Merge branch 'dev' of https://gitea.maskantech.in/gitea_admin/pdf into furqan

gateway/app/routers/documents.py

@@ -737,3 +737,104 @@ def export_document(document_id: str):
         )
     except Exception as e:
         raise HTTPException(status_code=status.HTTP_500_INTERNAL_SERVER_ERROR, detail=str(e))
+
+class TextObjectResponse(BaseModel):
+    text: str
+    fontName: str
+    fontSize: float
+    tm: list[float]
+
+@router.get("/{document_id}/pages/{page_index}/text_objects", response_model=list[TextObjectResponse])
+def get_text_objects(document_id: str, page_index: int) -> list[TextObjectResponse]:
+    if not engine.is_available():
+        raise HTTPException(status_code=501, detail="Engine not available")
+        
+    doc_info = document_store.get_document(document_id)
+    if not doc_info:
+        raise HTTPException(status_code=404, detail="Document not found")
+        
+    pdfengine = engine.require()
+    
+    import os
+    # For StreamEditor we need a real file path. If it's loaded from memory, we need to save it to a temp file.
+    # In this MVP, we assume the file was saved somewhere, but actually document_store keeps it in memory.
+    # Wait, doc_store has doc_info["filename"] and bytes_data. Let's write bytes to a temp file.
+    import tempfile
+    
+    with tempfile.NamedTemporaryFile(delete=False, suffix=".pdf") as tmp:
+        tmp.write(doc_info["bytes_data"])
+        tmp_path = tmp.name
+        
+    try:
+        editor = pdfengine.StreamEditor(tmp_path)
+        objects = editor.extract_text_objects(page_index)
+        
+        result = []
+        for obj in objects:
+            # obj["text"] is now returned as bytes from C++
+            text_str = obj["text"].decode("latin-1") if isinstance(obj["text"], bytes) else obj["text"]
+            result.append(TextObjectResponse(
+                text=text_str,
+                fontName=obj["fontName"],
+                fontSize=obj["fontSize"],
+                tm=obj["tm"]
+            ))
+        return result
+    except Exception as e:
+        raise HTTPException(status_code=500, detail=str(e))
+    finally:
+        if os.path.exists(tmp_path):
+            os.remove(tmp_path)
+
+
+class UpdateTextObjectRequest(BaseModel):
+    new_text: str
+
+
+@router.put("/{document_id}/pages/{page_index}/text_objects/{object_index}")
+def replace_text_object(document_id: str, page_index: int, object_index: int, req: UpdateTextObjectRequest) -> dict:
+    if not engine.is_available():
+        raise HTTPException(status_code=501, detail="Engine not available")
+        
+    doc_info = document_store.get_document(document_id)
+    if not doc_info:
+        raise HTTPException(status_code=404, detail="Document not found")
+        
+    pdfengine = engine.require()
+    
+    import tempfile
+    import os
+    
+    with tempfile.NamedTemporaryFile(delete=False, suffix=".pdf") as tmp:
+        tmp.write(doc_info["bytes_data"])
+        tmp_path = tmp.name
+        
+    out_path = tmp_path + ".out.pdf"
+        
+    try:
+        editor = pdfengine.StreamEditor(tmp_path)
+        # Convert frontend string back to exact bytes using latin-1
+        new_text_bytes = req.new_text.encode("latin-1")
+        success = editor.replace_text_object(page_index, object_index, new_text_bytes, out_path)
+        if not success:
+            raise HTTPException(status_code=400, detail="Failed to replace text object (not found or identical)")
+            
+        # Update the document store with the new bytes
+        with open(out_path, "rb") as f:
+            new_bytes = f.read()
+            
+        # Reload doc instance
+        doc_instance = pdfengine.PdfDocument.load_from_memory(new_bytes, "")
+        
+        # Update the store
+        doc_info["bytes_data"] = new_bytes
+        doc_info["doc_instance"] = doc_instance
+        
+        return {"success": True}
+    except Exception as e:
+        raise HTTPException(status_code=500, detail=str(e))
+    finally:
+        if os.path.exists(tmp_path):
+            os.remove(tmp_path)
+        if os.path.exists(out_path):
+            os.remove(out_path)