_start_and_stop(self, async_service):
+        """Test starting and stopping the service."""
+        async_service.start()
+
+        assert async_service._queue._started is True
+        assert async_service._cleanup_thread is not None
+        assert async_service._cleanup_thread.is_alive()
+
+        async_service.stop()
+
+        assert async_service._queue._started is False
+
+    def test_process_task_success(self, async_service, mock_db, mock_inference_service, sample_task):
+        """Test successful task processing."""
+        async_service._process_task(sample_task)
+
+        # Should update status to processing
+        mock_db.update_status.assert_called_with(sample_task.request_id, "processing")
+
+        # Should complete the request
+        mock_db.complete_request.assert_called_once()
+        call_kwargs = mock_db.complete_request.call_args[1]
+        assert call_kwargs["request_id"] == sample_task.request_id
+        assert "document_id" in call_kwargs
+
+    def test_process_task_pdf(self, async_service, mock_db, mock_inference_service, sample_task):
+        """Test processing a PDF task."""
+        async_service._process_task(sample_task)
+
+        # Should call process_pdf for .pdf files
+        mock_inference_service.process_pdf.assert_called_once()
+
+    def test_process_task_image(self, async_service, mock_db, mock_inference_service):
+        """Test processing an image task."""
+        with tempfile.NamedTemporaryFile(suffix=".png", delete=False) as f:
+            f.write(b"fake image content")
+            task = AsyncTask(
+                request_id="image-task",
+                api_key="test-api-key",
+                file_path=Path(f.name),
+                filename="test.png",
+            )
+
+        async_service._process_task(task)
+
+        # Should call process_image for image files
+        mock_inference_service.process_image.assert_called_once()
+
+    def test_process_task_failure(self, async_service, mock_db, mock_inference_service, sample_task):
+        """Test task processing failure."""
+        mock_inference_service.process_pdf.side_effect = Exception("Processing failed")
+
+        async_service._process_task(sample_task)
+
+        # Should update status to failed
+        mock_db.update_status.assert_called()
+        last_call = mock_db.update_status.call_args_list[-1]
+        assert last_call[0][1] == "failed"  # status
+        assert "Processing failed" in last_call[1]["error_message"]
+
+    def test_process_task_file_not_found(self, async_service, mock_db):
+        """Test task processing with missing file."""
+        task = AsyncTask(
+            request_id="missing-file-task",
+            api_key="test-api-key",
+            file_path=Path("/nonexistent/file.pdf"),
+            filename="test.pdf",
+        )
+
+        async_service._process_task(task)
+
+        # Should fail with file not found
+        mock_db.update_status.assert_called()
+        last_call = mock_db.update_status.call_args_list[-1]
+        assert last_call[0][1] == "failed"
+
+    def test_process_task_cleans_up_file(self, async_service, mock_db, mock_inference_service):
+        """Test that task processing cleans up the uploaded file."""
+        with tempfile.NamedTemporaryFile(suffix=".pdf", delete=False) as f:
+            f.write(b"fake pdf content")
+            file_path = Path(f.name)
+
+        task = AsyncTask(
+            request_id="cleanup-task",
+            api_key="test-api-key",
+            file_path=file_path,
+            filename="test.pdf",
+        )
+
+        async_service._process_task(task)
+
+        # File should be deleted
+        assert not file_path.exists()
+
+    def test_estimate_wait(self, async_service):
+        """Test wait time estimation."""
+        # Empty queue
+        wait = async_service._estimate_wait()
+        assert wait == 0
+
+    def test_cleanup_orphan_files(self, async_service, mock_db):
+        """Test cleanup of orphan files."""
+        # Create an orphan file
+        temp_dir = async_service._async_config.temp_upload_dir
+        orphan_file = temp_dir / "orphan-request.pdf"
+        orphan_file.write_bytes(b"orphan content")
+
+        # Set file mtime to old
+        import os
+        old_time = time.time() - 7200
+        os.utime(orphan_file, (old_time, old_time))
+
+        # Mock database to say file doesn't exist
+        mock_db.get_request.return_value = None
+
+        count = async_service._cleanup_orphan_files()
+
+        assert count == 1
+        assert not orphan_file.exists()
+
+    def test_save_upload(self, async_service):
+        """Test saving uploaded file."""
+        content = b"test content"
+
+        file_path = async_service._save_upload(
+            request_id="test-save",
+            filename="test.pdf",
+            content=content,
+        )
+
+        assert file_path.exists()
+        assert file_path.read_bytes() == content
+        assert file_path.suffix == ".pdf"
+
+        # Cleanup
+        file_path.unlink()
+
+    def test_save_upload_preserves_extension(self, async_service):
+        """Test that save_upload preserves file extension."""
+        content = b"test content"
+
+        # Test various extensions
+        for ext in [".pdf", ".png", ".jpg", ".jpeg"]:
+            file_path = async_service._save_upload(
+                request_id=f"test-{ext}",
+                filename=f"test{ext}",
+                content=content,
+            )
+
+            assert file_path.suffix == ext
+            file_path.unlink()

tests/web/test_autolabel_with_locks.py

@@ -0,0 +1,250 @@
+"""
+Tests for Auto-Label Service with Annotation Lock Integration (Phase 3.5).
+"""
+
+import pytest
+from datetime import datetime, timedelta, timezone
+from pathlib import Path
+from unittest.mock import Mock, MagicMock
+from uuid import uuid4
+
+from src.web.services.autolabel import AutoLabelService
+from src.data.admin_db import AdminDB
+
+
+class MockDocument:
+    """Mock document for testing."""
+
+    def __init__(self, document_id, annotation_lock_until=None):
+        self.document_id = document_id
+        self.annotation_lock_until = annotation_lock_until
+        self.status = "pending"
+        self.auto_label_status = None
+        self.auto_label_error = None
+
+
+class MockAdminDB:
+    """Mock AdminDB for testing."""
+
+    def __init__(self):
+        self.documents = {}
+        self.annotations = []
+        self.status_updates = []
+
+    def get_document(self, document_id):
+        """Get document by ID."""
+        return self.documents.get(str(document_id))
+
+    def update_document_status(
+        self,
+        document_id,
+        status=None,
+        auto_label_status=None,
+        auto_label_error=None,
+    ):
+        """Mock status update."""
+        self.status_updates.append({
+            "document_id": document_id,
+            "status": status,
+            "auto_label_status": auto_label_status,
+            "auto_label_error": auto_label_error,
+        })
+        doc = self.documents.get(str(document_id))
+        if doc:
+            if status:
+                doc.status = status
+            if auto_label_status:
+                doc.auto_label_status = auto_label_status
+            if auto_label_error:
+                doc.auto_label_error = auto_label_error
+
+    def delete_annotations_for_document(self, document_id, source=None):
+        """Mock delete annotations."""
+        return 0
+
+    def create_annotations_batch(self, annotations):
+        """Mock create annotations."""
+        self.annotations.extend(annotations)
+
+
+@pytest.fixture
+def mock_db():
+    """Create mock admin DB."""
+    return MockAdminDB()
+
+
+@pytest.fixture
+def auto_label_service(monkeypatch):
+    """Create auto-label service with mocked image processing."""
+    service = AutoLabelService()
+    # Mock the OCR engine to avoid dependencies
+    service._ocr_engine = Mock()
+    service._ocr_engine.extract_from_image = Mock(return_value=[])
+
+    # Mock the image processing methods to avoid file I/O errors
+    def mock_process_image(self, document_id, image_path, field_values, db, page_number=1):
+        return 0  # No annotations created (mocked)
+
+    monkeypatch.setattr(AutoLabelService, "_process_image", mock_process_image)
+
+    return service
+
+
+class TestAutoLabelWithLocks:
+    """Tests for auto-label service with lock integration."""
+
+    def test_auto_label_unlocked_document_succeeds(self, auto_label_service, mock_db, tmp_path):
+        """Test auto-labeling succeeds on unlocked document."""
+        # Create test document (unlocked)
+        document_id = str(uuid4())
+        mock_db.documents[document_id] = MockDocument(
+            document_id=document_id,
+            annotation_lock_until=None,
+        )
+
+        # Create dummy file
+        test_file = tmp_path / "test.png"
+        test_file.write_text("dummy")
+
+        # Attempt auto-label
+        result = auto_label_service.auto_label_document(
+            document_id=document_id,
+            file_path=str(test_file),
+            field_values={"invoice_number": "INV-001"},
+            db=mock_db,
+        )
+
+        # Should succeed
+        assert result["status"] == "completed"
+        # Verify status was updated to running and then completed
+        assert len(mock_db.status_updates) >= 2
+        assert mock_db.status_updates[0]["auto_label_status"] == "running"
+
+    def test_auto_label_locked_document_fails(self, auto_label_service, mock_db, tmp_path):
+        """Test auto-labeling fails on locked document."""
+        # Create test document (locked for 1 hour)
+        document_id = str(uuid4())
+        lock_until = datetime.now(timezone.utc) + timedelta(hours=1)
+        mock_db.documents[document_id] = MockDocument(
+            document_id=document_id,
+            annotation_lock_until=lock_until,
+        )
+
+        # Create dummy file
+        test_file = tmp_path / "test.png"
+        test_file.write_text("dummy")
+
+        # Attempt auto-label (should fail)
+        result = auto_label_service.auto_label_document(
+            document_id=document_id,
+            file_path=str(test_file),
+            field_values={"invoice_number": "INV-001"},
+            db=mock_db,
+        )
+
+        # Should fail
+        assert result["status"] == "failed"
+        assert "locked for annotation" in result["error"]
+        assert result["annotations_created"] == 0
+
+        # Verify status was updated to failed
+        assert any(
+            update["auto_label_status"] == "failed"
+            for update in mock_db.status_updates
+        )
+
+    def test_auto_label_expired_lock_succeeds(self, auto_label_service, mock_db, tmp_path):
+        """Test auto-labeling succeeds when lock has expired."""
+        # Create test document (lock expired 1 hour ago)
+        document_id = str(uuid4())
+        lock_until = datetime.now(timezone.utc) - timedelta(hours=1)
+        mock_db.documents[document_id] = MockDocument(
+            document_id=document_id,
+            annotation_lock_until=lock_until,
+        )
+
+        # Create dummy file
+        test_file = tmp_path / "test.png"
+        test_file.write_text("dummy")
+
+        # Attempt auto-label
+        result = auto_label_service.auto_label_document(
+            document_id=document_id,
+            file_path=str(test_file),
+            field_values={"invoice_number": "INV-001"},
+            db=mock_db,
+        )
+
+        # Should succeed (lock expired)
+        assert result["status"] == "completed"
+
+    def test_auto_label_skip_lock_check(self, auto_label_service, mock_db, tmp_path):
+        """Test auto-labeling with skip_lock_check=True bypasses lock."""
+        # Create test document (locked)
+        document_id = str(uuid4())
+        lock_until = datetime.now(timezone.utc) + timedelta(hours=1)
+        mock_db.documents[document_id] = MockDocument(
+            document_id=document_id,
+            annotation_lock_until=lock_until,
+        )
+
+        # Create dummy file
+        test_file = tmp_path / "test.png"
+        test_file.write_text("dummy")
+
+        # Attempt auto-label with skip_lock_check=True
+        result = auto_label_service.auto_label_document(
+            document_id=document_id,
+            file_path=str(test_file),
+            field_values={"invoice_number": "INV-001"},
+            db=mock_db,
+            skip_lock_check=True,  # Bypass lock check
+        )
+
+        # Should succeed even though document is locked
+        assert result["status"] == "completed"
+
+    def test_auto_label_document_not_found(self, auto_label_service, mock_db, tmp_path):
+        """Test auto-labeling fails when document doesn't exist."""
+        # Create dummy file
+        test_file = tmp_path / "test.png"
+        test_file.write_text("dummy")
+
+        # Attempt auto-label on non-existent document
+        result = auto_label_service.auto_label_document(
+            document_id=str(uuid4()),
+            file_path=str(test_file),
+            field_values={"invoice_number": "INV-001"},
+            db=mock_db,
+        )
+
+        # Should fail
+        assert result["status"] == "failed"
+        assert "not found" in result["error"]
+
+    def test_auto_label_respects_lock_by_default(self, auto_label_service, mock_db, tmp_path):
+        """Test that lock check is enabled by default."""
+        # Create test document (locked)
+        document_id = str(uuid4())
+        lock_until = datetime.now(timezone.utc) + timedelta(minutes=30)
+        mock_db.documents[document_id] = MockDocument(
+            document_id=document_id,
+            annotation_lock_until=lock_until,
+        )
+
+        # Create dummy file
+        test_file = tmp_path / "test.png"
+        test_file.write_text("dummy")
+
+        # Call without explicit skip_lock_check (defaults to False)
+        result = auto_label_service.auto_label_document(
+            document_id=document_id,
+            file_path=str(test_file),
+            field_values={"invoice_number": "INV-001"},
+            db=mock_db,
+            # skip_lock_check not specified, should default to False
+        )
+
+        # Should fail due to lock
+        assert result["status"] == "failed"
+        assert "locked" in result["error"].lower()

