updoot

app_hf.py

@@ -57,19 +57,34 @@ def handle_query():
         # Add web search results to sources
         for i, web_result in enumerate(web_search_results, len(sources_for_response) + 1):
             urls = web_result.get('urls', [])
-            sources_for_response.append({
+            # Ensure URLs is a list (not None or empty)
+            if not urls:
+                urls = []
+            
+            source_entry = {
                 "id": i,
                 "score": 0.0,  # Web search results don't have similarity scores
-                "url": urls[0] if urls else None,  # Primary URL for easy access
-                "urls": urls,  # All URLs at top level for easy access
                 "metadata": {
                     "source": web_result.get('source', 'Web Search'),
                     "type": web_result.get('type', 'web_search'),
                     "query": web_result.get('query', 'Unknown'),
                     "content": web_result.get('content', '')[:500],  # Preview
-                    "urls": urls  # Also in metadata for consistency
                 }
-            })
+            }
+            
+            # Add URLs at multiple levels for maximum compatibility
+            if urls:
+                source_entry["url"] = urls[0]  # Primary URL for easy access
+                source_entry["urls"] = urls  # All URLs at top level
+                source_entry["metadata"]["url"] = urls[0]  # Primary URL in metadata
+                source_entry["metadata"]["urls"] = urls  # All URLs in metadata
+            else:
+                source_entry["url"] = None
+                source_entry["urls"] = []
+                source_entry["metadata"]["url"] = None
+                source_entry["metadata"]["urls"] = []
+            
+            sources_for_response.append(source_entry)
 
         return jsonify({
             "response": response_text,

queryrun.py

@@ -566,8 +566,8 @@ Guidelines:
                 formatted_results.append(f"Source {i}: {doc['content'][:500]}...")
             return "\n".join(formatted_results)
         
-        def get_context_summary() -> str:
-            """Get a summary of available context documents"""
+        def get_context_summary(_=None) -> str:
+            """Get a summary of available context documents. Takes no arguments (ignores any provided)."""
             if not context_docs:
                 return "No context documents available."
             summary = f"Found {len(context_docs)} relevant documents:\n"
@@ -586,31 +586,55 @@ Guidelines:
                     # Get search results with URLs
                     ddg_results = list(ddgs.text(search_query, max_results=5))
                     
+                    # Debug: Print first result structure to understand format
+                    if ddg_results and len(ddg_results) > 0:
+                        print(f"🔍 [DEBUG] First result keys: {list(ddg_results[0].keys()) if isinstance(ddg_results[0], dict) else 'Not a dict'}")
+                    
                     # Format results
                     formatted_results = []
                     urls = []
                     for r in ddg_results:
+                        if not isinstance(r, dict):
+                            continue
                         title = r.get('title', '')
                         body = r.get('body', '')
-                        href = r.get('href', '')
+                        # Try different possible field names for URL
+                        href = r.get('href', '') or r.get('url', '') or r.get('link', '') or r.get('href', '')
                         if title or body:
                             formatted_results.append(f"{title}: {body}")
                         if href:
                             urls.append(href)
+                        # Also check if the result itself is a dict with URL info
+                        if not href:
+                            # Try to find any URL-like value in all fields
+                            for key, value in r.items():
+                                if isinstance(value, str) and ('http://' in value or 'https://' in value):
+                                    if value not in urls:
+                                        urls.append(value)
+                                        print(f"🔍 [DEBUG] Found URL in field '{key}': {value}")
+                                        break
                     
                     results = "\n".join(formatted_results) if formatted_results else "No results found."
                     
                 print(f"✅ [AGENT] Web search completed. Found {len(ddg_results)} results with {len(urls)} URLs.")
+                if urls:
+                    print(f"🔍 [DEBUG] URLs found: {urls[:3]}")  # Show first 3 URLs for debugging
+                else:
+                    print(f"⚠️ [DEBUG] No URLs found in search results!")
                 
                 # Store search results with metadata for sources
+                # Always include urls field, even if empty
                 web_search_results.append({
                     'query': search_query,
                     'content': results,
                     'type': 'web_search',
                     'source': f'Web Search: {search_query}',
-                    'urls': urls[:5] if urls else []  # Store up to 5 URLs
+                    'urls': urls[:5] if urls else []  # Store up to 5 URLs, empty list if none
                 })
                 
+                # Debug: Verify URLs are stored
+                print(f"🔍 [DEBUG] Stored web_search_result with {len(web_search_results[-1].get('urls', []))} URLs")
+                
                 # Format results with URLs if found
                 if urls:
                     url_list = "\n".join([f"  - {url}" for url in urls[:5]])
@@ -642,7 +666,7 @@ Guidelines:
             Tool(
                 name="get_context_summary",
                 func=get_context_summary,
-                description="Get a summary of the currently available context documents. Use this to understand what information is available."
+                description="Get a summary of the currently available context documents. Takes no arguments. Use this to understand what information is available."
             )
         ]