Update app.py

app.py

@@ -1,9 +1,9 @@
 #!/usr/bin/env python3
 
-import json
 import logging
 import os
 import subprocess
+import tempfile
 import time
 from typing import Any, List
 
@@ -28,6 +28,25 @@ _parser = None
 _vllm_proc = None
 
 
+def _render_pdf_pages_to_images(pdf_path: str) -> List[str]:
+    import pymupdf as fitz
+
+    doc = fitz.open(pdf_path)
+    page_images: List[str] = []
+    try:
+        for i in range(len(doc)):
+            page = doc[i]
+            pix = page.get_pixmap(matrix=fitz.Matrix(2.5, 2.5), alpha=False)
+            img_path = os.path.join(
+                tempfile.gettempdir(), f"glmocr_page_{os.getpid()}_{i}_{int(time.time()*1000)}.png"
+            )
+            pix.save(img_path)
+            page_images.append(img_path)
+    finally:
+        doc.close()
+    return page_images
+
+
 def _start_vllm_if_needed() -> None:
     global _vllm_proc
     if _vllm_proc is not None and _vllm_proc.poll() is None:
@@ -142,30 +161,34 @@ def _extract_md(result: Any) -> str:
     return str(md).strip() if md else ""
 
 
-def _extract_json(result: Any) -> str:
-    if result is None:
-        return "{}"
-    if isinstance(result, list):
-        payload = [getattr(item, "json_result", None) for item in result]
-        return json.dumps(payload, ensure_ascii=False, indent=2)
-    return json.dumps(getattr(result, "json_result", None), ensure_ascii=False, indent=2)
-
-
 def run_ocr(file_obj):
     if file_obj is None:
-        return "Please upload a file.", "{}"
+        return "Please upload a file."
     path = file_obj.name if hasattr(file_obj, "name") else str(file_obj)
+    page_images: List[str] = []
     try:
         parser = get_parser()
-        result = parser.parse(path)
+        is_pdf = path.lower().endswith(".pdf")
+        if is_pdf:
+            page_images = _render_pdf_pages_to_images(path)
+            # Parse one rendered page per request item so we can preserve page boundaries.
+            result = parser.parse(page_images)
+        else:
+            result = parser.parse(path)
         md = _extract_md(result) or "(No content)"
-        jr = _extract_json(result)
-        return md, jr
+        return md
     except Exception as e:
         import traceback
 
         log.exception("run_ocr failed: %s", e)
-        return f"Error: {e}\n\n{traceback.format_exc()}", "{}"
+        return f"Error: {e}\n\n{traceback.format_exc()}"
+    finally:
+        for p in page_images:
+            try:
+                if isinstance(p, str) and p.endswith(".png") and "glmocr_page_" in os.path.basename(p):
+                    os.unlink(p)
+            except Exception:
+                pass
 
 
 def build_demo():
@@ -181,8 +204,7 @@ def build_demo():
         )
         run_btn = gr.Button("Run OCR", variant="primary")
         out_md = gr.Textbox(label="Markdown", lines=30)
-        out_json = gr.Textbox(label="JSON (layout details)", lines=20)
-        run_btn.click(fn=run_ocr, inputs=file_in, outputs=[out_md, out_json])
+        run_btn.click(fn=run_ocr, inputs=file_in, outputs=out_md)
     return demo