Update app.py

app.py

@@ -19,7 +19,7 @@ from sklearn.multiclass import OneVsRestClassifier
 from sklearn.pipeline import Pipeline
 from sklearn.metrics import f1_score
 
-# ---------------- Storage paths (robust local vs. HF Spaces) ----------------
+# ---------------- Storage paths ----------------
 def _pick_data_dir():
     if os.path.isdir("/data") and os.access("/data", os.W_OK):
         return "/data"
@@ -28,8 +28,8 @@ def _pick_data_dir():
 DATA_DIR = os.getenv("MM_DATA_DIR", _pick_data_dir())
 os.makedirs(DATA_DIR, exist_ok=True)
 DB_PATH = os.path.join(DATA_DIR, "moodmirror.db")
-MODEL_PATH = os.path.join(DATA_DIR, "goemo_sklearn.joblib")     # pipeline + mlb
-MODEL_VERSION = "v1-tfidf-lr-ovr"                               # bump if you change training
+MODEL_PATH = os.path.join(DATA_DIR, "goemo_sklearn.joblib")
+MODEL_VERSION = "v1-tfidf-lr-ovr"
 
 print(f"[MM] Using data dir: {DATA_DIR}")
 print(f"[MM] SQLite path:   {DB_PATH}")
@@ -48,7 +48,7 @@ CRISIS_NUMBERS = {
     "Other / Not listed": "Call your local emergency number (**112/911**) or search “suicide crisis hotline” + your country.",
 }
 
-# --- Deep, varied emotional advice & comfort phrases ---
+# ---------------- Advice & Quotes ----------------
 SUGGESTIONS = {
     "sadness": [
         "Be gentle with yourself. Cry if you need to — that’s healing, not weakness.",
@@ -162,66 +162,53 @@ SUGGESTIONS = {
     ],
 }
 
-# --- Inspirational / comforting quotes & affirmations ---
 QUOTES = {
     "sadness": [
         "“Even the darkest night will end and the sun will rise.” – Victor Hugo",
-        "“You don’t have to feel better to start healing.”",
-        "“It’s okay to be lost for a while.”",
-        "“Tears are words the heart can’t express.” – Paulo Coelho",
         "“You have survived every hard day so far.”",
+        "“You don’t have to feel better to start healing.”",
     ],
     "fear": [
         "“Feel the fear and do it anyway.” – Susan Jeffers",
-        "“Courage is not the absence of fear, but acting in spite of it.”",
-        "“You’ve faced hard things before — you can again.”",
         "“This moment will not last forever.”",
+        "“You’ve faced hard things before — you can again.”",
     ],
     "joy": [
         "“Happiness is not out there, it’s in you.”",
         "“Let joy be your rebellion.”",
         "“Enjoy the little things — one day you’ll realize they were the big things.”",
-        "“Joy shared is joy doubled.”",
     ],
     "anger": [
-        "“Speak when you are angry and you’ll make the best speech you’ll ever regret.” – Ambrose Bierce",
         "“Peace begins with a pause.”",
-        "“Anger is energy — learn to guide it, not suppress it.”",
+        "“Anger is energy — guide it, don’t suppress it.”",
     ],
     "boredom": [
         "“Boredom is the beginning of imagination.” – Jules Renard",
         "“Curiosity is the cure for boredom.” – Dorothy Parker",
-        "“The small things done repeatedly change everything.”",
     ],
     "grief": [
         "“Grief is love that has nowhere to go.”",
-        "“What we once enjoyed we can never lose; all that we love deeply becomes part of us.” – Helen Keller",
         "“Love doesn’t end, it changes form.”",
     ],
     "love": [
         "“Where there is love, there is life.” – Mahatma Gandhi",
         "“You are loved just for being who you are.” – Ram Dass",
-        "“Love quietly transforms everything it touches.”",
     ],
     "nervousness": [
-        "“You don’t have to control your thoughts; just stop letting them control you.” – Dan Millman",
         "“Breathe. You are doing enough.”",
         "“This worry does not define you.”",
     ],
     "curiosity": [
         "“Stay curious — it’s the mind’s way of loving life.”",
-        "“Wonder is wisdom’s beginning.” – Socrates",
         "“Every question plants a seed.”",
     ],
     "gratitude": [
         "“Gratitude turns what we have into enough.”",
-        "“The more grateful I am, the more beauty I see.” – Mary Davis",
         "“Thankfulness unlocks joy.”",
     ],
     "neutral": [
         "“Be present — even a calm moment can be a quiet victory.”",
         "“Peace is not the absence of chaos, but the presence of inner calm.”",
-        "“Slow is smooth, smooth is peaceful.”",
     ],
 }
 
