Update src/streamlit_app.py

src/streamlit_app.py

@@ -6,6 +6,13 @@ import re
 import streamlit as st
 import torch
 
+# ==========================================================
+# 🌐 Language Detection Imports
+# ==========================================================
+from fasttext_langdetect import detect_langs
+from langdetect import detect
+
+
 
 # ==========================================================
 # ✅ PAGE CONFIGS
@@ -221,11 +228,12 @@ else:
             text, toc, toc_source = extract_text_from_pdf(temp_path)
             chunks = chunk_text(text, chunk_size=chunk_size, overlap=overlap)
 
-            # 🔠 Detect document language (Hindi or English)
-            doc_sample = " ".join(chunks[:3])[:1000]
-            doc_lang = detect_language(doc_sample)
+            # 🌐 Detect document language (robust multilingual)
+            doc_sample = " ".join(chunks[:3])[:3000]
+            doc_lang = detect_document_language(doc_sample)
             st.session_state["doc_lang"] = doc_lang
-            st.caption(f"🈹 Detected document language: {'Hindi' if doc_lang == 'hi' else 'English'}")
+            lang_label = "Hindi" if doc_lang.startswith("hi") else "English"
+            st.caption(f"🈹 Detected document language: {lang_label}")
 
         with st.spinner("⚙️ Building search index..."):
             embeddings = cache_embeddings(os.path.basename(temp_path), chunks, embed_chunks)
@@ -236,12 +244,10 @@ else:
             query_suggestions = generate_dynamic_suggestions_from_toc(toc, chunks, doc_name)
             st.session_state["query_suggestions_fixed"] = query_suggestions
             st.session_state["last_doc"] = doc_name
-
-            # 🧹 Reset query when new document is loaded
             st.session_state["user_query_input"] = ""
             st.session_state["selected_suggestion"] = None
             st.session_state["show_more"] = False
-            st.experimental_rerun()
+            st.rerun()
         else:
             query_suggestions = st.session_state["query_suggestions_fixed"]
 
@@ -260,7 +266,7 @@ else:
             toggle_text = "Show less ▲" if st.session_state["show_more"] else "Show more ▼"
             if st.button(toggle_text, help="Show or hide more suggestions"):
                 st.session_state["show_more"] = not st.session_state["show_more"]
-                st.experimental_rerun()
+                st.rerun()
 
         user_query = st.text_input("Type your question or click one above:", key="user_query_input")
 
@@ -271,15 +277,14 @@ else:
             reasoning_mode = mode == "Extended (Document + general)"
             with st.spinner("💭 Generating your answer..."):
                 retrieved = retrieve_chunks(user_query, index, chunks, top_k=top_k, embeddings=embeddings)
-                doc_lang = st.session_state.get("doc_lang", "en")  # 🆕 Pass language info
+                doc_lang = st.session_state.get("doc_lang", "en")
+                print("🧠 Document language used for GPT prompt:", doc_lang)
                 answer = generate_answer(user_query, retrieved, reasoning_mode=reasoning_mode, doc_lang=doc_lang)
 
             st.markdown("### 🤖 Assistant’s Answer")
 
-            # ✅ Apply bullet formatting only for strict factual mode (not reasoning)
             if not reasoning_mode and not answer.startswith("⚠️"):
-                # Remove Markdown bold markers (**text**) and convert hyphen bullets
-                answer = re.sub(r"\*\*(.*?)\*\*", r"\1", answer)      # strip **bold**
+                answer = re.sub(r"\*\*(.*?)\*\*", r"\1", answer)
                 answer = re.sub(r"(^|\n)-\s*", r"\1<br>• ", answer)
             st.markdown(f"<div class='answer-box'>{answer}</div>", unsafe_allow_html=True)