add in title info

app.py

@@ -204,12 +204,15 @@ class RAGChatbot:
             try:
                 results = self.ingestion.search(query, k=num_results)
                 if results:
-                    # Build context from retrieved chunks
+                    # Build context from retrieved chunks; include source/title so the model can cite it
                     context_parts = []
                     for i, result in enumerate(results, 1):
                         text = result['text'].strip()
-                        if text:
-                            context_parts.append(f"[Context {i}]\n{text}")
+                        if not text:
+                            continue
+                        meta = result.get('metadata') or {}
+                        source_label = meta.get('document_title') or meta.get('source') or f"Source {i}"
+                        context_parts.append(f"[Context {i}] (Source: {source_label})\n{text}")
                     
                     context = "\n\n".join(context_parts)
                     
@@ -229,8 +232,8 @@ Do not reveal your internal reasoning. Provide only the final answer.
 
 Structure your answer in the following format:
 Summary — A brief, high‑level answer.
-Supporting Details — Explain using information only from the provided context.
-Context References — Quote or cite the exact context segments that support your answer (if applicable). Context reference should include the title of the document referred to.
+Supporting Details — Explain using information only from the provided context. When citing, use the Source label shown for that context (e.g. the document title or name in parentheses after [Context N]).
+Context References — List each reference with the exact Source shown for that context (e.g. "CAN/CGSB-32.312-2018" or the document title). Include section name or page when that information appears in the context text. Format: document/source, section or location if available, and a short quote or paraphrase. Do not use only "Context 1" or "Context 5" as the reference; always include the document title/source.
 
 Context:
 {context}
@@ -260,7 +263,8 @@ Answer:"""
                         response_parts = []
                         response_parts.append("I retrieved relevant information, but couldn't generate a synthesized answer. Here are the relevant chunks:\n\n")
                         for i, result in enumerate(results, 1):
-                            source = result['metadata']['source']
+                            meta = result.get('metadata') or {}
+                            source = meta.get('document_title') or meta.get('source', '')
                             text = result['text'].strip()
                             if text:
                                 response_parts.append(f"**Relevant information {i}** (from {source}):\n{text}\n")

ingestion.py

@@ -34,6 +34,30 @@ class DocumentIngestion:
         self.embeddings = None
         self.index = None
         
+    def get_pdf_document_title(self, file_path: str) -> str:
+        """
+        Get a human-readable document title for a PDF (from metadata or filename).
+        
+        Args:
+            file_path: Path to the PDF file
+            
+        Returns:
+            Document title (e.g. standard name or filename without extension)
+        """
+        try:
+            reader = PdfReader(file_path)
+            if reader.metadata and getattr(reader.metadata, "title", None):
+                title = reader.metadata.title
+                if title and title.strip():
+                    return title.strip()
+        except Exception:
+            pass
+        # Fallback: filename without extension, cleaned for standards (e.g. CAN-CGSB-32.312 -> CAN/CGSB-32.312)
+        base = os.path.splitext(os.path.basename(file_path))[0]
+        if base:
+            return base.replace("-", "/") if "CGSB" in base or "CAN" in base else base
+        return file_path
+
     def read_pdf(self, file_path: str) -> str:
         """
         Extract text from a PDF file.
@@ -108,14 +132,14 @@ class DocumentIngestion:
                 if not os.path.exists(pdf_path):
                     print(f"Warning: PDF file not found: {pdf_path}")
                     continue
-                    
+                document_title = self.get_pdf_document_title(pdf_path)
                 text = self.read_pdf(pdf_path)
                 chunks = self.text_splitter.split_text(text)
-                
                 for i, chunk in enumerate(chunks):
                     all_texts.append(chunk)
                     all_metadata.append({
                         'source': pdf_path,
+                        'document_title': document_title,
                         'type': 'pdf',
                         'chunk_index': i
                     })
@@ -126,11 +150,20 @@ class DocumentIngestion:
                 try:
                     text = self.read_url(url)
                     chunks = self.text_splitter.split_text(text)
-                    
+                    # Use a short label for URL (domain + path hint) as document_title
+                    try:
+                        from urllib.parse import urlparse
+                        parsed = urlparse(url)
+                        document_title = parsed.netloc or url
+                        if parsed.path and parsed.path != "/":
+                            document_title += " " + parsed.path.strip("/")[:50]
+                    except Exception:
+                        document_title = url
                     for i, chunk in enumerate(chunks):
                         all_texts.append(chunk)
                         all_metadata.append({
                             'source': url,
+                            'document_title': document_title,
                             'type': 'url',
                             'chunk_index': i
                         })