Update app.py

app.py

@@ -89,14 +89,17 @@ def clean_source_name(book_name: str) -> str:
     return name or "Professor Handouts"
 
 
+def source_priority_label(book_name: str) -> str:
+    return "Primary source" if clean_source_name(book_name) == "Professor Handouts" else "Supporting textbook"
+
+
 def prioritize_professor_handouts(records):
-    """
-    Professor Handouts are always shown and used first.
-    Other textbooks are used only as supporting material.
-    """
     return sorted(
         records,
-        key=lambda r: 0 if clean_source_name(r.get("book", "")) == "Professor Handouts" else 1
+        key=lambda r: (
+            0 if clean_source_name(r.get("book", "")) == "Professor Handouts" else 1,
+            -float(r.get("final_score", r.get("similarity_score", 0)))
+        )
     )
 
 
@@ -107,21 +110,10 @@ def is_general_chat(text: str) -> bool:
     t = text.lower().strip()
 
     general_phrases = [
-        "hi",
-        "hello",
-        "hola",
-        "hey",
-        "good morning",
-        "good afternoon",
-        "good evening",
-        "thanks",
-        "thank you",
-        "gracias",
-        "ok",
-        "okay",
-        "who are you",
-        "what can you do",
-        "help"
+        "hi", "hello", "hola", "hey",
+        "good morning", "good afternoon", "good evening",
+        "thanks", "thank you", "gracias",
+        "ok", "okay", "who are you", "what can you do", "help"
     ]
 
     return t in general_phrases
@@ -158,9 +150,9 @@ def general_chat_reply(text: str, language_mode: str) -> str:
 
 
 # =====================================================
-# RETRIEVAL WITH SIMILARITY SCORES
+# RETRIEVAL WITH PROFESSOR HANDOUT BOOST
 # =====================================================
-def search_hybrid(query: str, shortlist_k: int = 20, final_k: int = 4):
+def search_hybrid(query: str, shortlist_k: int = 30, final_k: int = 5):
     ensure_loaded()
 
     q_tokens = tokenize(query)
@@ -172,18 +164,28 @@ def search_hybrid(query: str, shortlist_k: int = 20, final_k: int = 4):
     qvec = EMBED_MODEL.encode([query], normalize_embeddings=True).astype("float32")[0]
     dense_scores = shortlist_emb @ qvec
 
-    rerank = np.argsort(dense_scores)[::-1][:final_k]
-    final_idx = shortlist_idx[rerank]
-    final_scores = dense_scores[rerank]
-
     results = []
 
-    for idx, score in zip(final_idx, final_scores):
+    for idx, score in zip(shortlist_idx, dense_scores):
         record = CHUNKS[int(idx)].copy()
+        clean_book = clean_source_name(record.get("book", ""))
+
+        priority_boost = 0.15 if clean_book == "Professor Handouts" else 0.0
+        final_score = float(score) + priority_boost
+
         record["similarity_score"] = float(score)
+        record["final_score"] = final_score
+        record["source_priority"] = (
+            "Professor Handouts"
+            if clean_book == "Professor Handouts"
+            else "Supporting textbook"
+        )
+
         results.append(record)
 
-    return prioritize_professor_handouts(results)
+    results = sorted(results, key=lambda r: r["final_score"], reverse=True)
+
+    return prioritize_professor_handouts(results[:final_k])
 
 
 def build_context(records):
@@ -197,10 +199,11 @@ def build_context(records):
         blocks.append(
             f"""[Source {i}]
 Book: {clean_book}
-Source priority: {"Primary Professor Handouts" if clean_book == "Professor Handouts" else "Supporting textbook"}
+Source priority: {source_priority_label(clean_book)}
 Section: {r.get('section_title','')}
 Pages: {r.get('page_start','')}-{r.get('page_end','')}
 Similarity Score: {r.get('similarity_score', 0):.3f}
+Final Score: {r.get('final_score', r.get('similarity_score', 0)):.3f}
 Text:
 {r.get('text','')}"""
         )
@@ -232,7 +235,7 @@ def make_sources(records):
         section = r.get("section_title", "Course Material")
         page_start = r.get("page_start", "")
         page_end = r.get("page_end", "")
-        score = r.get("similarity_score", 0)
+        score = r.get("final_score", r.get("similarity_score", 0))
 
         if page_start and page_end and page_start != page_end:
             page_text = f"pages {page_start}-{page_end}"
@@ -241,13 +244,10 @@ def make_sources(records):
         else:
             page_text = "page not specified"
 
