Update app.py

app.py

@@ -7,151 +7,97 @@ import gradio as gr
 from dotenv import load_dotenv
 import os
 
-load_dotenv()  
+import faiss
+import gradio as gr
+from sentence_transformers import SentenceTransformer
+from google import genai
+from google.genai import types
+
+# ✅ Take Gemini API key from user
+api_key = input("🔑 Enter your Gemini API Key: ")
+client = genai.Client(api_key=api_key)
 
-# ==========================
-# Globals
-# ==========================
-MODEL_NAME = "gemini-2.5-flash"
+# ✅ Load embedding model
 embedder = SentenceTransformer("sentence-transformers/all-MiniLM-L6-v2")
 
-index = None
-meta = None
-client = None
+MODEL_NAME = "gemini-2.5-flash"   # use stable model
 
+# ✅ Read file (assumes filename already set from upload)
+with open(filename, "r", encoding="utf-8") as f:
+    raw_text = f.read()
 
-# ==========================
-# Helper Functions
-# ==========================
+# ✅ Split into chunks
 def chunk_text(text, chunk_chars=800, overlap=100):
-    """Split text into overlapping chunks"""
     chunks, start = [], 0
     while start < len(text):
         end = min(start + chunk_chars, len(text))
         chunks.append(text[start:end].strip())
         start = end - overlap
-        if start < 0:
-            start = 0
-        if end == len(text):
-            break
+        if start < 0: start = 0
+        if end == len(text): break
     return [c for c in chunks if c]
 
+chunks = chunk_text(raw_text)
+embeddings = embedder.encode(chunks, convert_to_numpy=True, normalize_embeddings=True)
 
-def build_index(file_obj):
-    """Build FAISS index from uploaded file"""
-    global index, meta
-
-    raw_text = file_obj.read().decode("utf-8")
-    chunks = chunk_text(raw_text)
-
-    embeddings = embedder.encode(
-        chunks, convert_to_numpy=True, normalize_embeddings=True
-    )
-
-    dim = embeddings.shape[1]
-    index = faiss.IndexFlatIP(dim)
-    index.add(embeddings)
-    meta = [{"id": i, "chunk": c} for i, c in enumerate(chunks)]
-
-    return "✅ File processed! You can now ask questions."
-
+dim = embeddings.shape[1]
+index = faiss.IndexFlatIP(dim)
+index.add(embeddings)
+meta = [{"id": i, "chunk": c} for i, c in enumerate(chunks)]
 
-def set_api_key(api_key):
-    """Initialize Gemini client with API key"""
-    global client
-    try:
-        client = genai.Client(api_key=api_key)
-        return "✅ API Key set successfully!"
-    except Exception as e:
-        return f"❌ Error setting API Key: {e}"
+# ✅ Prompt templates
+SYSTEM_TEMPLATE = """You are a professional Arduino freelancer.
+Answer ONLY using the provided Fiverr chat context.
+If the answer is not in the context, reply:
+"I don’t know based on the provided chat."
+"""
+USER_TEMPLATE = """Client asked:
+{question}
 
+Relevant chat context:
+{context}
+"""
 
+# ✅ Retrieve top chunks
 def retrieve(query, top_k=3):
-    """Retrieve top matching chunks"""
-    if index is None:
-        return []
-
     q_emb = embedder.encode([query], convert_to_numpy=True, normalize_embeddings=True)
     scores, idxs = index.search(q_emb, top_k)
-
     results = []
     for i in idxs[0]:
-        if i == -1:
-            continue
+        if i == -1: continue
         results.append(meta[i]["chunk"])
     return results
 
-
+# ✅ Call Gemini (new client API)
 def call_gemini(system_prompt, user_prompt):
-    """Query Gemini API"""
-    if client is None:
-        return "❌ Please enter your Gemini API Key first."
-
     try:
         response = client.models.generate_content(
             model=MODEL_NAME,
             contents=f"{system_prompt}\n\n{user_prompt}",
             config=types.GenerateContentConfig(
-                thinking_config=types.ThinkingConfig(thinking_budget=0)
+                thinking_config=types.ThinkingConfig(thinking_budget=0)  # disables thinking
             ),
         )
         return response.text
     except Exception as e:
         return f"❌ Gemini error: {e}"
 
-
+# ✅ Answer function
 def answer_question(question, top_k=3):
-    """Answer user query based on retrieved context"""
     context_chunks = retrieve(question, top_k)
     if not context_chunks:
-        return "⚠️ No relevant context found in your file."
-
+        return "⚠️ No relevant context found in your chat file."
     context = "\n---\n".join(context_chunks)
-    system_prompt = """You are a professional Arduino freelancer.
-Answer ONLY using the provided Fiverr chat context.
-If the answer is not in the context, reply:
-"I don’t know based on the provided chat."
-"""
-    user_prompt = f"""Client asked:
-{question}
-
-Relevant chat context:
-{context}
-"""
-    return call_gemini(system_prompt, user_prompt)
-
+    user_prompt = USER_TEMPLATE.format(question=question, context=context)
+    return call_gemini(SYSTEM_TEMPLATE, user_prompt)
 
+# ✅ Gradio Chatbot
 def chatbot_fn(message, history):
-    """Gradio chatbot function"""
     reply = answer_question(message)
     return reply
 
-
-# ==========================
-# Gradio UI
-# ==========================
 with gr.Blocks() as demo:
     gr.Markdown("## 🤖 Arduino Fiverr Chatbot (Gemini Powered)")
+    gr.ChatInterface(fn=chatbot_fn, title="Arduino Client Assistant")
 
-    with gr.Row():
-        api_key_box = gr.Textbox(label="🔑 Enter Gemini API Key", type="password")
-        api_key_btn = gr.Button("Set API Key")
-        api_status = gr.Label(label="API Key Status")
-
-    api_key_btn.click(set_api_key, inputs=api_key_box, outputs=api_status)
-
-    with gr.Row():
-        file_upload = gr.File(label="📂 Upload your Fiverr chat (.txt)", type="file")
-        file_status = gr.Label(label="File Status")
-
-    file_upload.upload(build_index, inputs=file_upload, outputs=file_status)
-
-    gr.ChatInterface(
-        fn=chatbot_fn,
-        title="Arduino Client Assistant",
-        textbox="Ask your client-related question here...",
-    )
-
-# Run app
-if __name__ == "__main__":
-    demo.launch()
+demo.launch()