tests/web/test_batch_queue.py

@@ -0,0 +1,282 @@
+"""
+Tests for Batch Upload Queue
+"""
+
+import time
+from datetime import datetime
+from threading import Event
+from uuid import uuid4
+
+import pytest
+
+from src.web.workers.batch_queue import BatchTask, BatchTaskQueue
+
+
+class MockBatchService:
+    """Mock batch upload service for testing."""
+
+    def __init__(self):
+        self.processed_tasks = []
+        self.process_delay = 0.1  # Simulate processing time
+        self.should_fail = False
+
+    def process_zip_upload(self, admin_token, zip_filename, zip_content, upload_source):
+        """Mock process_zip_upload method."""
+        if self.should_fail:
+            raise Exception("Simulated processing error")
+
+        time.sleep(self.process_delay)  # Simulate work
+
+        self.processed_tasks.append({
+            "admin_token": admin_token,
+            "zip_filename": zip_filename,
+            "upload_source": upload_source,
+        })
+
+        return {
+            "status": "completed",
+            "successful_files": 1,
+            "failed_files": 0,
+        }
+
+
+class TestBatchTask:
+    """Tests for BatchTask dataclass."""
+
+    def test_batch_task_creation(self):
+        """BatchTask can be created with required fields."""
+        task = BatchTask(
+            batch_id=uuid4(),
+            admin_token="test-token",
+            zip_content=b"test",
+            zip_filename="test.zip",
+            upload_source="ui",
+            auto_label=True,
+            created_at=datetime.utcnow(),
+        )
+
+        assert task.batch_id is not None
+        assert task.admin_token == "test-token"
+        assert task.zip_filename == "test.zip"
+        assert task.upload_source == "ui"
+        assert task.auto_label is True
+
+
+class TestBatchTaskQueue:
+    """Tests for batch task queue functionality."""
+
+    def test_queue_initialization(self):
+        """Queue initializes with correct defaults."""
+        queue = BatchTaskQueue(max_size=10, worker_count=1)
+
+        assert queue.get_queue_depth() == 0
+        assert queue.is_running is False
+        assert queue._worker_count == 1
+
+    def test_start_queue(self):
+        """Queue starts with batch service."""
+        queue = BatchTaskQueue(max_size=10, worker_count=1)
+        service = MockBatchService()
+
+        queue.start(service)
+
+        assert queue.is_running is True
+        assert len(queue._workers) == 1
+
+        queue.stop()
+
+    def test_stop_queue(self):
+        """Queue stops gracefully."""
+        queue = BatchTaskQueue(max_size=10, worker_count=1)
+        service = MockBatchService()
+
+        queue.start(service)
+        assert queue.is_running is True
+
+        queue.stop(timeout=5.0)
+
+        assert queue.is_running is False
+        assert len(queue._workers) == 0
+
+    def test_submit_task_success(self):
+        """Task is submitted to queue successfully."""
+        queue = BatchTaskQueue(max_size=10, worker_count=1)
+
+        task = BatchTask(
+            batch_id=uuid4(),
+            admin_token="test-token",
+            zip_content=b"test",
+            zip_filename="test.zip",
+            upload_source="ui",
+            auto_label=True,
+            created_at=datetime.utcnow(),
+        )
+
+        result = queue.submit(task)
+
+        assert result is True
+        assert queue.get_queue_depth() == 1
+
+    def test_submit_task_queue_full(self):
+        """Returns False when queue is full."""
+        queue = BatchTaskQueue(max_size=2, worker_count=1)
+
+        # Fill the queue
+        for i in range(2):
+            task = BatchTask(
+                batch_id=uuid4(),
+                admin_token="test-token",
+                zip_content=b"test",
+                zip_filename=f"test{i}.zip",
+                upload_source="ui",
+                auto_label=True,
+                created_at=datetime.utcnow(),
+            )
+            assert queue.submit(task) is True
+
+        # Try to add one more (should fail)
+        extra_task = BatchTask(
+            batch_id=uuid4(),
+            admin_token="test-token",
+            zip_content=b"test",
+            zip_filename="extra.zip",
+            upload_source="ui",
+            auto_label=True,
+            created_at=datetime.utcnow(),
+        )
+
+        result = queue.submit(extra_task)
+
+        assert result is False
+        assert queue.get_queue_depth() == 2
+
+    def test_worker_processes_task(self):
+        """Worker thread processes queued tasks."""
+        queue = BatchTaskQueue(max_size=10, worker_count=1)
+        service = MockBatchService()
+
+        queue.start(service)
+
+        task = BatchTask(
+            batch_id=uuid4(),
+            admin_token="test-token",
+            zip_content=b"test",
+            zip_filename="test.zip",
+            upload_source="ui",
+            auto_label=True,
+            created_at=datetime.utcnow(),
+        )
+
+        queue.submit(task)
+
+        # Wait for processing
+        time.sleep(0.5)
+
+        assert len(service.processed_tasks) == 1
+        assert service.processed_tasks[0]["zip_filename"] == "test.zip"
+
+        queue.stop()
+
+    def test_multiple_tasks_processed(self):
+        """Multiple tasks are processed in order."""
+        queue = BatchTaskQueue(max_size=10, worker_count=1)
+        service = MockBatchService()
+
+        queue.start(service)
+
+        # Submit multiple tasks
+        for i in range(3):
+            task = BatchTask(
+                batch_id=uuid4(),
+                admin_token="test-token",
+                zip_content=b"test",
+                zip_filename=f"test{i}.zip",
+                upload_source="ui",
+                auto_label=True,
+                created_at=datetime.utcnow(),
+            )
+            queue.submit(task)
+
+        # Wait for all to process
+        time.sleep(1.0)
+
+        assert len(service.processed_tasks) == 3
+
+        queue.stop()
+
+    def test_get_queue_depth(self):
+        """Returns correct queue depth."""
+        queue = BatchTaskQueue(max_size=10, worker_count=1)
+
+        assert queue.get_queue_depth() == 0
+
+        # Add tasks
+        for i in range(3):
+            task = BatchTask(
+                batch_id=uuid4(),
+                admin_token="test-token",
+                zip_content=b"test",
+                zip_filename=f"test{i}.zip",
+                upload_source="ui",
+                auto_label=True,
+                created_at=datetime.utcnow(),
+            )
+            queue.submit(task)
+
+        assert queue.get_queue_depth() == 3
+
+    def test_is_running_property(self):
+        """is_running reflects queue state."""
+        queue = BatchTaskQueue(max_size=10, worker_count=1)
+        service = MockBatchService()
+
+        assert queue.is_running is False
+
+        queue.start(service)
+        assert queue.is_running is True
+
+        queue.stop()
+        assert queue.is_running is False
+
+    def test_double_start_ignored(self):
+        """Starting queue twice is safely ignored."""
+        queue = BatchTaskQueue(max_size=10, worker_count=1)
+        service = MockBatchService()
+
+        queue.start(service)
+        worker_count_after_first_start = len(queue._workers)
+
+        queue.start(service)  # Second start
+        worker_count_after_second_start = len(queue._workers)
+
+        assert worker_count_after_first_start == worker_count_after_second_start
+
+        queue.stop()
+
+    def test_error_handling_in_worker(self):
+        """Worker handles processing errors gracefully."""
+        queue = BatchTaskQueue(max_size=10, worker_count=1)
+        service = MockBatchService()
+        service.should_fail = True  # Cause errors
+
+        queue.start(service)
+
+        task = BatchTask(
+            batch_id=uuid4(),
+            admin_token="test-token",
+            zip_content=b"test",
+            zip_filename="test.zip",
+            upload_source="ui",
+            auto_label=True,
+            created_at=datetime.utcnow(),
+        )
+
+        queue.submit(task)
+
+        # Wait for processing attempt
+        time.sleep(0.5)
+
+        # Worker should still be running
+        assert queue.is_running is True
+
+        queue.stop()

