Preload Paddle OCR models and stabilize runtime deps

Dockerfile

@@ -3,12 +3,15 @@ FROM python:3.10-slim
 ENV PYTHONDONTWRITEBYTECODE=1
 ENV PYTHONUNBUFFERED=1
 ENV PADDLE_PDX_DISABLE_MODEL_SOURCE_CHECK=True
+ENV PADDLEOCR_HOME=/app/.paddleocr
 
 WORKDIR /app
 
 COPY requirements.txt .
 RUN pip install --no-cache-dir -r requirements.txt
 
+RUN mkdir -p /app/.paddleocr && python -c "from paddleocr import PaddleOCR; PaddleOCR(use_angle_cls=True, lang='en', use_gpu=False, show_log=False); print('PaddleOCR models ready')"
+
 COPY . .
 
 EXPOSE 7860

requirements.txt

@@ -11,5 +11,6 @@ pyarrow
 pytest
 python-dotenv
 python-multipart
+requests>=2.32.3
 reportlab
 uvicorn[standard]

simpliscribe/web.py

@@ -1,3 +1,4 @@
+import logging
 import re
 import uuid
 from datetime import datetime, timezone
@@ -15,6 +16,9 @@ from .reporting import build_pdf_report
 from .storage import append_history, get_analysis_record, load_history
 
 
+logger = logging.getLogger(__name__)
+
+
 def utc_now_iso() -> str:
     return datetime.now(timezone.utc).isoformat()
 
@@ -109,6 +113,7 @@ async def analyze(file: UploadFile = File(...)) -> JSONResponse:
     except HTTPException:
         raise
     except Exception as exc:
+        logger.exception("Prescription analysis failed.")
         message = str(exc)
         if "PDX has already been initialized" in message:
             message = "OCR engine is warming up. Please retry in a few seconds."