Update app.py

app.py

@@ -1,114 +1,119 @@
 # app.py
 import json
 import traceback
-import tempfile # Not strictly needed by process_upload anymore, but good to keep if other parts use it.
+import tempfile
 import os
 import gradio as gr
-from typing import Tuple, Optional, Any 
+from typing import Tuple, Optional, Any # Added Any for file_data_binary
 
 # Import config first to ensure JAVA_HOME is set early
 import config 
 
 from main_analyzer import analyze_pdf 
-# language_tool_python needed for the test in __main__
-import language_tool_python
+# Import language_tool_python only for the test in __main__ if needed
+# import language_tool_python 
+
+def process_upload(file_data_binary: Optional[Any]) -> Tuple[str, Optional[str]]: # Use Optional[Any] for Gradio File type="binary"
+    if file_data_binary is None or not hasattr(file_data_binary, 'read'): # Check if it's a file-like object
+        # Gradio's binary type for gr.File returns a tempfile._TemporaryFileWrapper object
+        # If it's None, no file was uploaded.
+        # If it's not None but doesn't have 'read', it's an unexpected type.
+        # However, gradio usually passes the bytes directly if type="binary" was used in older versions
+        # or a TemporaryFileWrapper which is file-like.
+        # For robustness, let's check if it's bytes.
+        if isinstance(file_data_binary, bytes):
+             pass # Good, it's bytes
+        elif file_data_binary is None:
+             return json.dumps({"error": "No file uploaded or file data is None"}, indent=2), None
+        elif not hasattr(file_data_binary, 'read'): # It's not None, not bytes, not file-like
+             return json.dumps({"error": f"Unexpected file data type: {type(file_data_binary)}"}), None
+        # If it has 'read', it's a file-like object, proceed.
+
+    # analyze_pdf now handles stream-to-temp-file logic internally via original_pdf_access_path
+    # So we can pass the file_data_binary (which is a file-like object from Gradio) directly.
+
+
+
+
+
 
 
-def process_upload(uploaded_file_input: Optional[Any]) -> Tuple[str, Optional[str]]:
-    """
-    Processes the uploaded file from Gradio.
-    'uploaded_file_input' is expected to be an object from gr.File(),
-    which could be a path string (like gradio.utils.NamedString) or a file-like object.
-    """
-    if uploaded_file_input is None:
-        print("App: No file uploaded.")
-        return json.dumps({"error": "No file uploaded."}, indent=2), None
 
     try:
-        # Pass the uploaded_file_input directly to analyze_pdf.
-        # analyze_pdf is responsible for determining if it's a path or a stream
-        # and handling it accordingly (including creating a temp file for streams if needed).
-        print(f"App: Received file input of type: {type(uploaded_file_input)}. Passing to analyzer.")
+        print(f"App: Processing uploaded file...")
+        # If file_data_binary is bytes, wrap it in BytesIO for file-like interface
+        # analyze_pdf expects a path or a file-like object with read() and seek()
         
-        results_dict, _ = analyze_pdf(uploaded_file_input) 
+        # Gradio with type="binary" gives a tempfile._TemporaryFileWrapper.
+        # This object is already file-like and can be passed directly.
+        # No need to create another temp file here in app.py if main_analyzer handles it.
         
-        # Check if analyze_pdf returned an error (it returns a dict with "error" key in that case)
-        if isinstance(results_dict, dict) and "error" in results_dict:
-            print(f"App: Analysis returned an error: {results_dict['error']}")
-            # Return the error JSON directly
-            return json.dumps(results_dict, indent=2, ensure_ascii=False), None
+        # analyze_pdf will create its own temp file if it receives a stream.
+        results_dict, _ = analyze_pdf(file_data_binary) 
+        
+
+
+
+
+
 
         results_json = json.dumps(results_dict, indent=2, ensure_ascii=False)
         return results_json, None 
 
     except Exception as e:
-        # This catches unexpected errors during the call to analyze_pdf or JSON dumping.
-        print(f"App: Error in process_upload: {e}\n{traceback.format_exc()}")
-        error_detail = str(e)
-        # If the error is due to an invalid input type that analyze_pdf couldn't handle:
-        if "Invalid PDF input type" in error_detail: # Check for specific error from analyze_pdf
-             error_detail = f"Invalid PDF input type received from uploader: {type(uploaded_file_input)}. Details: {str(e)}"
-        
-        error_message = json.dumps({"error": error_detail, "traceback": traceback.format_exc()}, indent=2)
+        print(f"Error in process_upload: {e}\n{traceback.format_exc()}")
+        error_message = json.dumps({"error": str(e), "traceback": traceback.format_exc()}, indent=2)
+
+
+
+
+
+
         return error_message, None
-    # No explicit temp file cleanup needed here in process_upload for the Gradio-provided file object.
-    # Gradio manages its own temporary files.
-    # analyze_pdf manages any internal temporary files it creates.
+    # No finally block needed here for temp file, as analyze_pdf handles its own if it creates one
+    # and Gradio handles the temp file wrapper it provides.
+
 
 def create_interface():
     with gr.Blocks(title="PDF Analyzer") as interface:
-        gr.Markdown("# PDF Analyzer")
-        gr.Markdown(
-            "Upload a PDF document to analyze its structure, references, language, and more. "
-            "Language issues are derived from font-filtered text. "
-            "Regex issues and general document checks use unfiltered text from the original PDF. "
-            "All issue coordinates (if found) are mapped back to the original PDF."
-        )
-        
         with gr.Row():
             file_input = gr.File(
                 label="Upload PDF",
-                file_types=[".pdf"]
-                # Default type: Gradio provides a path-like object (e.g., NamedString)
-                # or a TemporaryFileWrapper. Both should be acceptable by analyze_pdf.
+                file_types=[".pdf"],
+                # type="binary" # Gradio's File component returns a TemporaryFileWrapper which is file-like
+
             )
         
         with gr.Row():
-            analyze_btn = gr.Button("Analyze PDF")
         
         with gr.Row():
             results_output = gr.JSON(
-                label="Analysis Results",
+                label="Analysis Results", # Simplified label
                 show_label=True
             )
         
-        with gr.Row():
             pdf_output = gr.File( 
                 label="Annotated PDF (Placeholder - View Coordinates in JSON)",
                 show_label=True,
-                interactive=False 
+                interactive=False # Not interactive as it's a placeholder
             )
         
         analyze_btn.click(
-            fn=process_upload,
-            inputs=[file_input],
-            outputs=[results_output, pdf_output]
-        )
-    return interface
-
-if __name__ == "__main__":
     print("\n--- Launching Gradio Interface ---")
     # config.set_java_home() is called when config.py is imported.
 
+    # Optional: Test LanguageTool initialization
     try:
+        import language_tool_python # Import here for the test
         lt_test = language_tool_python.LanguageTool('en-US')
         lt_test.close()
         print("App: LanguageTool initialized successfully for test.")
     except Exception as lt_e:
-        print(f"App: Warning: Could not initialize LanguageTool for test. Language checks might fail: {lt_e}\n{traceback.format_exc(limit=1)}")
+        print(f"App: Warning: Could not initialize LanguageTool for test. Language checks might fail: {lt_e}")
         print("Please ensure Java is installed and JAVA_HOME is correctly set (see config.py).")
     
     app_interface = create_interface()
     app_interface.launch(
-        share=False,
+        share=False, # Set to True for public link if ngrok is installed and desired
+        # server_port=7860 # Optionally specify a port
     )