feat: add log for export api

.claude/settings.local.json

@@ -8,7 +8,8 @@
       "WebFetch(domain:raw.githubusercontent.com)",
       "Bash(python -c \"\nfrom app.services.glm_postprocess import GLMResultFormatter, clean_repeated_content, clean_formula_number\nf = GLMResultFormatter\\(\\)\nprint\\('GLMResultFormatter OK'\\)\nprint\\('clean_formula_number:', clean_formula_number\\('\\(2.1\\)'\\)\\)\nregions = [\n    {'index': 0, 'label': 'text', 'native_label': 'doc_title', 'content': 'Introduction', 'bbox_2d': [10,10,990,50]},\n    {'index': 1, 'label': 'formula', 'native_label': 'display_formula', 'content': r'\\\\frac{a}{b}', 'bbox_2d': [10,60,990,200]},\n    {'index': 2, 'label': 'text', 'native_label': 'formula_number', 'content': '\\(1\\)', 'bbox_2d': [900,60,990,200]},\n]\nmd = f.process\\(regions\\)\nprint\\('process output:'\\)\nprint\\(md\\)\n\" 2>&1 | grep -v \"^$\")",
       "Bash(python3 -c \"\nfrom app.services.glm_postprocess import GLMResultFormatter, clean_repeated_content, clean_formula_number\nf = GLMResultFormatter\\(\\)\nprint\\('GLMResultFormatter OK'\\)\nprint\\('clean_formula_number:', clean_formula_number\\('\\(2.1\\)'\\)\\)\nregions = [\n    {'index': 0, 'label': 'text', 'native_label': 'doc_title', 'content': 'Introduction', 'bbox_2d': [10,10,990,50]},\n    {'index': 1, 'label': 'formula', 'native_label': 'display_formula', 'content': r'\\\\frac{a}{b}', 'bbox_2d': [10,60,990,200]},\n    {'index': 2, 'label': 'text', 'native_label': 'formula_number', 'content': '\\(1\\)', 'bbox_2d': [900,60,990,200]},\n]\nmd = f.process\\(regions\\)\nprint\\('process output:'\\)\nprint\\(repr\\(md\\)\\)\n\" 2>&1)",
-      "Bash(ls .venv 2>/dev/null || ls venv 2>/dev/null || echo \"no venv found\" && find . -name \"activate\" -path \"*/bin/activate\" 2>/dev/null | head -3)"
+      "Bash(ls .venv 2>/dev/null || ls venv 2>/dev/null || echo \"no venv found\" && find . -name \"activate\" -path \"*/bin/activate\" 2>/dev/null | head -3)",
+      "Bash(ruff check:*)"
     ]
   }
 }

app/api/v1/endpoints/convert.py

@@ -1,5 +1,7 @@
 """Format conversion endpoints."""
 
+import logging
+
 from fastapi import APIRouter, Depends, HTTPException
 from fastapi.responses import Response
 
@@ -7,6 +9,8 @@ from app.core.dependencies import get_converter
 from app.schemas.convert import LatexToOmmlRequest, LatexToOmmlResponse, MarkdownToDocxRequest
 from app.services.converter import Converter
 
+logger = logging.getLogger(__name__)
+
 router = APIRouter()
 
 
@@ -19,14 +23,25 @@ async def convert_markdown_to_docx(
 
     Returns the generated DOCX file as a binary response.
     """
+    logger.info(
+        "Converting markdown to DOCX, filename=%s, content_length=%d",
+        request.filename,
+        len(request.markdown),
+    )
     try:
         docx_bytes = converter.export_to_file(request.markdown, export_type="docx")
+        logger.info(
+            "DOCX conversion successful, filename=%s, size=%d bytes",
+            request.filename,
+            len(docx_bytes),
+        )
         return Response(
             content=docx_bytes,
             media_type="application/vnd.openxmlformats-officedocument.wordprocessingml.document",
             headers={"Content-Disposition": f'attachment; filename="{request.filename}.docx"'},
         )
     except Exception as e:
+        logger.exception("DOCX conversion failed, filename=%s: %s", request.filename, e)
         raise HTTPException(status_code=500, detail=f"Conversion failed: {e}")
 
 
@@ -55,12 +70,17 @@ async def convert_latex_to_omml(
         ```
     """
     if not request.latex or not request.latex.strip():
+        logger.warning("LaTeX to OMML request received with empty formula")
         raise HTTPException(status_code=400, detail="LaTeX formula cannot be empty")
 