@@ -233,216 +220,118 @@ COLOR_MAP = {
     "neutral": "#F5F5F5",
 }
 
-# Map GoEmotions label -> your UI buckets
 GOEMO_TO_APP = {
-    "admiration": "gratitude",
-    "amusement": "joy",
-    "anger": "anger",
-    "annoyance": "anger",
-    "approval": "gratitude",
-    "caring": "love",
-    "confusion": "nervousness",
-    "curiosity": "curiosity",
-    "desire": "joy",
-    "disappointment": "sadness",
-    "disapproval": "anger",
-    "disgust": "anger",
-    "embarrassment": "nervousness",
-    "excitement": "joy",
-    "fear": "fear",
-    "gratitude": "gratitude",
-    "grief": "grief",
-    "joy": "joy",
-    "love": "love",
-    "nervousness": "nervousness",
-    "optimism": "joy",
-    "pride": "joy",
-    "realization": "neutral",
-    "relief": "gratitude",
-    "remorse": "grief",
-    "sadness": "sadness",
-    "surprise": "neutral",
-    "neutral": "neutral",
+    "admiration": "gratitude", "amusement": "joy", "anger": "anger", "annoyance": "anger",
+    "approval": "gratitude", "caring": "love", "confusion": "nervousness",
+    "curiosity": "curiosity", "desire": "joy", "disappointment": "sadness",
+    "disapproval": "anger", "disgust": "anger", "embarrassment": "nervousness",
+    "excitement": "joy", "fear": "fear", "gratitude": "gratitude", "grief": "grief",
+    "joy": "joy", "love": "love", "nervousness": "nervousness", "optimism": "joy",
+    "pride": "joy", "realization": "neutral", "relief": "gratitude", "remorse": "grief",
+    "sadness": "sadness", "surprise": "neutral", "neutral": "neutral",
 }
 
-THRESHOLD = 0.30  # probability threshold for selecting labels
+THRESHOLD = 0.30
 
 # ---------------- SQLite helpers ----------------
 def get_conn():
     return sqlite3.connect(DB_PATH, check_same_thread=False, timeout=10)
 
 def init_db():
-    conn = None
-    try:
-        conn = get_conn()
-        c = conn.cursor()
-        c.execute("""
-        CREATE TABLE IF NOT EXISTS sessions(
-            id INTEGER PRIMARY KEY AUTOINCREMENT,
-            ts TEXT,
-            country TEXT,
-            user_text TEXT,
-            main_emotion TEXT
-        )
-        """)
-        conn.commit()
-    finally:
-        try:
-            if conn: conn.close()
-        except Exception:
-            pass
+    conn = get_conn()
+    c = conn.cursor()
+    c.execute("""
+    CREATE TABLE IF NOT EXISTS sessions(
+        id INTEGER PRIMARY KEY AUTOINCREMENT,
+        ts TEXT,
+        country TEXT,
+        user_text TEXT,
+        main_emotion TEXT
+    )
+    """)
+    conn.commit()
+    conn.close()
 
