Initial commit: Invoice field extraction system using YOLO + OCR

Features:
- Auto-labeling pipeline: CSV values -> PDF search -> YOLO annotations
- Flexible date matching: year-month match, nearby date tolerance
- PDF text extraction with PyMuPDF
- OCR support for scanned documents (PaddleOCR)
- YOLO training and inference pipeline
- 7 field types: InvoiceNumber, InvoiceDate, InvoiceDueDate, OCR, Bankgiro, Plusgiro, Amount

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>

scripts/run_train.sh

@@ -0,0 +1,67 @@
+#!/bin/bash
+# 训练运行脚本
+# 使用方法: bash scripts/run_train.sh
+
+set -e
+
+# 项目根目录
+PROJECT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")/.." && pwd)"
+cd "$PROJECT_DIR"
+
+# 激活虚拟环境
+if [ -f "venv/bin/activate" ]; then
+    source venv/bin/activate
+else
+    echo "错误: 虚拟环境不存在，请先运行 setup_wsl.sh"
+    exit 1
+fi
+
+# 默认参数
+DATA_YAML="${DATA_YAML:-data/dataset/dataset.yaml}"
+MODEL="${MODEL:-yolov8s.pt}"
+EPOCHS="${EPOCHS:-100}"
+BATCH_SIZE="${BATCH_SIZE:-16}"
+IMG_SIZE="${IMG_SIZE:-1280}"
+DEVICE="${DEVICE:-0}"
+
+# 检查数据集是否存在
+if [ ! -f "$DATA_YAML" ]; then
+    echo "错误: 数据集配置文件不存在: $DATA_YAML"
+    echo "请先运行自动标注: bash scripts/run_autolabel.sh"
+    exit 1
+fi
+
+# 显示配置
+echo "=========================================="
+echo "训练配置"
+echo "=========================================="
+echo "数据集:      $DATA_YAML"
+echo "基础模型:    $MODEL"
+echo "Epochs:      $EPOCHS"
+echo "Batch Size:  $BATCH_SIZE"
+echo "图像尺寸:    $IMG_SIZE"
+echo "设备:        $DEVICE"
+echo "=========================================="
+echo ""
+
+# 检查 GPU
+if command -v nvidia-smi &> /dev/null; then
+    echo "GPU 状态:"
+    nvidia-smi --query-gpu=name,memory.used,memory.total --format=csv,noheader
+    echo ""
+else
+    echo "警告: 未检测到 GPU，将使用 CPU 训练 (较慢)"
+    DEVICE="cpu"
+fi
+
+# 运行训练
+python -m src.cli.train \
+    --data "$DATA_YAML" \
+    --model "$MODEL" \
+    --epochs "$EPOCHS" \
+    --batch "$BATCH_SIZE" \
+    --imgsz "$IMG_SIZE" \
+    --device "$DEVICE"
+
+echo ""
+echo "训练完成! 模型保存在: runs/train/invoice_fields/weights/"