Update app.py

app.py

@@ -75,32 +75,35 @@ def set_custom_css():
     """, unsafe_allow_html=True)
 
 def get_docparser_data(file, api_key, parser_id) -> Optional[dict]:
-    # First, upload the document
-    upload_url = "https://api.docparser.com/v1/document/upload"
-    headers = {
-        'Authorization': f'Basic {api_key}'
-    }
-    files = {
-        'file': (file.name, file, 'application/pdf')
-    }
-    params = {
-        'parser_id': parser_id
-    }
-    
     try:
+        # First, upload the document
+        upload_url = f"https://api.docparser.com/v1/document/upload/{parser_id}"
+        
+        # Create proper headers with base64 encoded API key
+        import base64
+        auth_string = base64.b64encode(f"{api_key}:".encode()).decode()
+        headers = {
+            'Authorization': f'Basic {auth_string}',
+            'Content-Type': 'multipart/form-data'
+        }
+        
+        # Prepare the file for upload
+        files = {
+            'file': (file.name, file, 'application/pdf')
+        }
+        
         # Upload document
         upload_response = requests.post(
-            upload_url, 
-            headers=headers,
-            files=files,
-            params=params
+            upload_url,
+            headers={'Authorization': f'Basic {auth_string}'},
+            files=files
         )
         upload_response.raise_for_status()
         
         # Get document ID from upload response
         upload_data = upload_response.json()
         if not isinstance(upload_data, list) or len(upload_data) == 0:
-            st.error("Invalid response from Docparser upload")
+            st.error(f"Invalid response from Docparser upload: {upload_data}")
             return None
             
         document_id = upload_data[0].get('id')
@@ -108,26 +111,51 @@ def get_docparser_data(file, api_key, parser_id) -> Optional[dict]:
             st.error("Failed to get document ID from upload response")
             return None
 
+        # Wait a moment for processing
+        import time
+        time.sleep(2)  # Give Docparser time to process the document
+
         # Get parsed results
         results_url = f"https://api.docparser.com/v1/results/{parser_id}/{document_id}"
         results_response = requests.get(
             results_url,
-            headers=headers
+            headers={'Authorization': f'Basic {auth_string}'}
         )
         results_response.raise_for_status()
         
         # Handle results
         results_data = results_response.json()
+        
+        # Debug information
+        st.write("Debug - API Response:", results_data)
+        
         if isinstance(results_data, list) and len(results_data) > 0:
-            return results_data[0]
+            # Extract the relevant fields based on your Docparser parser configuration
+            parsed_data = {
+                'name': results_data[0].get('full_name', 'Unknown'),
+                'email': results_data[0].get('email', 'Unknown'),
+                'phone': results_data[0].get('phone', 'Unknown'),
+                'skills': [skill.strip() for skill in results_data[0].get('skills', '').split(',') if skill.strip()],
+                'certifications': results_data[0].get('certifications', []),
+                'experience_years': float(results_data[0].get('experience_years', 0)),
+                'degree': results_data[0].get('degree', 'Not specified'),
+                'institution': results_data[0].get('institution', 'Not specified'),
+                'year': results_data[0].get('graduation_year', 'Not specified'),
+                'summary': results_data[0].get('summary', 'No summary available'),
+                'projects': results_data[0].get('projects', [])
+            }
+            return parsed_data
         else:
-            st.error("No parsed data received from Docparser")
+            st.error(f"No parsed data received from Docparser: {results_data}")
             return None
 
     except requests.exceptions.HTTPError as http_err:
         st.error(f"HTTP error occurred: {http_err}")
-        if http_err.response is not None:
+        if hasattr(http_err, 'response') and http_err.response is not None:
             st.error(f"Response content: {http_err.response.content}")
+    except json.JSONDecodeError as json_err:
+        st.error(f"JSON decode error: {json_err}")
+        st.error("Raw response content: " + str(upload_response.content if 'upload_response' in locals() else 'No response'))
     except Exception as e:
         st.error(f"Error fetching data from Docparser: {e}")
     return None