-        if clean_book == "Professor Handouts":
-            source_type = "Primary source"
-        else:
-            source_type = "Supporting textbook"
+        source_type = source_priority_label(clean_book)
 
         lines.append(
-            f"• {clean_book} ({source_type}) | {section} | {page_text} | similarity: {score:.2f}"
+            f"• {clean_book} ({source_type}) | {section} | {page_text} | relevance: {score:.2f}"
         )
 
     return "\n".join(lines)
@@ -275,36 +275,49 @@ def compute_confidence(records, answer: str):
             "score": 0.0,
         }
 
-    scores = [float(r.get("similarity_score", 0)) for r in records]
-
-    if not scores:
+    if not records:
         return {
             "level": "red",
             "label": "Not found",
             "score": 0.0,
         }
 
+    scores = [float(r.get("final_score", r.get("similarity_score", 0))) for r in records]
+    raw_scores = [float(r.get("similarity_score", 0)) for r in records]
+
     top_score = max(scores)
-    avg_score = sum(scores) / len(scores)
+    top_raw = max(raw_scores)
+
+    professor_found = any(
+        clean_source_name(r.get("book", "")) == "Professor Handouts"
+        for r in records[:3]
+    )
+
+    if professor_found and top_score >= 0.48:
+        return {
+            "level": "green",
+            "label": "High confidence",
+            "score": top_raw,
+        }
 
-    if top_score >= 0.55 and avg_score >= 0.38:
+    if top_score >= 0.52:
         return {
             "level": "green",
             "label": "High confidence",
-            "score": top_score,
+            "score": top_raw,
         }
 
-    if top_score >= 0.38:
+    if top_score >= 0.35:
         return {
             "level": "orange",
             "label": "Medium confidence",
-            "score": top_score,
+            "score": top_raw,
         }
 
     return {
         "level": "red",
         "label": "Low confidence",
-        "score": top_score,
+        "score": top_raw,
     }
 
 
@@ -739,9 +752,7 @@ def respond(user_msg, history, mode, language_mode, quiz_count_mode, show_source
     try:
         history = history + [{"role": "user", "content": text}]
 
-        # ---------------------------------------------
         # General greeting / normal conversation
-        # ---------------------------------------------
         if is_general_chat(text):
             reply = general_chat_reply(text, language_mode)
 
@@ -770,9 +781,7 @@ def respond(user_msg, history, mode, language_mode, quiz_count_mode, show_source
 
             return "", history, render_chat(history), quiz_state, render_dashboard()
 
-        # ---------------------------------------------
         # Quiz evaluation mode
-        # ---------------------------------------------
         if quiz_state.get("active", False):
             evaluation = oai_json(
                 build_quiz_eval_prompt(
@@ -837,16 +846,11 @@ def respond(user_msg, history, mode, language_mode, quiz_count_mode, show_source
 
             return "", history, render_chat(history), quiz_state, render_dashboard()
 
-        # ---------------------------------------------
         # Retrieval
-        # ---------------------------------------------
-        records = search_hybrid(text, shortlist_k=20, final_k=4)
-        records = prioritize_professor_handouts(records)
+        records = search_hybrid(text, shortlist_k=30, final_k=5)
         context = build_context(records)
 
-        # ---------------------------------------------
         # Quiz generation
-        # ---------------------------------------------
         if mode == "Quiz Me":
             n_questions = choose_quiz_count(text, quiz_count_mode)
 
@@ -899,9 +903,7 @@ def respond(user_msg, history, mode, language_mode, quiz_count_mode, show_source
 
             return "", history, render_chat(history), quiz_state, render_dashboard()
 
-        # ---------------------------------------------
         # Normal answer
-        # ---------------------------------------------
         answer = oai_text(
             build_tutor_prompt(
                 mode,
@@ -972,18 +974,12 @@ def clear_all():
 CSS = """
 :root{
   --page-bg: #d9d9dd;
-  --panel-bg: #555765;
-  --chat-bg: #4a4c59;
-  --grad-top: #e8c7d4;
-  --grad-mid: #a55ca2;
-  --grad-bot: #5a2d77;
-  --accent: #f4eb4b;
-  --accent-soft: #f5ef9a;
-  --user-bubble: #ffffff;
-  --bot-bubble: #f5efad;
+  --uva-purple: #5a2d77;
+  --uva-purple-light: #7b3f98;
+  --uva-gold: #c7a008;
+  --uva-gold-light: #fff8cc;
+  --uva-soft-purple: #efe7f6;
   --text-dark: #241336;
-  --text-dark-strong: #170c25;
-  --text-light: #ffffff;
   --shadow: rgba(30,20,50,0.18);
 }
 