tests/web/test_batch_upload_routes.py

@@ -0,0 +1,368 @@
+"""
+Tests for Batch Upload Routes
+"""
+
+import io
+import zipfile
+from datetime import datetime
+from uuid import uuid4
+
+import pytest
+from fastapi import FastAPI
+from fastapi.testclient import TestClient
+
+from src.web.api.v1.batch.routes import router
+from src.web.core.auth import validate_admin_token, get_admin_db
+from src.web.workers.batch_queue import init_batch_queue, shutdown_batch_queue
+from src.web.services.batch_upload import BatchUploadService
+
+
+class MockAdminDB:
+    """Mock AdminDB for testing."""
+
+    def __init__(self):
+        self.batches = {}
+        self.batch_files = {}
+
+    def create_batch_upload(self, admin_token, filename, file_size, upload_source):
+        batch_id = uuid4()
+        batch = type('BatchUpload', (), {
+            'batch_id': batch_id,
+            'admin_token': admin_token,
+            'filename': filename,
+            'file_size': file_size,
+            'upload_source': upload_source,
+            'status': 'processing',
+            'total_files': 0,
+            'processed_files': 0,
+            'successful_files': 0,
+            'failed_files': 0,
+            'csv_filename': None,
+            'csv_row_count': None,
+            'error_message': None,
+            'created_at': datetime.utcnow(),
+            'completed_at': None,
+        })()
+        self.batches[batch_id] = batch
+        return batch
+
+    def update_batch_upload(self, batch_id, **kwargs):
+        if batch_id in self.batches:
+            batch = self.batches[batch_id]
+            for key, value in kwargs.items():
+                setattr(batch, key, value)
+
+    def create_batch_upload_file(self, batch_id, filename, **kwargs):
+        file_id = uuid4()
+        defaults = {
+            'file_id': file_id,
+            'batch_id': batch_id,
+            'filename': filename,
+            'status': 'pending',
+            'error_message': None,
+            'annotation_count': 0,
+            'csv_row_data': None,
+        }
+        defaults.update(kwargs)
+        file_record = type('BatchUploadFile', (), defaults)()
+        if batch_id not in self.batch_files:
+            self.batch_files[batch_id] = []
+        self.batch_files[batch_id].append(file_record)
+        return file_record
+
+    def update_batch_upload_file(self, file_id, **kwargs):
+        for files in self.batch_files.values():
+            for file_record in files:
+                if file_record.file_id == file_id:
+                    for key, value in kwargs.items():
+                        setattr(file_record, key, value)
+                    return
+
+    def get_batch_upload(self, batch_id):
+        return self.batches.get(batch_id, type('BatchUpload', (), {
+            'batch_id': batch_id,
+            'admin_token': 'test-token',
+            'filename': 'test.zip',
+            'status': 'completed',
+            'total_files': 2,
+            'processed_files': 2,
+            'successful_files': 2,
+            'failed_files': 0,
+            'csv_filename': None,
+            'csv_row_count': None,
+            'error_message': None,
+            'created_at': datetime.utcnow(),
+            'completed_at': datetime.utcnow(),
+        })())
+
+    def get_batch_upload_files(self, batch_id):
+        return self.batch_files.get(batch_id, [])
+
+    def get_batch_uploads_by_token(self, admin_token, limit=50, offset=0):
+        """Get batches filtered by admin token with pagination."""
+        token_batches = [b for b in self.batches.values() if b.admin_token == admin_token]
+        total = len(token_batches)
+        return token_batches[offset:offset+limit], total
+
+
+@pytest.fixture(scope="class")
+def app():
+    """Create test FastAPI app with mocked dependencies."""
+    app = FastAPI()
+
+    # Create mock admin DB
+    mock_admin_db = MockAdminDB()
+
+    # Override dependencies
+    app.dependency_overrides[validate_admin_token] = lambda: "test-token"
+    app.dependency_overrides[get_admin_db] = lambda: mock_admin_db
+
+    # Initialize batch queue with mock service
+    batch_service = BatchUploadService(mock_admin_db)
+    init_batch_queue(batch_service)
+
+    app.include_router(router)
+
+    yield app
+
+    # Cleanup: shutdown batch queue after all tests in class
+    shutdown_batch_queue()
+
+
+@pytest.fixture
+def client(app):
+    """Create test client."""
+    return TestClient(app)
+
+
+def create_test_zip(files):
+    """Create a test ZIP file."""
+    zip_buffer = io.BytesIO()
+    with zipfile.ZipFile(zip_buffer, 'w', zipfile.ZIP_DEFLATED) as zip_file:
+        for filename, content in files.items():
+            zip_file.writestr(filename, content)
+    zip_buffer.seek(0)
+    return zip_buffer
+
+
+class TestBatchUploadRoutes:
+    """Tests for batch upload API routes."""
+
+    def test_upload_batch_success(self, client):
+        """Test successful batch upload (defaults to async mode)."""
+        files = {
+            "INV001.pdf": b"%PDF-1.4 test content",
+            "INV002.pdf": b"%PDF-1.4 test content 2",
+        }
+        zip_file = create_test_zip(files)
+
+        response = client.post(
+            "/api/v1/admin/batch/upload",
+            files={"file": ("test.zip", zip_file, "application/zip")},
+            data={"upload_source": "ui"},
+        )
+
+        # Async mode is default, should return 202
+        assert response.status_code == 202
+        result = response.json()
+        assert "batch_id" in result
+        assert result["status"] == "accepted"
+
+    def test_upload_batch_non_zip_file(self, client):
+        """Test uploading non-ZIP file."""
+        response = client.post(
+            "/api/v1/admin/batch/upload",
+            files={"file": ("test.pdf", io.BytesIO(b"test"), "application/pdf")},
+            data={"upload_source": "ui"},
+        )
+
+        assert response.status_code == 400
+        assert "Only ZIP files" in response.json()["detail"]
+
+    def test_upload_batch_with_csv(self, client):
+        """Test batch upload with CSV (defaults to async)."""
+        csv_content = """DocumentId,InvoiceNumber,Amount
+INV001,F2024-001,1500.00
+INV002,F2024-002,2500.00
+"""
+        files = {
+            "INV001.pdf": b"%PDF-1.4 test",
+            "INV002.pdf": b"%PDF-1.4 test 2",
+            "metadata.csv": csv_content.encode('utf-8'),
+        }
+        zip_file = create_test_zip(files)
+
+        response = client.post(
+            "/api/v1/admin/batch/upload",
+            files={"file": ("batch.zip", zip_file, "application/zip")},
+            data={"upload_source": "api"},
+        )
+
+        # Async mode is default, should return 202
+        assert response.status_code == 202
+        result = response.json()
+        assert "batch_id" in result
+        assert result["status"] == "accepted"
+
+    def test_get_batch_status(self, client):
+        """Test getting batch status."""
+        batch_id = str(uuid4())
+        response = client.get(f"/api/v1/admin/batch/status/{batch_id}")
+
+        assert response.status_code == 200
+        result = response.json()
+        assert result["batch_id"] == batch_id
+        assert "status" in result
+        assert "total_files" in result
+
+    def test_list_batch_uploads(self, client):
+        """Test listing batch uploads."""
+        response = client.get("/api/v1/admin/batch/list")
+
+        assert response.status_code == 200
+        result = response.json()
+        assert "batches" in result
+        assert "total" in result
+        assert "limit" in result
+        assert "offset" in result
+
+    def test_upload_batch_async_mode_default(self, client):
+        """Test async mode is default (async_mode=True)."""
+        files = {
+            "INV001.pdf": b"%PDF-1.4 test content",
+        }
+        zip_file = create_test_zip(files)
+
+        response = client.post(
+            "/api/v1/admin/batch/upload",
+            files={"file": ("test.zip", zip_file, "application/zip")},
+            data={"upload_source": "ui"},
+        )
+
+        # Async mode should return 202 Accepted
+        assert response.status_code == 202
+        result = response.json()
+        assert result["status"] == "accepted"
+        assert "batch_id" in result
+        assert "status_url" in result
+        assert "queue_depth" in result
+        assert result["message"] == "Batch upload queued for processing"
+
+    def test_upload_batch_async_mode_explicit(self, client):
+        """Test explicit async mode (async_mode=True)."""
+        files = {
+            "INV001.pdf": b"%PDF-1.4 test content",
+        }
+        zip_file = create_test_zip(files)
+
+        response = client.post(
+            "/api/v1/admin/batch/upload",
+            files={"file": ("test.zip", zip_file, "application/zip")},
+            data={"upload_source": "ui", "async_mode": "true"},
+        )
+
+        assert response.status_code == 202
+        result = response.json()
+        assert result["status"] == "accepted"
+        assert "batch_id" in result
+        assert "status_url" in result
+
+    def test_upload_batch_sync_mode(self, client):
+        """Test sync mode (async_mode=False)."""
+        files = {
+            "INV001.pdf": b"%PDF-1.4 test content",
+        }
+        zip_file = create_test_zip(files)
+
+        response = client.post(
+            "/api/v1/admin/batch/upload",
+            files={"file": ("test.zip", zip_file, "application/zip")},
+            data={"upload_source": "ui", "async_mode": "false"},
+        )
+
+        # Sync mode should return 200 OK with full results
+        assert response.status_code == 200
+        result = response.json()
+        assert "batch_id" in result
+        assert result["status"] in ["completed", "partial", "failed"]
+        assert "successful_files" in result
+
+    def test_upload_batch_async_with_auto_label(self, client):
+        """Test async mode with auto_label flag."""
+        files = {
+            "INV001.pdf": b"%PDF-1.4 test content",
+        }
+        zip_file = create_test_zip(files)
+
+        response = client.post(
+            "/api/v1/admin/batch/upload",
+            files={"file": ("test.zip", zip_file, "application/zip")},
+            data={
+                "upload_source": "ui",
+                "async_mode": "true",
+                "auto_label": "true",
+            },
+        )
+
+        assert response.status_code == 202
+        result = response.json()
+        assert result["status"] == "accepted"
+        assert "batch_id" in result
+
+    def test_upload_batch_async_without_auto_label(self, client):
+        """Test async mode with auto_label disabled."""
+        files = {
+            "INV001.pdf": b"%PDF-1.4 test content",
+        }
+        zip_file = create_test_zip(files)
+
+        response = client.post(
+            "/api/v1/admin/batch/upload",
+            files={"file": ("test.zip", zip_file, "application/zip")},
+            data={
+                "upload_source": "ui",
+                "async_mode": "true",
+                "auto_label": "false",
+            },
+        )
+
+        assert response.status_code == 202
+        result = response.json()
+        assert result["status"] == "accepted"
+
+    def test_upload_batch_queue_full(self, client):
+        """Test handling queue full scenario."""
+        # This test would require mocking the queue to return False on submit
+        # For now, we verify the endpoint accepts the request
+        files = {
+            "INV001.pdf": b"%PDF-1.4 test content",
+        }
+        zip_file = create_test_zip(files)
+
+        response = client.post(
+            "/api/v1/admin/batch/upload",
+            files={"file": ("test.zip", zip_file, "application/zip")},
+            data={"upload_source": "ui", "async_mode": "true"},
+        )
+
+        # Should either accept (202) or reject if queue full (503)
+        assert response.status_code in [202, 503]
+
+    def test_async_status_url_format(self, client):
+        """Test async response contains correctly formatted status URL."""
+        files = {
+            "INV001.pdf": b"%PDF-1.4 test content",
+        }
+        zip_file = create_test_zip(files)
+
+        response = client.post(
+            "/api/v1/admin/batch/upload",
+            files={"file": ("test.zip", zip_file, "application/zip")},
+            data={"async_mode": "true"},
+        )
+
+        assert response.status_code == 202
+        result = response.json()
+        batch_id = result["batch_id"]
+        expected_url = f"/api/v1/admin/batch/status/{batch_id}"
+        assert result["status_url"] == expected_url

