Update app.py

app.py

@@ -8,35 +8,37 @@ from PIL import Image
 import PyPDF2
 from io import BytesIO
 import docx
+import ntpath
 
 # Initialize clients
 API_KEY = os.environ.get("HF_API_KEY")
 client = InferenceClient(token=API_KEY)
 
-def process_file(file):
-    """Handle different file types and extract text"""
-    if file is None:
+def process_file(filepath):
+    """Handle different file types and extract text."""
+    if not filepath:
         return ""
-        
+    
     # Get file extension
-    ext = file.name.split('.')[-1].lower()
+    # e.g., if filepath = "/somefolder/report.pdf", ext = ".pdf"
+    ext = os.path.splitext(filepath)[1].lower()
     
     try:
-        if ext in ['png', 'jpg', 'jpeg']:
+        if ext in ['.png', '.jpg', '.jpeg']:
             # OCR processing for images
-            image = Image.open(file.name)
+            image = Image.open(filepath)
             text = pytesseract.image_to_string(image)
             return f"IMAGE CONTENT:\n{text}"
             
-        elif ext == 'pdf':
+        elif ext == '.pdf':
             # PDF text extraction
-            pdf_reader = PyPDF2.PdfReader(file.name)
+            pdf_reader = PyPDF2.PdfReader(filepath)
             text = "\n".join([page.extract_text() for page in pdf_reader.pages])
             return f"PDF CONTENT:\n{text}"
             
-        elif ext == 'docx':
+        elif ext == '.docx':
             # Word document processing
-            doc = docx.Document(file.name)
+            doc = docx.Document(filepath)
             text = "\n".join([para.text for para in doc.paragraphs])
             return f"DOCUMENT CONTENT:\n{text}"
             
@@ -47,9 +49,9 @@ def process_file(file):
         print(f"File processing error: {e}")
         return "Error reading file"
 
-def chat(message, history, file):
+def chat(message, history, filepath):
     # Process uploaded file
-    file_content = process_file(file) if file else ""
+    file_content = process_file(filepath) if filepath else ""
     
     # Build enhanced prompt
     full_prompt = f"""
@@ -57,7 +59,8 @@ def chat(message, history, file):
     
     User Message: {message}
     
-    Please respond considering both the message and any attached documents:"""
+    Please respond considering both the message and any attached documents:
+    """
     
     # Configure generation parameters
     generate_kwargs = dict(
@@ -77,6 +80,7 @@ def chat(message, history, file):
     
     partial_message = ""
     for response in stream:
+        # Skip special tokens
         if response.token.special:
             continue
         partial_message += response.token.text
@@ -88,8 +92,10 @@ with gr.Blocks(theme="soft") as demo:
     gr.Markdown("Upload images, PDFs, or docs and chat about them!")
     
     with gr.Row():
-        file_input = gr.File(label="Upload File (PDF/Image/Doc)", type="file")
+        # Note the change: type="filepath"
+        file_input = gr.File(label="Upload File (PDF/Image/Doc)", type="filepath")
         
+    # The ChatInterface should pass the file path to `chat`
     chatbot = gr.ChatInterface(
         fn=chat,
         additional_inputs=[file_input],
@@ -101,4 +107,3 @@ with gr.Blocks(theme="soft") as demo:
 
 demo.launch()
 
-