 def log_session(country, msg, emotion):
-    conn = None
-    try:
-        conn = get_conn()
-        c = conn.cursor()
-        c.execute(
-            "INSERT INTO sessions(ts, country, user_text, main_emotion) VALUES(?,?,?,?)",
-            (datetime.utcnow().isoformat(timespec="seconds"), country, msg[:500], emotion),
-        )
-        conn.commit()
-    finally:
-        try:
-            if conn: conn.close()
-        except Exception:
-            pass
-
-# ---------------- Train / Load model from DATASET ONLY ----------------
+    conn = get_conn()
+    c = conn.cursor()
+    c.execute("INSERT INTO sessions(ts, country, user_text, main_emotion) VALUES(?,?,?,?)",
+              (datetime.utcnow().isoformat(timespec="seconds"), country, msg[:500], emotion))
+    conn.commit()
+    conn.close()
+
+# ---------------- Train / Load model ----------------
 def load_goemotions_dataset():
-    # "simplified" gives 'text' and 'labels' as list[int] indices
     ds = load_dataset("google-research-datasets/go_emotions", "simplified")
     label_names = ds["train"].features["labels"].feature.names
     return ds, label_names
 
-def _prepare_xy(split):
-    # Each example has text and labels (list of ints)
-    X = split["text"]
-    y = split["labels"]  # list[list[int]]
-    return X, y
-
 def train_or_load_model():
-    # Try cache first
     if os.path.isfile(MODEL_PATH):
         print("[MM] Loading cached classifier...")
         bundle = joblib.load(MODEL_PATH)
         if bundle.get("version") == MODEL_VERSION:
             return bundle["pipeline"], bundle["mlb"], bundle["label_names"]
-        else:
-            print("[MM] Cached model version mismatch; retraining...")
 
     print("[MM] Loading GoEmotions dataset...")
     ds, label_names = load_goemotions_dataset()
 
-    print("[MM] Preparing data...")
-    X_train, y_train_idx = _prepare_xy(ds["train"])
-    X_val, y_val_idx = _prepare_xy(ds["validation"])
+    X_train = ds["train"]["text"]; y_train = ds["train"]["labels"]
+    X_val = ds["validation"]["text"]; y_val = ds["validation"]["labels"]
 
-    # MultiLabelBinarizer to convert list[int] -> multi-hot
     mlb = MultiLabelBinarizer(classes=list(range(len(label_names))))
-    Y_train = mlb.fit_transform(y_train_idx)
-    Y_val   = mlb.transform(y_val_idx)
-
-    # Build pipeline
-    clf = Pipeline(steps=[
-        ("tfidf", TfidfVectorizer(
-            lowercase=True,
-            ngram_range=(1,2),
-            min_df=2,
-            max_df=0.9,
-            strip_accents="unicode",
-        )),
-        ("ovr", OneVsRestClassifier(
-            LogisticRegression(
-                solver="saga",
-                max_iter=1000,
-                n_jobs=-1,
-                class_weight="balanced",
-            ),
-            n_jobs=-1
-        ))
+    Y_train = mlb.fit_transform(y_train)
+    Y_val = mlb.transform(y_val)
+
+    clf = Pipeline([
+        ("tfidf", TfidfVectorizer(lowercase=True, ngram_range=(1, 2), min_df=2, max_df=0.9, strip_accents="unicode")),
+        ("ovr", OneVsRestClassifier(LogisticRegression(solver="saga", max_iter=1000, n_jobs=-1, class_weight="balanced"), n_jobs=-1))
     ])
 
-    print("[MM] Training classifier (this happens once; cached afterward)...")
+    print("[MM] Training classifier...")
     clf.fit(X_train, Y_train)
 
-    # Quick validation metric (macro F1 over labels present in val)
-    Y_val_pred = clf.predict(X_val)
-    macro_f1 = f1_score(Y_val, Y_val_pred, average="macro", zero_division=0)
-    print(f"[MM] Validation macro F1: {macro_f1:.3f}")
-
-    # Cache model
-    joblib.dump({
-        "version": MODEL_VERSION,
-        "pipeline": clf,
-        "mlb": mlb,
-        "label_names": label_names
-    }, MODEL_PATH)
-    print(f"[MM] Saved classifier to {MODEL_PATH}")
+    print(f"[MM] Validation macro F1: {f1_score(Y_val, clf.predict(X_val), average='macro', zero_division=0):.3f}")
 