tests/web/test_batch_upload_service.py

@@ -0,0 +1,221 @@
+"""
+Tests for Batch Upload Service
+"""
+
+import io
+import zipfile
+from pathlib import Path
+from uuid import uuid4
+
+import pytest
+
+from src.data.admin_db import AdminDB
+from src.web.services.batch_upload import BatchUploadService
+
+
+@pytest.fixture
+def admin_db():
+    """Mock admin database for testing."""
+    class MockAdminDB:
+        def __init__(self):
+            self.batches = {}
+            self.batch_files = {}
+
+        def create_batch_upload(self, admin_token, filename, file_size, upload_source):
+            batch_id = uuid4()
+            batch = type('BatchUpload', (), {
+                'batch_id': batch_id,
+                'admin_token': admin_token,
+                'filename': filename,
+                'file_size': file_size,
+                'upload_source': upload_source,
+                'status': 'processing',
+                'total_files': 0,
+                'processed_files': 0,
+                'successful_files': 0,
+                'failed_files': 0,
+                'csv_filename': None,
+                'csv_row_count': None,
+                'error_message': None,
+                'created_at': None,
+                'completed_at': None,
+            })()
+            self.batches[batch_id] = batch
+            return batch
+
+        def update_batch_upload(self, batch_id, **kwargs):
+            if batch_id in self.batches:
+                batch = self.batches[batch_id]
+                for key, value in kwargs.items():
+                    setattr(batch, key, value)
+
+        def create_batch_upload_file(self, batch_id, filename, **kwargs):
+            file_id = uuid4()
+            # Set defaults for attributes
+            defaults = {
+                'file_id': file_id,
+                'batch_id': batch_id,
+                'filename': filename,
+                'status': 'pending',
+                'error_message': None,
+                'annotation_count': 0,
+                'csv_row_data': None,
+            }
+            defaults.update(kwargs)
+            file_record = type('BatchUploadFile', (), defaults)()
+            if batch_id not in self.batch_files:
+                self.batch_files[batch_id] = []
+            self.batch_files[batch_id].append(file_record)
+            return file_record
+
+        def update_batch_upload_file(self, file_id, **kwargs):
+            for files in self.batch_files.values():
+                for file_record in files:
+                    if file_record.file_id == file_id:
+                        for key, value in kwargs.items():
+                            setattr(file_record, key, value)
+                        return
+
+        def get_batch_upload(self, batch_id):
+            return self.batches.get(batch_id)
+
+        def get_batch_upload_files(self, batch_id):
+            return self.batch_files.get(batch_id, [])
+
+    return MockAdminDB()
+
+
+@pytest.fixture
+def batch_service(admin_db):
+    """Batch upload service instance."""
+    return BatchUploadService(admin_db)
+
+
+def create_test_zip(files):
+    """Create a test ZIP file with given files.
+
+    Args:
+        files: Dictionary mapping filenames to content bytes
+
+    Returns:
+        ZIP file content as bytes
+    """
+    zip_buffer = io.BytesIO()
+    with zipfile.ZipFile(zip_buffer, 'w', zipfile.ZIP_DEFLATED) as zip_file:
+        for filename, content in files.items():
+            zip_file.writestr(filename, content)
+    return zip_buffer.getvalue()
+
+
+class TestBatchUploadService:
+    """Tests for BatchUploadService."""
+
+    def test_process_empty_zip(self, batch_service):
+        """Test processing an empty ZIP file."""
+        zip_content = create_test_zip({})
+        result = batch_service.process_zip_upload(
+            admin_token="test-token",
+            zip_filename="empty.zip",
+            zip_content=zip_content,
+        )
+
+        assert result["status"] == "failed"
+        assert "No PDF files" in result.get("error", "")
+
+    def test_process_zip_with_pdfs_only(self, batch_service):
+        """Test processing ZIP with PDFs but no CSV."""
+        files = {
+            "INV001.pdf": b"%PDF-1.4 test content",
+            "INV002.pdf": b"%PDF-1.4 test content 2",
+        }
+        zip_content = create_test_zip(files)
+
+        result = batch_service.process_zip_upload(
+            admin_token="test-token",
+            zip_filename="invoices.zip",
+            zip_content=zip_content,
+        )
+
+        assert result["status"] == "completed"
+        assert result["total_files"] == 2
+        assert result["successful_files"] == 2
+        assert result["failed_files"] == 0
+
+    def test_process_zip_with_csv(self, batch_service):
+        """Test processing ZIP with PDFs and CSV."""
+        csv_content = """DocumentId,InvoiceNumber,Amount,OCR
+INV001,F2024-001,1500.00,7350012345678
+INV002,F2024-002,2500.00,7350087654321
+"""
+        files = {
+            "INV001.pdf": b"%PDF-1.4 test content",
+            "INV002.pdf": b"%PDF-1.4 test content 2",
+            "metadata.csv": csv_content.encode('utf-8'),
+        }
+        zip_content = create_test_zip(files)
+
+        result = batch_service.process_zip_upload(
+            admin_token="test-token",
+            zip_filename="invoices.zip",
+            zip_content=zip_content,
+        )
+
+        assert result["status"] == "completed"
+        assert result["total_files"] == 2
+        assert result["csv_filename"] == "metadata.csv"
+        assert result["csv_row_count"] == 2
+
+    def test_process_invalid_zip(self, batch_service):
+        """Test processing invalid ZIP file."""
+        result = batch_service.process_zip_upload(
+            admin_token="test-token",
+            zip_filename="invalid.zip",
+            zip_content=b"not a zip file",
+        )
+
+        assert result["status"] == "failed"
+        assert "Invalid ZIP file" in result.get("error", "")
+
+    def test_csv_parsing(self, batch_service):
+        """Test CSV field parsing."""
+        csv_content = """DocumentId,InvoiceNumber,InvoiceDate,Amount,OCR,Bankgiro,customer_number
+INV001,F2024-001,2024-01-15,1500.00,7350012345678,123-4567,C123
+INV002,F2024-002,2024-01-16,2500.00,7350087654321,123-4567,C124
+"""
+        zip_file_content = create_test_zip({"metadata.csv": csv_content.encode('utf-8')})
+
+        with zipfile.ZipFile(io.BytesIO(zip_file_content)) as zip_file:
+            csv_file_info = [f for f in zip_file.filelist if f.filename.endswith('.csv')][0]
+            csv_data = batch_service._parse_csv_file(zip_file, csv_file_info)
+
+        assert len(csv_data) == 2
+        assert "INV001" in csv_data
+        assert csv_data["INV001"]["InvoiceNumber"] == "F2024-001"
+        assert csv_data["INV001"]["Amount"] == "1500.00"
+        assert csv_data["INV001"]["customer_number"] == "C123"
+
+    def test_get_batch_status(self, batch_service, admin_db):
+        """Test getting batch upload status."""
+        # Create a batch
+        zip_content = create_test_zip({"INV001.pdf": b"%PDF-1.4 test"})
+        result = batch_service.process_zip_upload(
+            admin_token="test-token",
+            zip_filename="test.zip",
+            zip_content=zip_content,
+        )
+
+        batch_id = result["batch_id"]
+
+        # Get status
+        status = batch_service.get_batch_status(batch_id)
+
+        assert status["batch_id"] == batch_id
+        assert status["filename"] == "test.zip"
+        assert status["status"] == "completed"
+        assert status["total_files"] == 1
+        assert len(status["files"]) == 1
+
+    def test_get_batch_status_not_found(self, batch_service):
+        """Test getting status for non-existent batch."""
+        status = batch_service.get_batch_status(str(uuid4()))
+        assert "error" in status