+    logger.info("Converting LaTeX to OMML, latex=%r", request.latex)
     try:
         omml = converter.convert_to_omml(request.latex)
+        logger.info("LaTeX to OMML conversion successful")
         return LatexToOmmlResponse(omml=omml)
     except ValueError as e:
+        logger.warning("LaTeX to OMML conversion invalid input: %s", e)
         raise HTTPException(status_code=400, detail=str(e))
     except RuntimeError as e:
+        logger.error("LaTeX to OMML conversion runtime error: %s", e)
         raise HTTPException(status_code=503, detail=str(e))

app/core/logging_config.py

@@ -2,11 +2,15 @@
 
 import logging
 import logging.handlers
+from contextvars import ContextVar
 from pathlib import Path
 from typing import Any
 
 from app.core.config import get_settings
 
+# Context variable to hold the current request_id across async boundaries
+request_id_ctx: ContextVar[str] = ContextVar("request_id", default="-")
+
 
 class TimedRotatingAndSizeFileHandler(logging.handlers.TimedRotatingFileHandler):
     """File handler that rotates by both time (daily) and size (100MB)."""
@@ -92,14 +96,13 @@ def setup_logging(log_dir: str | None = None) -> logging.Logger:
     # Remove existing handlers to avoid duplicates
     logger.handlers.clear()
 
-    # Create custom formatter that handles missing request_id
+    # Create custom formatter that automatically injects request_id from context
     class RequestIDFormatter(logging.Formatter):
-        """Formatter that handles request_id in log records."""
+        """Formatter that injects request_id from ContextVar into log records."""
 
         def format(self, record):
-            # Add request_id if not present
             if not hasattr(record, "request_id"):
-                record.request_id = getattr(record, "request_id", "unknown")
+                record.request_id = request_id_ctx.get()
             return super().format(record)
 
     formatter = RequestIDFormatter(

app/main.py

@@ -8,6 +8,7 @@ from app.api.v1.router import api_router
 from app.core.config import get_settings
 from app.core.dependencies import init_layout_detector
 from app.core.logging_config import setup_logging
+from app.middleware.request_id import RequestIDMiddleware
 
 settings = get_settings()
 
@@ -33,6 +34,8 @@ app = FastAPI(
     lifespan=lifespan,
 )
 
+app.add_middleware(RequestIDMiddleware)
+
 # Include API router
 app.include_router(api_router, prefix=settings.api_prefix)
 

app/middleware/request_id.py

@@ -0,0 +1,34 @@
+"""Middleware to propagate or generate request_id for every request."""
+
+import uuid
+
+from starlette.middleware.base import BaseHTTPMiddleware
+from starlette.requests import Request
+from starlette.responses import Response
+
+from app.core.logging_config import request_id_ctx
+
+REQUEST_ID_HEADER = "X-Request-ID"
+
+
+class RequestIDMiddleware(BaseHTTPMiddleware):
+    """Extract X-Request-ID from incoming request headers or generate one.
+
+    The request_id is stored in a ContextVar so that all log records emitted
+    during the request are automatically annotated with it, without needing to
+    pass it explicitly through every call.
+
+    The same request_id is also echoed back in the response header so that
+    callers can correlate logs.
+    """
+
+    async def dispatch(self, request: Request, call_next) -> Response:
+        request_id = request.headers.get(REQUEST_ID_HEADER) or str(uuid.uuid4())
+        token = request_id_ctx.set(request_id)
+        try:
+            response = await call_next(request)
+        finally:
+            request_id_ctx.reset(token)
+
+        response.headers[REQUEST_ID_HEADER] = request_id
+        return response

tests/tools/layout.py

@@ -0,0 +1,35 @@
+import cv2
+
+from app.core.config import get_settings
+from app.services.layout_detector import LayoutDetector
+
+settings = get_settings()
+
+
+def debug_layout_detector():
+    layout_detector = LayoutDetector()
+    image = cv2.imread("test/image2.png")
+
+    print(f"Image shape: {image.shape}")
+
+    # padded_image = ImageProcessor(padding_ratio=0.15).add_padding(image)
+    layout_info = layout_detector.detect(image)
+
+    # draw the layout info and label
+    for region in layout_info.regions:
+        x1, y1, x2, y2 = region.bbox
+        cv2.putText(
+            image,
+            region.native_label,
+            (int(x1), int(y1)),
+            cv2.FONT_HERSHEY_SIMPLEX,
+            0.5,
+            (0, 0, 255),
+            2,
+        )
+        cv2.rectangle(image, (int(x1), int(y1)), (int(x2), int(y2)), (0, 0, 255), 2)
+    cv2.imwrite("test/layout_debug.png", image)
+
+
+if __name__ == "__main__":
+    debug_layout_detector()