+    joblib.dump({"version": MODEL_VERSION, "pipeline": clf, "mlb": mlb, "label_names": label_names}, MODEL_PATH)
     return clf, mlb, label_names
 
-# Train/load at startup
 try:
     CLASSIFIER, MLB, LABEL_NAMES = train_or_load_model()
 except Exception as e:
     print(f"[WARN] Failed to train/load classifier: {e}")
     CLASSIFIER, MLB, LABEL_NAMES = None, None, None
 
-# ---------------- Inference using ONLY the trained classifier ----------------
+# ---------------- Emotion detection ----------------
 def classify_text(text: str):
-    """
-    Returns list of (label_name, prob) for labels above THRESHOLD, sorted desc.
-    """
-    if not CLASSIFIER or not MLB or not LABEL_NAMES:
-        return []
-
-    # predict_proba returns array shape (1, n_labels)
+    if not CLASSIFIER: return []
     try:
         proba = CLASSIFIER.predict_proba([text])[0]
     except AttributeError:
-        # If estimator doesn't support predict_proba (shouldn't happen with LR),
-        # fall back to decision_function -> sigmoid
         from scipy.special import expit
-        scores = CLASSIFIER.decision_function([text])[0]
-        proba = expit(scores)
-
+        proba = expit(CLASSIFIER.decision_function([text])[0])
     idxs = [i for i, p in enumerate(proba) if p >= THRESHOLD]
-    # Sort by probability desc
     idxs.sort(key=lambda i: proba[i], reverse=True)
     return [(LABEL_NAMES[i], float(proba[i])) for i in idxs]
 
 def detect_emotions(text: str):
     chosen = classify_text(text)
-    if not chosen:
-        return "neutral"
-    # Map to app buckets and take the strongest
+    if not chosen: return "neutral"
     bucket = {}
     for label, p in chosen:
         app = GOEMO_TO_APP.get(label.lower(), "neutral")
         bucket[app] = max(bucket.get(app, 0.0), p)
-    main = max(bucket, key=bucket.get) if bucket else "neutral"
-    return main
+    return max(bucket, key=bucket.get)
 
-# ---------------- Legacy-style reply composer (advice/quote/both) -----------
+# ---------------- Reply composer ----------------
 def compose_support_legacy(main_emotion: str, is_first_msg: bool) -> str:
-    tip = random.choice(SUGGESTIONS.get(
-        main_emotion,
-        ["Take a slow breath. One small act of kindness can shift your day."]
-    ))
-    quote = random.choice(QUOTES.get(
-        main_emotion,
-        ["“No matter what you feel right now, this moment will pass.”"]
-    ))
-
-    # 0 = advice only, 1 = quote only, 2 = both
-    mode = random.choice([0, 1, 2])
-    if mode == 0:
-        reply = tip
-    elif mode == 1:
-        reply = f"✨ {quote}"
-    else:
-        reply = f"{tip}\n\n💬 {quote}"
+    tip = random.choice(SUGGESTIONS.get(main_emotion, ["Take a slow breath. One small act of kindness can shift your day."]))
+    quote = random.choice(QUOTES.get(main_emotion, ["“No matter what you feel right now, this moment will pass.”"]))
+    include_quote = random.random() < 0.5
+
+    reply = tip
+    if include_quote:
+        reply += f"\n\n💬 {quote}"
 
     if is_first_msg:
         reply += "\n\n*Can you tell me a bit more about what’s behind that feeling?*"
@@ -452,88 +341,14 @@ def compose_support_legacy(main_emotion: str, is_first_msg: bool) -> str:
 # ---------------- Chat logic ----------------
 def crisis_block(country):
     msg = CRISIS_NUMBERS.get(country, CRISIS_NUMBERS["Other / Not listed"])
-    return (
-        "💛 I'm really sorry you're feeling like this. You matter.\n\n"
-        f"**If you might be in danger or thinking about harming yourself:**\n{msg}\n\n"
-        "Please reach out to someone now. You are not alone."
-    )
+    return ("💛 I'm really sorry you're feeling like this. You matter.\n\n"
+            f"**If you might be in danger or thinking about harming yourself:**\n{msg}\n\n"
+            "Please reach out to someone now. You are not alone.")
 