tests/web/test_inference_api.py

@@ -0,0 +1,298 @@
+"""
+Integration tests for inference API endpoints.
+
+Tests the /api/v1/infer endpoint to ensure it works end-to-end.
+"""
+
+import pytest
+from pathlib import Path
+from unittest.mock import Mock, patch
+from fastapi.testclient import TestClient
+from PIL import Image
+import io
+
+from src.web.app import create_app
+from src.web.config import ModelConfig, StorageConfig, AppConfig
+
+
+@pytest.fixture
+def test_app(tmp_path):
+    """Create test FastAPI application."""
+    # Setup test directories
+    upload_dir = tmp_path / "uploads"
+    result_dir = tmp_path / "results"
+    upload_dir.mkdir()
+    result_dir.mkdir()
+
+    # Create test config
+    app_config = AppConfig(
+        model=ModelConfig(
+            model_path=Path("runs/train/invoice_fields/weights/best.pt"),
+            confidence_threshold=0.5,
+            use_gpu=False,
+            dpi=150,
+        ),
+        storage=StorageConfig(
+            upload_dir=upload_dir,
+            result_dir=result_dir,
+            allowed_extensions={".pdf", ".png", ".jpg", ".jpeg"},
+            max_file_size_mb=50,
+        ),
+    )
+
+    # Create app
+    app = create_app(app_config)
+    return app
+
+
+@pytest.fixture
+def client(test_app):
+    """Create test client."""
+    return TestClient(test_app)
+
+
+@pytest.fixture
+def sample_png_bytes():
+    """Create sample PNG image bytes."""
+    img = Image.new('RGB', (800, 1200), color='white')
+    img_bytes = io.BytesIO()
+    img.save(img_bytes, format='PNG')
+    img_bytes.seek(0)
+    return img_bytes
+
+
+class TestHealthEndpoint:
+    """Test /api/v1/health endpoint."""
+
+    def test_health_check_returns_200(self, client):
+        """Test health check returns 200 OK."""
+        response = client.get("/api/v1/health")
+        assert response.status_code == 200
+
+    def test_health_check_response_structure(self, client):
+        """Test health check response has correct structure."""
+        response = client.get("/api/v1/health")
+        data = response.json()
+
+        assert "status" in data
+        assert "model_loaded" in data
+        assert "gpu_available" in data
+        assert "version" in data
+
+        assert data["status"] == "healthy"
+        assert isinstance(data["model_loaded"], bool)
+        assert isinstance(data["gpu_available"], bool)
+
+
+class TestInferEndpoint:
+    """Test /api/v1/infer endpoint."""
+
+    @patch('src.inference.pipeline.InferencePipeline')
+    @patch('src.inference.yolo_detector.YOLODetector')
+    def test_infer_accepts_png_file(
+        self,
+        mock_yolo_detector,
+        mock_pipeline,
+        client,
+        sample_png_bytes,
+    ):
+        """Test that /infer endpoint accepts PNG files."""
+        # Setup mocks
+        mock_detector_instance = Mock()
+        mock_pipeline_instance = Mock()
+        mock_yolo_detector.return_value = mock_detector_instance
+        mock_pipeline.return_value = mock_pipeline_instance
+
+        # Mock pipeline result
+        mock_result = Mock()
+        mock_result.fields = {"InvoiceNumber": "12345"}
+        mock_result.confidence = {"InvoiceNumber": 0.95}
+        mock_result.success = True
+        mock_result.errors = []
+        mock_result.raw_detections = []
+        mock_result.document_id = "test123"
+        mock_result.document_type = "invoice"
+        mock_result.processing_time_ms = 100.0
+        mock_result.visualization_path = None
+        mock_result.detections = []
+        mock_pipeline_instance.process_image.return_value = mock_result
+
+        # Make request
+        response = client.post(
+            "/api/v1/infer",
+            files={"file": ("test.png", sample_png_bytes, "image/png")},
+        )
+
+        # Verify response
+        assert response.status_code == 200
+        data = response.json()
+
+        assert data["status"] == "success"
+        assert "result" in data
+        assert data["result"]["fields"]["InvoiceNumber"] == "12345"
+        assert data["result"]["confidence"]["InvoiceNumber"] == 0.95
+
+    def test_infer_rejects_invalid_file_type(self, client):
+        """Test that /infer rejects unsupported file types."""
+        invalid_file = io.BytesIO(b"fake txt content")
+
+        response = client.post(
+            "/api/v1/infer",
+            files={"file": ("test.txt", invalid_file, "text/plain")},
+        )
+
+        assert response.status_code == 400
+        assert "Unsupported file type" in response.json()["detail"]
+
+    def test_infer_requires_file(self, client):
+        """Test that /infer requires a file parameter."""
+        response = client.post("/api/v1/infer")
+
+        assert response.status_code == 422  # Unprocessable Entity
+
+    @patch('src.inference.pipeline.InferencePipeline')
+    @patch('src.inference.yolo_detector.YOLODetector')
+    def test_infer_returns_cross_validation_if_available(
+        self,
+        mock_yolo_detector,
+        mock_pipeline,
+        client,
+        sample_png_bytes,
+    ):
+        """Test that cross-validation results are included if available."""
+        # Setup mocks
+        mock_detector_instance = Mock()
+        mock_pipeline_instance = Mock()
+        mock_yolo_detector.return_value = mock_detector_instance
+        mock_pipeline.return_value = mock_pipeline_instance
+
+        # Mock pipeline result with cross-validation
+        mock_result = Mock()
+        mock_result.fields = {
+            "InvoiceNumber": "12345",
+            "OCR": "1234567",
+            "Amount": "100.00",
+        }
+        mock_result.confidence = {
+            "InvoiceNumber": 0.95,
+            "OCR": 0.90,
+            "Amount": 0.88,
+        }
+        mock_result.success = True
+        mock_result.errors = []
+        mock_result.raw_detections = []
+        mock_result.document_id = "test123"
+        mock_result.document_type = "invoice"
+        mock_result.processing_time_ms = 100.0
+        mock_result.visualization_path = None
+        mock_result.detections = []
+
+        # Add cross-validation result
+        mock_cv = Mock()
+        mock_cv.is_valid = True
+        mock_cv.payment_line_ocr = "1234567"
+        mock_cv.ocr_match = True
+        mock_result.cross_validation = mock_cv
+
+        mock_pipeline_instance.process_image.return_value = mock_result
+
+        # Make request
+        response = client.post(
+            "/api/v1/infer",
+            files={"file": ("test.png", sample_png_bytes, "image/png")},
+        )
+
+        # Verify response includes cross-validation
+        assert response.status_code == 200
+        data = response.json()
+
+        # Note: cross_validation is not currently in the response schema
+        # This test documents that it should be added
+
+
+    @patch('src.inference.pipeline.InferencePipeline')
+    @patch('src.inference.yolo_detector.YOLODetector')
+    def test_infer_handles_processing_errors_gracefully(
+        self,
+        mock_yolo_detector,
+        mock_pipeline,
+        client,
+        sample_png_bytes,
+    ):
+        """Test that processing errors are handled gracefully."""
+        # Setup mocks
+        mock_detector_instance = Mock()
+        mock_pipeline_instance = Mock()
+        mock_yolo_detector.return_value = mock_detector_instance
+        mock_pipeline.return_value = mock_pipeline_instance
+
+        # Make pipeline raise an error
+        mock_pipeline_instance.process_image.side_effect = Exception("Model inference failed")
+
+        # Make request
+        response = client.post(
+            "/api/v1/infer",
+            files={"file": ("test.png", sample_png_bytes, "image/png")},
+        )
+
+        # Verify error handling - service catches exceptions and returns partial results
+        assert response.status_code == 200
+        data = response.json()
+        assert data["status"] == "partial"
+        assert data["result"]["success"] is False
+        assert len(data["result"]["errors"]) > 0
+        assert "Model inference failed" in data["result"]["errors"][0]
+
+
+class TestResultsEndpoint:
+    """Test /api/v1/results/{filename} endpoint."""
+
+    def test_get_result_image_returns_404_if_not_found(self, client):
+        """Test that getting non-existent result returns 404."""
+        response = client.get("/api/v1/results/nonexistent.png")
+        assert response.status_code == 404
+
+    def test_get_result_image_returns_file_if_exists(self, client, test_app, tmp_path):
+        """Test that existing result file is returned."""
+        # Get storage config from app
+        storage_config = test_app.extra.get("storage_config")
+        if not storage_config:
+            pytest.skip("Storage config not available in test app")
+
+        # Create a test result file
+        result_file = storage_config.result_dir / "test_result.png"
+        img = Image.new('RGB', (100, 100), color='red')
+        img.save(result_file)
+
+        # Request the file
+        response = client.get("/api/v1/results/test_result.png")
+
+        assert response.status_code == 200
+        assert response.headers["content-type"] == "image/png"
+
+
+class TestInferenceServiceImports:
+    """Critical test to catch import errors."""
+
+    def test_inference_service_can_import_modules(self):
+        """
+        Test that InferenceService can import its dependencies.
+
+        This test will fail if there are ImportError issues like:
+        - from ..inference.pipeline (wrong relative import)
+        - from src.web.inference (non-existent module)
+
+        It ensures the imports are correct before runtime.
+        """
+        from src.web.services.inference import InferenceService
+
+        # Import the modules that InferenceService tries to import
+        from src.inference.pipeline import InferencePipeline
+        from src.inference.yolo_detector import YOLODetector
+        from src.pdf.renderer import render_pdf_to_images
+
+        # If we got here, all imports work correctly
+        assert InferencePipeline is not None
+        assert YOLODetector is not None
+        assert render_pdf_to_images is not None
+        assert InferenceService is not None

