Create app.py

app.py

@@ -0,0 +1,112 @@
+import gradio as gr
+import joblib
+import os
+import re
+import emoji
+import demoji
+import numpy as np
+
+# ==========================================================
+# 📦 Load all models
+# ==========================================================
+vectorizer_en = joblib.load("tfidf_vectorizer_en.pkl")
+le_en = joblib.load("label_encoder_en.pkl")
+stacking_en = joblib.load("stacking_en.pkl")
+
+vectorizer_fa = joblib.load("tfidf_vectorizer_fa.pkl")
+le_fa = joblib.load("label_encoder_fa.pkl")
+stacking_fa = joblib.load("stacking_fa.pkl")
+
+# ==========================================================
+# 🧹 Text cleaning functions
+# ==========================================================
+import nltk
+from nltk.corpus import stopwords
+from nltk.tokenize import word_tokenize
+from nltk.stem import WordNetLemmatizer
+from hazm import Normalizer, Lemmatizer as HazmLemmatizer, word_tokenize as hazm_tokenize
+nltk.download("punkt")
+nltk.download("stopwords")
+nltk.download("wordnet")
+
+# English preprocess
+lemmatizer = WordNetLemmatizer()
+STOPWORDS = set(stopwords.words("english"))
+RE_URL = re.compile(r"http\S+|www\.\S+")
+RE_HTML = re.compile(r"<.*?>")
+RE_NONALPHA = re.compile(r"[^a-zA-Z\s]")
+
+def preprocess_english(text):
+    text = str(text).lower()
+    text = emoji.demojize(text)
+    text = demoji.replace(text, "")
+    text = RE_URL.sub(" ", text)
+    text = RE_HTML.sub(" ", text)
+    text = RE_NONALPHA.sub(" ", text)
+    text = re.sub(r"\s+", " ", text).strip()
+    tokens = word_tokenize(text)
+    tokens = [lemmatizer.lemmatize(t) for t in tokens if t not in STOPWORDS and len(t) > 2]
+    return " ".join(tokens)
+
+# Persian preprocess
+normalizer = Normalizer()
+hazm_lemmatizer = HazmLemmatizer()
+RE_URL_FA = re.compile(r"http\S+|www\.\S+")
+RE_NONPERSIAN = re.compile(r"[^\u0600-\u06FFA-Za-z\s]")
+
+def preprocess_persian(text):
+    text = str(text)
+    text = normalizer.normalize(text)
+    text = emoji.demojize(text)
+    text = demoji.replace(text, "")
+    text = RE_URL_FA.sub(" ", text)
+    text = re.sub(r"@\w+|#\w+|\d+", " ", text)
+    text = RE_NONPERSIAN.sub(" ", text)
+    text = re.sub(r"\s+", " ", text).strip()
+    tokens = hazm_tokenize(text)
+    tokens = [hazm_lemmatizer.lemmatize(t) for t in tokens if len(t) > 1]
+    return " ".join(tokens)
+
+# ==========================================================
+# 🔮 Prediction function
+# ==========================================================
+def predict_sentiment(comment, language):
+    if language == "English":
+        clean_text = preprocess_english(comment)
+        X = vectorizer_en.transform([clean_text])
+        pred = stacking_en.predict(X)[0]
+        probs = stacking_en.predict_proba(X)[0]
+        classes = le_en.classes_
+    else:
+        clean_text = preprocess_persian(comment)
+        X = vectorizer_fa.transform([clean_text])
+        pred = stacking_fa.predict(X)[0]
+        probs = stacking_fa.predict_proba(X)[0]
+        classes = le_fa.classes_
+
+    result_str = f"🔹 **Predicted Sentiment:** {pred}\n\n"
+    prob_table = "\n".join([f"{cls}: {round(p,3)}" for cls, p in zip(classes, probs)])
+    return f"🗣️ **Input:** {comment}\n\n{result_str}**Prediction Probabilities:**\n{prob_table}"
+
+# ==========================================================
+# 🎨 Gradio UI
+# ==========================================================
+lang_dropdown = gr.Dropdown(["English", "Persian"], label="Select Language", value="English")
+input_box = gr.Textbox(label="Enter your comment here")
+output_box = gr.Markdown()
+
+iface = gr.Interface(
+    fn=predict_sentiment,
+    inputs=[input_box, lang_dropdown],
+    outputs=output_box,
+    title="🌍 Multilingual Sentiment Analyzer (English + Persian)",
+    description="Enter a comment in English or Persian to see the predicted sentiment and probabilities.",
+    examples=[
+        ["I loved the show! It was amazing!", "English"],
+        ["برنامه خیلی عالی بود و مجری هم خوب بود", "Persian"],
+        ["It was an average episode, not too bad.", "English"],
+    ]
+)
+
+if __name__ == "__main__":
+    iface.launch()