 def chat_step(message, history, country, save_session):
     if CRISIS_RE.search(message):
         return crisis_block(country), "#FFD6E7"
-
     if CLOSING_RE.search(message):
         return ("Thank you 💛 Take care of yourself. Small steps matter. 🌿", "#FFFFFF")
 
-    recent = " ".join(message.split()[-100:])
-    main = detect_emotions(recent)
-    color = COLOR_MAP.get(main, "#FFFFFF")
-
-    if save_session:
-        log_session(country, message, main)
-
-    reply = compose_support_legacy(main, is_first_msg=not bool(history))
-    return reply, color
-
-# ---------------- Gradio UI ----------------
-init_db()
-
-custom_css = """
-:root, body, .gradio-container { transition: background-color 0.8s ease !important; }
-.typing { font-style: italic; opacity: 0.8; animation: blink 1s infinite; }
-@keyframes blink { 50% {opacity: 0.4;} }
-"""
-
-with gr.Blocks(css=custom_css, title="🪞 MoodMirror+ (Dataset-only Edition)") as demo:
-    style_injector = gr.HTML("")
-    gr.Markdown(
-        "### 🪞 MoodMirror+ — Emotional Support & Inspiration 🌸\n"
-        "Powered only by the **GoEmotions dataset** (trained locally on startup).\n\n"
-        "_Not medical advice. If you feel unsafe, please reach out for help immediately._"
-    )
-
-    with gr.Row():
-        country = gr.Dropdown(choices=list(CRISIS_NUMBERS.keys()), value="Other / Not listed", label="Country")
-        save_ok = gr.Checkbox(value=False, label="Save anonymized session (no personal data)")
-
-    chat = gr.Chatbot(height=360)
-    msg = gr.Textbox(placeholder="Type how you feel...", label="Your message")
-    send = gr.Button("Send")
-    typing = gr.Markdown("", elem_classes="typing")
-
-    # Optional: dataset preview (for transparency)
-    with gr.Accordion("🔎 Preview GoEmotions samples", open=False):
-        with gr.Row():
-            n_examples = gr.Slider(1, 10, value=5, step=1, label="Number of examples")
-            split = gr.Dropdown(["train", "validation", "test"], value="train", label="Split")
-            refresh = gr.Button("Show samples")
-        table = gr.Dataframe(headers=["text", "labels"], row_count=5, wrap=True)
-
-        def refresh_samples(n, split_name):
-            try:
-                ds = load_dataset("google-research-datasets/go_emotions", "simplified")
-                names = ds["train"].features["labels"].feature.names
-                rows = ds[split_name].shuffle(seed=42).select(range(min(int(n), len(ds[split_name]))))
-                return [[t, ", ".join([names[i] for i in labs])] for t, labs in zip(rows["text"], rows["labels"])]
-            except Exception as e:
-                return [[f"Dataset load error: {e}", ""]]
-
-        refresh.click(refresh_samples, inputs=[n_examples, split], outputs=[table])
-
-    def respond(user_msg, chat_hist, country_choice, save_flag):
-        if not user_msg or not user_msg.strip():
-            yield chat_hist + [[user_msg, "Please share a short sentence about how you feel 🙂"]], "", "", ""
-            return
-        yield chat_hist, "💭 MoodMirror is thinking...", "", ""
-        reply, color = chat_step(user_msg, chat_hist, country_choice, bool(save_flag))
-        style_tag = f"<style>:root,body,.gradio-container{{background:{color}!important;}}</style>"
-        yield chat_hist + [[user_msg, reply]], "", style_tag, ""
-
-    send.click(respond, inputs=[msg, chat, country, save_ok],
-               outputs=[chat, typing, style_injector, msg], queue=True)
-    msg.submit(respond, inputs=[msg, chat, country, save_ok],
-               outputs=[chat, typing, style_injector, msg], queue=True)
-
-if __name__ == "__main__":
-    demo.queue()
-    demo.launch()
+    emotion = detect_emotions(" ".join