tests/web/test_inference_service.py

@@ -0,0 +1,297 @@
+"""
+Integration tests for inference service.
+
+Tests the full initialization and processing flow to catch import errors.
+"""
+
+import pytest
+from pathlib import Path
+from unittest.mock import Mock, patch
+from PIL import Image
+import io
+
+from src.web.services.inference import InferenceService
+from src.web.config import ModelConfig, StorageConfig
+
+
+@pytest.fixture
+def model_config(tmp_path):
+    """Create model configuration for testing."""
+    return ModelConfig(
+        model_path=Path("runs/train/invoice_fields/weights/best.pt"),
+        confidence_threshold=0.5,
+        use_gpu=False,  # Use CPU for tests
+        dpi=150,
+    )
+
+
+@pytest.fixture
+def storage_config(tmp_path):
+    """Create storage configuration for testing."""
+    upload_dir = tmp_path / "uploads"
+    result_dir = tmp_path / "results"
+    upload_dir.mkdir()
+    result_dir.mkdir()
+
+    return StorageConfig(
+        upload_dir=upload_dir,
+        result_dir=result_dir,
+        allowed_extensions={".pdf", ".png", ".jpg", ".jpeg"},
+        max_file_size_mb=50,
+    )
+
+
+@pytest.fixture
+def sample_image(tmp_path):
+    """Create a sample test image."""
+    image_path = tmp_path / "test_invoice.png"
+    img = Image.new('RGB', (800, 1200), color='white')
+    img.save(image_path)
+    return image_path
+
+
+@pytest.fixture
+def inference_service(model_config, storage_config):
+    """Create inference service instance."""
+    return InferenceService(
+        model_config=model_config,
+        storage_config=storage_config,
+    )
+
+
+class TestInferenceServiceInitialization:
+    """Test inference service initialization to catch import errors."""
+
+    def test_service_creation(self, inference_service):
+        """Test that service can be created without errors."""
+        assert inference_service is not None
+        assert not inference_service.is_initialized
+
+    def test_gpu_available_check(self, inference_service):
+        """Test GPU availability check (should not crash)."""
+        gpu_available = inference_service.gpu_available
+        assert isinstance(gpu_available, bool)
+
+    @patch('src.inference.pipeline.InferencePipeline')
+    @patch('src.inference.yolo_detector.YOLODetector')
+    def test_initialize_imports_correctly(
+        self,
+        mock_yolo_detector,
+        mock_pipeline,
+        inference_service,
+    ):
+        """
+        Test that initialize() imports modules correctly.
+
+        This test ensures that the import statements in initialize()
+        use correct paths and don't fail with ImportError.
+        """
+        # Mock the constructors to avoid actually loading models
+        mock_detector_instance = Mock()
+        mock_pipeline_instance = Mock()
+        mock_yolo_detector.return_value = mock_detector_instance
+        mock_pipeline.return_value = mock_pipeline_instance
+
+        # Initialize should not raise ImportError
+        inference_service.initialize()
+
+        # Verify initialization succeeded
+        assert inference_service.is_initialized
+
+        # Verify imports were called with correct parameters
+        mock_yolo_detector.assert_called_once()
+        mock_pipeline.assert_called_once()
+
+    @patch('src.inference.pipeline.InferencePipeline')
+    @patch('src.inference.yolo_detector.YOLODetector')
+    def test_initialize_sets_up_pipeline(
+        self,
+        mock_yolo_detector,
+        mock_pipeline,
+        inference_service,
+        model_config,
+    ):
+        """Test that initialize sets up pipeline with correct config."""
+        mock_detector_instance = Mock()
+        mock_pipeline_instance = Mock()
+        mock_yolo_detector.return_value = mock_detector_instance
+        mock_pipeline.return_value = mock_pipeline_instance
+
+        inference_service.initialize()
+
+        # Check YOLO detector was initialized correctly
+        mock_yolo_detector.assert_called_once_with(
+            str(model_config.model_path),
+            confidence_threshold=model_config.confidence_threshold,
+            device="cpu",  # use_gpu=False in fixture
+        )
+
+        # Check pipeline was initialized correctly
+        mock_pipeline.assert_called_once_with(
+            model_path=str(model_config.model_path),
+            confidence_threshold=model_config.confidence_threshold,
+            use_gpu=False,
+            dpi=150,
+            enable_fallback=True,
+        )
+
+    @patch('src.inference.pipeline.InferencePipeline')
+    @patch('src.inference.yolo_detector.YOLODetector')
+    def test_initialize_idempotent(
+        self,
+        mock_yolo_detector,
+        mock_pipeline,
+        inference_service,
+    ):
+        """Test that calling initialize() multiple times is safe."""
+        mock_detector_instance = Mock()
+        mock_pipeline_instance = Mock()
+        mock_yolo_detector.return_value = mock_detector_instance
+        mock_pipeline.return_value = mock_pipeline_instance
+
+        # Call initialize twice
+        inference_service.initialize()
+        inference_service.initialize()
+
+        # Should only be called once due to is_initialized check
+        assert mock_yolo_detector.call_count == 1
+        assert mock_pipeline.call_count == 1
+
+
+class TestInferenceServiceProcessing:
+    """Test inference processing methods."""
+
+    @patch('src.inference.pipeline.InferencePipeline')
+    @patch('src.inference.yolo_detector.YOLODetector')
+    @patch('ultralytics.YOLO')
+    def test_process_image_basic_flow(
+        self,
+        mock_yolo_class,
+        mock_yolo_detector,
+        mock_pipeline,
+        inference_service,
+        sample_image,
+    ):
+        """Test basic image processing flow."""
+        # Setup mocks
+        mock_detector_instance = Mock()
+        mock_pipeline_instance = Mock()
+        mock_yolo_detector.return_value = mock_detector_instance
+        mock_pipeline.return_value = mock_pipeline_instance
+
+        # Mock pipeline result
+        mock_result = Mock()
+        mock_result.fields = {"InvoiceNumber": "12345"}
+        mock_result.confidence = {"InvoiceNumber": 0.95}
+        mock_result.success = True
+        mock_result.errors = []
+        mock_result.raw_detections = []
+        mock_pipeline_instance.process_image.return_value = mock_result
+
+        # Process image
+        result = inference_service.process_image(sample_image)
+
+        # Verify result
+        assert result.success
+        assert result.fields == {"InvoiceNumber": "12345"}
+        assert result.confidence == {"InvoiceNumber": 0.95}
+        assert result.processing_time_ms > 0
+
+    @patch('src.inference.pipeline.InferencePipeline')
+    @patch('src.inference.yolo_detector.YOLODetector')
+    def test_process_image_handles_errors(
+        self,
+        mock_yolo_detector,
+        mock_pipeline,
+        inference_service,
+        sample_image,
+    ):
+        """Test that processing errors are handled gracefully."""
+        # Setup mocks
+        mock_detector_instance = Mock()
+        mock_pipeline_instance = Mock()
+        mock_yolo_detector.return_value = mock_detector_instance
+        mock_pipeline.return_value = mock_pipeline_instance
+
+        # Make pipeline raise an error
+        mock_pipeline_instance.process_image.side_effect = Exception("Test error")
+
+        # Process should not crash
+        result = inference_service.process_image(sample_image)
+
+        # Verify error handling
+        assert not result.success
+        assert len(result.errors) > 0
+        assert "Test error" in result.errors[0]
+
+
+class TestInferenceServicePDFRendering:
+    """Test PDF rendering imports."""
+
+    @patch('src.inference.pipeline.InferencePipeline')
+    @patch('src.inference.yolo_detector.YOLODetector')
+    @patch('src.pdf.renderer.render_pdf_to_images')
+    @patch('ultralytics.YOLO')
+    def test_pdf_visualization_imports_correctly(
+        self,
+        mock_yolo_class,
+        mock_render_pdf,
+        mock_yolo_detector,
+        mock_pipeline,
+        inference_service,
+        tmp_path,
+    ):
+        """
+        Test that _save_pdf_visualization imports render_pdf_to_images correctly.
+
+        This catches the import error we had with:
+        from ..pdf.renderer (wrong) vs from src.pdf.renderer (correct)
+        """
+        # Setup mocks
+        mock_detector_instance = Mock()
+        mock_pipeline_instance = Mock()
+        mock_yolo_detector.return_value = mock_detector_instance
+        mock_pipeline.return_value = mock_pipeline_instance
+
+        # Create a fake PDF path
+        pdf_path = tmp_path / "test.pdf"
+        pdf_path.touch()
+
+        # Mock render_pdf_to_images to return an image
+        image_bytes = io.BytesIO()
+        img = Image.new('RGB', (800, 1200), color='white')
+        img.save(image_bytes, format='PNG')
+        mock_render_pdf.return_value = [(1, image_bytes.getvalue())]
+
+        # Mock YOLO
+        mock_model_instance = Mock()
+        mock_result = Mock()
+        mock_result.save = Mock()
+        mock_model_instance.predict.return_value = [mock_result]
+        mock_yolo_class.return_value = mock_model_instance
+
+        # This should not raise ImportError
+        result_path = inference_service._save_pdf_visualization(pdf_path, "test123")
+
+        # Verify import was successful
+        mock_render_pdf.assert_called_once()
+        assert result_path is not None
+
+
+@pytest.mark.skipif(
+    not Path("runs/train/invoice_fields/weights/best.pt").exists(),
+    reason="Model file not available"
+)
+class TestInferenceServiceRealModel:
+    """Integration tests with real model (skip if model not available)."""
+
+    def test_real_initialization(self, model_config, storage_config):
+        """Test real initialization with actual model."""
+        service = InferenceService(model_config, storage_config)
+
+        # This should work with the real imports
+        service.initialize()
+
+        assert service.is_initialized
+        assert service._pipeline is not None
+        assert service._detector is not None

