Create app.py

app.py

@@ -0,0 +1,105 @@
+import os
+import pandas as pd
+import numpy as np
+import gradio as gr
+from openai import OpenAI
+
+# OpenRouter API client (API key ko Hugging Face Space ke secrets mein add karna hai)
+client = OpenAI(
+    base_url="https://openrouter.ai/api/v1",
+    api_key=os.getenv("OPENROUTER_API_KEY")  # HF Secret
+)
+
+# --- Sentiment Function ---
+def get_sentiment(review_text: str) -> str:
+    prompt = f"""
+Classify sentiment of this review in English or Roman Urdu.
+Return EXACTLY one word: positive, negative, or neutral.
+
+Rules (priority order, MUST follow strictly):
+
+1. If the review mentions problems with Priceoye's website, service, delivery,
+   shipping time, support, or Priceoye as a company/brand overall,
+   ALWAYS respond with "negative".
+   This overrides everything else, even if the review also praises products.
+
+2. If both positive and negative opinions are present but not directly about
+   Priceoye's website, service, or delivery → respond "neutral".
+
+3. Otherwise:
+   - Respond "positive" if the review is praising.
+   - Respond "negative" if the review is complaining.
+   - Respond "neutral" if it is neither.
+
+4. respond with only one word.
+
+Do NOT explain. Do NOT add anything else.
+
+Review: {review_text}
+Sentiment:
+"""
+
+    try:
+        response = client.chat.completions.create(
+            model="openai/gpt-5-mini",
+            messages=[{"role": "user", "content": prompt}],
+            temperature=0.2,
+            max_tokens=10
+        )
+
+        raw = response.choices[0].message.content or ""
+        raw_lower = raw.strip().lower()
+
+        if "positive" in raw_lower:
+            return "positive"
+        elif "negative" in raw_lower:
+            return "negative"
+        elif "neutral" in raw_lower:
+            return "neutral"
+        else:
+            return "neutral"
+
+    except Exception as e:
+        print("Error from API:", str(e))
+        return "neutral"
+
+# --- Validation Function (short version) ---
+def validate_review(description, rating):
+    LLM = get_sentiment(description)
+
+    data = pd.DataFrame({
+        'LLM': [LLM],
+        'description': [description],
+        'rating': [rating]
+    })
+
+    # Simplified rules (your full rules can be pasted here if needed)
+    if rating in ["4", "5"] and LLM == "positive":
+        decision = "accepted"
+    elif rating in ["1", "2"] and LLM == "negative":
+        decision = "rejected"
+    else:
+        decision = "ignored"
+
+    return LLM, decision
+
+# --- Gradio UI ---
+def classify_review(description, rating):
+    sentiment, decision = validate_review(description, rating)
+    return sentiment, decision
+
+with gr.Blocks() as demo:
+    gr.Markdown("## 📊 Review Sentiment & Validation Tool")
+
+    with gr.Row():
+        review_input = gr.Textbox(label="Enter Review", placeholder="Write review in English or Roman Urdu...")
+        rating_input = gr.Dropdown(choices=["1","2","3","4","5"], label="Rating (1-5)", value="5")
+
+    with gr.Row():
+        sentiment_output = gr.Textbox(label="Predicted Sentiment")
+        decision_output = gr.Textbox(label="Final Decision")
+
+    btn = gr.Button("Classify Review")
+    btn.click(fn=classify_review, inputs=[review_input, rating_input], outputs=[sentiment_output, decision_output])
+
+demo.launch()