@@ -1037,13 +1033,15 @@ footer{
   border-left: 5px solid #c7a008;
 }
 
+/* CHAT WINDOW - same family as analytics dashboard */
 .bc-phone{
   position: relative;
-  background: linear-gradient(180deg, var(--grad-top) 0%, var(--grad-mid) 48%, var(--grad-bot) 100%);
+  background: #ffffff;
   border-radius: 30px;
   padding: 92px 14px 14px 14px;
-  box-shadow: 0 16px 34px var(--shadow);
+  box-shadow: 0 16px 34px rgba(0,0,0,0.22);
   min-height: 620px;
+  border-top: 8px solid #5a2d77;
 }
 
 .bc-logo-holder{
@@ -1054,7 +1052,7 @@ footer{
   width: 104px;
   height: 104px;
   border-radius: 999px;
-  background: var(--accent);
+  background: #c7a008;
   display: flex;
   align-items: center;
   justify-content: center;
@@ -1078,17 +1076,17 @@ footer{
   text-align:center;
   font-size: 13px;
   font-weight: 900;
-  color: var(--text-dark-strong);
-  background: rgba(255,255,255,0.40);
+  color: #241336;
+  background: rgba(255,255,255,0.55);
   line-height: 1.05;
 }
 
 .bc-chat-shell{
-  background: rgba(74,76,89,0.92);
+  background: #ffffff;
   border-radius: 20px;
   padding: 16px;
   min-height: 460px;
-  box-shadow: inset 0 1px 0 rgba(255,255,255,0.06);
+  box-shadow: inset 0 0 0 2px #e5d8ef;
 }
 
 .bc-chat-wrap{
@@ -1105,7 +1103,7 @@ footer{
 }
 
 .bc-chat-wrap::-webkit-scrollbar-thumb{
-  background: rgba(255,255,255,0.28);
+  background: #c7a008;
   border-radius: 999px;
 }
 
@@ -1134,19 +1132,21 @@ footer{
 }
 
 .bc-user-bubble{
-  background: var(--user-bubble);
-  color: var(--text-dark-strong) !important;
+  background: #efe7f6;
+  color: #241336 !important;
+  border: 2px solid #d8c6e8;
   border-bottom-left-radius: 8px;
 }
 
 .bc-bot-bubble{
-  background: var(--bot-bubble);
-  color: var(--text-dark-strong) !important;
+  background: #fff8cc;
+  color: #241336 !important;
+  border: 2px solid #c7a008;
   border-bottom-right-radius: 8px;
 }
 
 .bc-bubble strong{
-  color: var(--text-dark-strong) !important;
+  color: #241336 !important;
 }
 
 .bc-confidence{
@@ -1155,10 +1155,11 @@ footer{
   gap:8px;
   margin-bottom:10px;
   padding:7px 10px;
-  background:rgba(255,255,255,0.65);
+  background:rgba(255,255,255,0.75);
   border-radius:999px;
   font-size:13px;
   color:#111827;
+  border:1px solid #e5d8ef;
 }
 
 .bc-dot{
@@ -1177,16 +1178,17 @@ footer{
 }
 
 .bc-empty-text{
-  color: white;
+  color: #5a2d77;
   text-align:center;
   opacity: 0.96;
   font-size: 16px;
   line-height: 1.6;
+  font-weight:600;
 }
 
 .bc-input-bar{
   margin-top: 12px;
-  background: var(--accent);
+  background: #5a2d77;
   border-radius: 999px;
   padding: 8px 10px;
   display:flex;
@@ -1199,22 +1201,22 @@ footer{
   width: 38px;
   height: 38px;
   border-radius: 999px;
-  background: rgba(255,255,255,0.34);
+  background: #c7a008;
   display:flex;
   align-items:center;
   justify-content:center;
   font-size: 30px;
   font-weight: 900;
-  color: var(--text-dark-strong);
+  color: #ffffff;
   user-select:none;
 }
 
 #bc_msg textarea{
-  background: rgba(255,255,255,0.42) !important;
-  border: none !important;
+  background: #ffffff !important;
+  border: 2px solid #c7a008 !important;
   box-shadow: none !important;
   border-radius: 999px !important;
-  color: var(--text-dark-strong) !important;
+  color: #241336 !important;
   padding: 11px 14px !important;
   min-height: 42px !important;
 }
@@ -1228,15 +1230,15 @@ footer{
   height: 42px !important;
   border-radius: 999px !important;
   border: none !important;
-  background: rgba(255,255,255,0.34) !important;
-  color: var(--text-dark-strong) !important;
+  background: #c7a008 !important;
+  color: #ffffff !important;
   font-size: 20px !important;
   font-weight: 900 !important;
   box-shadow: none !important;
 }
 
 #bc_send button:hover{
-  background: rgba(255,255,255,0.52) !important;
+  background: #9f8006 !important;
 }
 
 #bc_clear button, #bc_refresh button, #bc_clear_analytics button{