tests/web/test_rate_limiter.py

@@ -0,0 +1,154 @@
+"""
+Tests for the RateLimiter class.
+"""
+
+import time
+from datetime import datetime, timedelta
+from unittest.mock import MagicMock
+
+import pytest
+
+from src.data.async_request_db import ApiKeyConfig
+from src.web.rate_limiter import RateLimiter, RateLimitConfig, RateLimitStatus
+
+
+class TestRateLimiter:
+    """Tests for RateLimiter."""
+
+    def test_check_submit_limit_allowed(self, rate_limiter, mock_db):
+        """Test that requests are allowed under the limit."""
+        status = rate_limiter.check_submit_limit("test-api-key")
+
+        assert status.allowed is True
+        assert status.remaining_requests >= 0
+        assert status.retry_after_seconds is None
+
+    def test_check_submit_limit_rate_exceeded(self, rate_limiter, mock_db):
+        """Test rate limit exceeded when too many requests."""
+        # Record 10 requests (the default limit)
+        for _ in range(10):
+            rate_limiter.record_request("test-api-key")
+
+        status = rate_limiter.check_submit_limit("test-api-key")
+
+        assert status.allowed is False
+        assert status.remaining_requests == 0
+        assert status.retry_after_seconds is not None
+        assert status.retry_after_seconds > 0
+        assert "rate limit" in status.reason.lower()
+
+    def test_check_submit_limit_concurrent_jobs_exceeded(self, rate_limiter, mock_db):
+        """Test rejection when max concurrent jobs reached."""
+        # Mock active jobs at the limit
+        mock_db.count_active_jobs.return_value = 3  # Max is 3
+
+        status = rate_limiter.check_submit_limit("test-api-key")
+
+        assert status.allowed is False
+        assert "concurrent" in status.reason.lower()
+
+    def test_record_request(self, rate_limiter, mock_db):
+        """Test that recording a request works."""
+        rate_limiter.record_request("test-api-key")
+
+        # Should have called the database
+        mock_db.record_rate_limit_event.assert_called_once_with("test-api-key", "request")
+
+    def test_check_poll_limit_allowed(self, rate_limiter, mock_db):
+        """Test that polling is allowed initially."""
+        status = rate_limiter.check_poll_limit("test-api-key", "request-123")
+
+        assert status.allowed is True
+
+    def test_check_poll_limit_too_frequent(self, rate_limiter, mock_db):
+        """Test that rapid polling is rejected."""
+        # First poll should succeed
+        status1 = rate_limiter.check_poll_limit("test-api-key", "request-123")
+        assert status1.allowed is True
+
+        # Immediate second poll should fail
+        status2 = rate_limiter.check_poll_limit("test-api-key", "request-123")
+        assert status2.allowed is False
+        assert "polling" in status2.reason.lower()
+        assert status2.retry_after_seconds is not None
+
+    def test_check_poll_limit_different_requests(self, rate_limiter, mock_db):
+        """Test that different request_ids have separate poll limits."""
+        # Poll request 1
+        status1 = rate_limiter.check_poll_limit("test-api-key", "request-1")
+        assert status1.allowed is True
+
+        # Poll request 2 should also be allowed
+        status2 = rate_limiter.check_poll_limit("test-api-key", "request-2")
+        assert status2.allowed is True
+
+    def test_sliding_window_expires(self, rate_limiter, mock_db):
+        """Test that requests expire from the sliding window."""
+        # Record requests
+        for _ in range(5):
+            rate_limiter.record_request("test-api-key")
+
+        # Check status - should have 5 remaining
+        status1 = rate_limiter.check_submit_limit("test-api-key")
+        assert status1.allowed is True
+        assert status1.remaining_requests == 4  # 10 - 5 - 1 (for this check)
+
+    def test_get_rate_limit_headers(self, rate_limiter):
+        """Test rate limit header generation."""
+        status = RateLimitStatus(
+            allowed=False,
+            remaining_requests=0,
+            reset_at=datetime.utcnow() + timedelta(seconds=30),
+            retry_after_seconds=30,
+        )
+
+        headers = rate_limiter.get_rate_limit_headers(status)
+
+        assert "X-RateLimit-Remaining" in headers
+        assert headers["X-RateLimit-Remaining"] == "0"
+        assert "Retry-After" in headers
+        assert headers["Retry-After"] == "30"
+
+    def test_cleanup_poll_timestamps(self, rate_limiter, mock_db):
+        """Test cleanup of old poll timestamps."""
+        # Add some poll timestamps
+        rate_limiter.check_poll_limit("test-api-key", "old-request")
+
+        # Manually age the timestamp
+        rate_limiter._poll_timestamps[("test-api-key", "old-request")] = time.time() - 7200
+
+        # Run cleanup with 1 hour max age
+        cleaned = rate_limiter.cleanup_poll_timestamps(max_age_seconds=3600)
+
+        assert cleaned == 1
+        assert ("test-api-key", "old-request") not in rate_limiter._poll_timestamps
+
+    def test_cleanup_request_windows(self, rate_limiter, mock_db):
+        """Test cleanup of empty request windows."""
+        # Add some old requests
+        rate_limiter._request_windows["old-key"] = [time.time() - 120]
+
+        # Run cleanup
+        rate_limiter.cleanup_request_windows()
+
+        # Old entries should be removed
+        assert "old-key" not in rate_limiter._request_windows
+
+    def test_config_caching(self, rate_limiter, mock_db):
+        """Test that API key configs are cached."""
+        # First call should query database
+        rate_limiter._get_config("test-api-key")
+        assert mock_db.get_api_key_config.call_count == 1
+
+        # Second call should use cache
+        rate_limiter._get_config("test-api-key")
+        assert mock_db.get_api_key_config.call_count == 1  # Still 1
+
+    def test_default_config_for_unknown_key(self, rate_limiter, mock_db):
+        """Test that unknown API keys get default config."""
+        mock_db.get_api_key_config.return_value = None
+
+        config = rate_limiter._get_config("unknown-key")
+
+        assert config.requests_per_minute == 10  # Default
+        assert config.max_concurrent_jobs == 3  # Default

tests/web/test_training_phase4.py

@@ -0,0 +1,384 @@
+"""
+Tests for Phase 4: Training Data Management
+"""
+
+import pytest
+from datetime import datetime
+from uuid import uuid4
+
+from fastapi import FastAPI
+from fastapi.testclient import TestClient
+
+from src.web.api.v1.admin.training import create_training_router
+from src.web.core.auth import validate_admin_token, get_admin_db
+
+
+class MockTrainingTask:
+    """Mock TrainingTask for testing."""
+
+    def __init__(self, **kwargs):
+        self.task_id = kwargs.get('task_id', uuid4())
+        self.admin_token = kwargs.get('admin_token', 'test-token')
+        self.name = kwargs.get('name', 'Test Training')
+        self.description = kwargs.get('description', None)
+        self.status = kwargs.get('status', 'completed')
+        self.task_type = kwargs.get('task_type', 'train')
+        self.config = kwargs.get('config', {})
+        self.scheduled_at = kwargs.get('scheduled_at', None)
+        self.cron_expression = kwargs.get('cron_expression', None)
+        self.is_recurring = kwargs.get('is_recurring', False)
+        self.started_at = kwargs.get('started_at', datetime.utcnow())
+        self.completed_at = kwargs.get('completed_at', datetime.utcnow())
+        self.error_message = kwargs.get('error_message', None)
+        self.result_metrics = kwargs.get('result_metrics', {})
+        self.model_path = kwargs.get('model_path', 'runs/train/test/weights/best.pt')
+        self.document_count = kwargs.get('document_count', 0)
+        self.metrics_mAP = kwargs.get('metrics_mAP', 0.935)
+        self.metrics_precision = kwargs.get('metrics_precision', 0.92)
+        self.metrics_recall = kwargs.get('metrics_recall', 0.88)
+        self.created_at = kwargs.get('created_at', datetime.utcnow())
+        self.updated_at = kwargs.get('updated_at', datetime.utcnow())
+
+
+class MockTrainingDocumentLink:
+    """Mock TrainingDocumentLink for testing."""
+
+    def __init__(self, **kwargs):
+        self.link_id = kwargs.get('link_id', uuid4())
+        self.task_id = kwargs.get('task_id')
+        self.document_id = kwargs.get('document_id')
+        self.annotation_snapshot = kwargs.get('annotation_snapshot', None)
+        self.created_at = kwargs.get('created_at', datetime.utcnow())
+
+
+class MockAdminDocument:
+    """Mock AdminDocument for testing."""
+
+    def __init__(self, **kwargs):
+        self.document_id = kwargs.get('document_id', uuid4())
+        self.admin_token = kwargs.get('admin_token', 'test-token')
+        self.filename = kwargs.get('filename', 'test.pdf')
+        self.file_size = kwargs.get('file_size', 100000)
+        self.content_type = kwargs.get('content_type', 'application/pdf')
+        self.file_path = kwargs.get('file_path', 'data/admin_docs/test.pdf')
+        self.page_count = kwargs.get('page_count', 1)
+        self.status = kwargs.get('status', 'labeled')
+        self.auto_label_status = kwargs.get('auto_label_status', None)
+        self.auto_label_error = kwargs.get('auto_label_error', None)
+        self.upload_source = kwargs.get('upload_source', 'ui')
+        self.batch_id = kwargs.get('batch_id', None)
+        self.csv_field_values = kwargs.get('csv_field_values', None)
+        self.auto_label_queued_at = kwargs.get('auto_label_queued_at', None)
+        self.annotation_lock_until = kwargs.get('annotation_lock_until', None)
+        self.created_at = kwargs.get('created_at', datetime.utcnow())
+        self.updated_at = kwargs.get('updated_at', datetime.utcnow())
+
+
+class MockAnnotation:
+    """Mock AdminAnnotation for testing."""
+
+    def __init__(self, **kwargs):
+        self.annotation_id = kwargs.get('annotation_id', uuid4())
+        self.document_id = kwargs.get('document_id')
+        self.page_number = kwargs.get('page_number', 1)
+        self.class_id = kwargs.get('class_id', 0)
+        self.class_name = kwargs.get('class_name', 'invoice_number')
+        self.bbox_x = kwargs.get('bbox_x', 100)
+        self.bbox_y = kwargs.get('bbox_y', 100)
+        self.bbox_width = kwargs.get('bbox_width', 200)
+        self.bbox_height = kwargs.get('bbox_height', 50)
+        self.x_center = kwargs.get('x_center', 0.5)
+        self.y_center = kwargs.get('y_center', 0.5)
+        self.width = kwargs.get('width', 0.3)
+        self.height = kwargs.get('height', 0.1)
+        self.text_value = kwargs.get('text_value', 'INV-001')
+        self.confidence = kwargs.get('confidence', 0.95)
+        self.source = kwargs.get('source', 'manual')
+        self.is_verified = kwargs.get('is_verified', False)
+        self.verified_at = kwargs.get('verified_at', None)
+        self.verified_by = kwargs.get('verified_by', None)
+        self.override_source = kwargs.get('override_source', None)
+        self.original_annotation_id = kwargs.get('original_annotation_id', None)
+        self.created_at = kwargs.get('created_at', datetime.utcnow())
+        self.updated_at = kwargs.get('updated_at', datetime.utcnow())
+
+
+class MockAdminDB:
+    """Mock AdminDB for testing Phase 4."""
+
+    def __init__(self):
+        self.documents = {}
+        self.annotations = {}
+        self.training_tasks = {}
+        self.training_links = {}
+
+    def get_documents_for_training(
+        self,
+        admin_token,
+        status="labeled",
+        has_annotations=True,
+        min_annotation_count=None,
+        exclude_used_in_training=False,
+        limit=100,
+        offset=0,
+    ):
+        """Get documents for training."""
+        # Filter documents by criteria
+        filtered = []
+        for doc in self.documents.values():
+            if doc.admin_token != admin_token or doc.status != status:
+                continue
+
+            # Check annotations
+            annotations = self.annotations.get(str(doc.document_id), [])
+            if has_annotations and len(annotations) == 0:
+                continue
+            if min_annotation_count and len(annotations) < min_annotation_count:
+                continue
+
+            # Check if used in training
+            if exclude_used_in_training:
+                links = self.training_links.get(str(doc.document_id), [])
+                if links:
+                    continue
+
+            filtered.append(doc)
+
+        total = len(filtered)
+        return filtered[offset:offset+limit], total
+
+    def get_annotations_for_document(self, document_id):
+        """Get annotations for document."""
+        return self.annotations.get(str(document_id), [])
+
+    def get_document_training_tasks(self, document_id):
+        """Get training tasks that used this document."""
+        return self.training_links.get(str(document_id), [])
+
+    def get_training_tasks_by_token(
+        self,
+        admin_token,
+        status=None,
+        limit=20,
+        offset=0,
+    ):
+        """Get training tasks filtered by token."""
+        tasks = [t for t in self.training_tasks.values() if t.admin_token == admin_token]
+        if status:
+            tasks = [t for t in tasks if t.status == status]
+
+        total = len(tasks)
+        return tasks[offset:offset+limit], total
+
+    def get_training_task(self, task_id):
+        """Get training task by ID."""
+        return self.training_tasks.get(str(task_id))
+
+
+@pytest.fixture
+def app():
+    """Create test FastAPI app."""
+    app = FastAPI()
+
+    # Create mock DB
+    mock_db = MockAdminDB()
+
+    # Add test documents
+    doc1 = MockAdminDocument(
+        filename="DOC001.pdf",
+        status="labeled",
+    )
+    doc2 = MockAdminDocument(
+        filename="DOC002.pdf",
+        status="labeled",
+    )
+    doc3 = MockAdminDocument(
+        filename="DOC003.pdf",
+        status="labeled",
+    )
+
+    mock_db.documents[str(doc1.document_id)] = doc1
+    mock_db.documents[str(doc2.document_id)] = doc2
+    mock_db.documents[str(doc3.document_id)] = doc3
+
+    # Add annotations
+    mock_db.annotations[str(doc1.document_id)] = [
+        MockAnnotation(document_id=doc1.document_id, source="manual"),
+        MockAnnotation(document_id=doc1.document_id, source="auto"),
+    ]
+    mock_db.annotations[str(doc2.document_id)] = [
+        MockAnnotation(document_id=doc2.document_id, source="auto"),
+        MockAnnotation(document_id=doc2.document_id, source="auto"),
+        MockAnnotation(document_id=doc2.document_id, source="auto"),
+    ]
+    # doc3 has no annotations
+
+    # Add training tasks
+    task1 = MockTrainingTask(
+        name="Training Run 2024-01",
+        status="completed",
+        document_count=500,
+        metrics_mAP=0.935,
+        metrics_precision=0.92,
+        metrics_recall=0.88,
+    )
+    task2 = MockTrainingTask(
+        name="Training Run 2024-02",
+        status="completed",
+        document_count=600,
+        metrics_mAP=0.951,
+        metrics_precision=0.94,
+        metrics_recall=0.92,
+    )
+
+    mock_db.training_tasks[str(task1.task_id)] = task1
+    mock_db.training_tasks[str(task2.task_id)] = task2
+
+    # Add training links (doc1 used in task1)
+    link1 = MockTrainingDocumentLink(
+        task_id=task1.task_id,
+        document_id=doc1.document_id,
+    )
+    mock_db.training_links[str(doc1.document_id)] = [link1]
+
+    # Override dependencies
+    app.dependency_overrides[validate_admin_token] = lambda: "test-token"
+    app.dependency_overrides[get_admin_db] = lambda: mock_db
+
+    # Include router
+    router = create_training_router()
+    app.include_router(router)
+
+    return app
+
+
+@pytest.fixture
+def client(app):
+    """Create test client."""
+    return TestClient(app)
+
+
+class TestTrainingDocuments:
+    """Tests for GET /admin/training/documents endpoint."""
+
+    def test_get_training_documents_success(self, client):
+        """Test getting documents for training."""
+        response = client.get("/admin/training/documents")
+
+        assert response.status_code == 200
+        data = response.json()
+        assert "total" in data
+        assert "documents" in data
+        assert data["total"] >= 0
+        assert isinstance(data["documents"], list)
+
+    def test_get_training_documents_with_annotations(self, client):
+        """Test filtering documents with annotations."""
+        response = client.get("/admin/training/documents?has_annotations=true")
+
+        assert response.status_code == 200
+        data = response.json()
+        # Should return doc1 and doc2 (both have annotations)
+        assert data["total"] == 2
+
+    def test_get_training_documents_min_annotation_count(self, client):
+        """Test filtering by minimum annotation count."""
+        response = client.get("/admin/training/documents?min_annotation_count=3")
+
+        assert response.status_code == 200
+        data = response.json()
+        # Should return only doc2 (has 3 annotations)
+        assert data["total"] == 1
+
+    def test_get_training_documents_exclude_used(self, client):
+        """Test excluding documents already used in training."""
+        response = client.get("/admin/training/documents?exclude_used_in_training=true")
+
+        assert response.status_code == 200
+        data = response.json()
+        # Should exclude doc1 (used in training)
+        assert data["total"] == 1  # Only doc2 (doc3 has no annotations)
+
+    def test_get_training_documents_annotation_sources(self, client):
+        """Test that annotation sources are included."""
+        response = client.get("/admin/training/documents?has_annotations=true")
+
+        assert response.status_code == 200
+        data = response.json()
+        # Check that documents have annotation_sources field
+        for doc in data["documents"]:
+            assert "annotation_sources" in doc
+            assert isinstance(doc["annotation_sources"], dict)
+            assert "manual" in doc["annotation_sources"]
+            assert "auto" in doc["annotation_sources"]
+
+    def test_get_training_documents_pagination(self, client):
+        """Test pagination parameters."""
+        response = client.get("/admin/training/documents?limit=1&offset=0")
+
+        assert response.status_code == 200
+        data = response.json()
+        assert data["limit"] == 1
+        assert data["offset"] == 0
+        assert len(data["documents"]) <= 1
+
+
+class TestTrainingModels:
+    """Tests for GET /admin/training/models endpoint."""
+
+    def test_get_training_models_success(self, client):
+        """Test getting trained models list."""
+        response = client.get("/admin/training/models")
+
+        assert response.status_code == 200
+        data = response.json()
+        assert "total" in data
+        assert "models" in data
+        assert data["total"] == 2
+        assert len(data["models"]) == 2
+
+    def test_get_training_models_includes_metrics(self, client):
+        """Test that models include metrics."""
+        response = client.get("/admin/training/models")
+
+        assert response.status_code == 200
+        data = response.json()
+        # Check first model has metrics
+        model = data["models"][0]
+        assert "metrics" in model
+        assert "mAP" in model["metrics"]
+        assert model["metrics"]["mAP"] is not None
+        assert "precision" in model["metrics"]
+        assert "recall" in model["metrics"]
+
+    def test_get_training_models_includes_download_url(self, client):
+        """Test that completed models have download URLs."""
+        response = client.get("/admin/training/models")
+
+        assert response.status_code == 200
+        data = response.json()
+        # Check completed models have download URLs
+        for model in data["models"]:
+            if model["status"] == "completed":
+                assert "download_url" in model
+                assert model["download_url"] is not None
+
+    def test_get_training_models_filter_by_status(self, client):
+        """Test filtering models by status."""
+        response = client.get("/admin/training/models?status=completed")
+
+        assert response.status_code == 200
+        data = response.json()
+        # All returned models should be completed
+        for model in data["models"]:
+            assert model["status"] == "completed"
+
+    def test_get_training_models_pagination(self, client):
+        """Test pagination for models."""
+        response = client.get("/admin/training/models?limit=1&offset=0")
+
+        assert response.status_code == 200
+        data = response.json()
+        assert data["limit"] == 1
+        assert data["offset"] == 0
+        assert len(data["models"]) == 1