Update app.py

app.py

@@ -1,47 +1,23 @@
 import os
 import asyncio
 import time
-from typing import Optional
-from datetime import datetime, timezone
+from typing import Optional, List
+from datetime import datetime
 import httpx
 import trafilatura
 import gradio as gr
 from dateutil import parser as dateparser
+from limits import parse
+from limits.aio.storage import MemoryStorage
+from limits.aio.strategies import MovingWindowRateLimiter
+from analytics import record_request, last_n_days_df, last_n_days_avg_time_df
 from fastapi import FastAPI
 from pydantic import BaseModel
-import google.generativeai as genai
-
-# <<< NEW: Dono research modes ke liye alag prompts >>>
-PROMPT_NORMAL = """
-Based on the user's original query, provide a concise summary (3-4 important bullet points) of the following text. Focus only on the most critical information.
-
-USER'S QUERY: "{query}"
-
-TEXT TO SUMMARIZE:
----
-{context_text}
----
-"""
-
-PROMPT_DEEP = """
-As a meticulous research analyst, your task is to synthesize the information from the provided web search results into a detailed and comprehensive report.
-**Current Date:** {current_date}.
-**VERY IMPORTANT:** Your top priority is to provide information relevant to this current date and the future. If the user's query is about a recurring event (like an exam), you MUST focus on the upcoming or current event.
-**User's Original Query:** "{query}"
-**Instructions:**
-1.  Combine information from different sources to create a coherent and detailed report.
-2.  Cite source URLs inline, like this: "(Source: http://...)." The URL is provided in the text.
-3.  At the end of your report, create a "## Sources" section and list all the unique URLs you used.
-4.  Use clear markdown with headings and bold text.
-
-**Provided Search Results:**
----
-{context_text}
----
-"""
+import google.generativeai as genai # <<< NEW IMPORT
 
 # --- Core Search Logic (No Changes) ---
 async def search_web_logic(query: str, serper_api_key: str, search_type: str, num_results: int) -> str:
+    # ... (Ye function bilkul waisa hi hai jaisa pehle tha, isme koi badlav nahi)
     start_time = time.time()
     if not serper_api_key: return "Error: Serper API Key is required."
     num_results = max(1, min(20, num_results))
@@ -70,104 +46,73 @@ async def search_web_logic(query: str, serper_api_key: str, search_type: str, nu
             chunks.append(chunk)
         if not chunks: return "Found results, but couldn't extract content."
         summary = f"Successfully extracted content from {successful_extractions}/{len(results)} results.\n\n---\n\n"
-        # Removed the record_request call to simplify, can be added back if needed
+        await record_request(time.time() - start_time, num_results)
         return summary + "\n---\n".join(chunks)
     except Exception as e:
         return f"An error occurred during web search: {str(e)}"
 
-# --- Gemini Summarization Logic ---
-async def summarize_with_gemini(text_to_summarize: str, query: str, gemini_key: str, model_name: str, research_mode: str) -> str:
+# <<< NEW: Gemini Summarization Logic >>>
+async def summarize_with_gemini(text_to_summarize: str, query: str, gemini_key: str, model_name: str) -> str:
     try:
         genai.configure(api_key=gemini_key)
         model = genai.GenerativeModel(model_name)
-        current_date = datetime.now(timezone.utc).strftime("%Y-%m-%d")
-
-        # <<< CHANGE: Research mode ke hisab se prompt select hoga >>>
-        if research_mode == 'deep':
-            prompt_template = PROMPT_DEEP
-        else: # Default to normal
-            prompt_template = PROMPT_NORMAL
-            
-        prompt = prompt_template.format(query=query, context_text=text_to_summarize, current_date=current_date)
+        
+        prompt = f"""
+        Based on the user's original query, provide a concise summary of the following text.
+        Focus on the information that directly answers or relates to the query.
+        Filter out any irrelevant details, advertisements, or boilerplate content.
+        Present the final summary in clean, easy-to-read markdown format.
+
+        USER'S QUERY: "{query}"
+
+        TEXT TO SUMMARIZE:
+        ---
+        {text_to_summarize}
+        ---
+        """
         
         response = await model.generate_content_async(prompt)
         return response.text
     except Exception as e:
         return f"\n\n--- ⚠️ Gemini Summarization Failed ---\nError: {str(e)}\nReturning raw text instead."
 
-# --- Main Orchestrator Function ---
-async def search_and_summarize(query, serper_api_key, search_type, num_results, gemini_api_key, gemini_model, research_mode):
-    scraped_text = await search_web_logic(query, serper_api_key, search_type, num_results)
-    
-    if gemini_api_key and "Error:" not in scraped_text:
-        summarized_text = await summarize_with_gemini(scraped_text, query, gemini_api_key, gemini_model, research_mode)
-        if "⚠️ Gemini Summarization Failed" in summarized_text:
-             return scraped_text + summarized_text
-        else:
-             return summarized_text
-    return scraped_text
-
 # --- FastAPI App ---
 app = FastAPI()
 
+# <<< CHANGE: FastAPI request model ko naye fields ke sath update kiya gaya >>>
 class SearchRequest(BaseModel):
     query: str
     serper_api_key: str
     search_type: str = "search"
     num_results: int = 4
-    gemini_api_key: Optional[str] = None
-    gemini_model: Optional[str] = "gemini-1.5-flash-latest"
-    research_mode: str = "normal" # <<< NEW: Research mode field
+    gemini_api_key: Optional[str] = None # Optional field
+    gemini_model: Optional[str] = "gemini-1.5-flash-latest" # Optional field with default
 
 @app.post("/api/search")
 async def api_search(request: SearchRequest):
-    result = await search_and_summarize(
-        request.query, request.serper_api_key, request.search_type, request.num_results,
-        request.gemini_api_key, request.gemini_model, request.research_mode
+    # Step 1: Web Search
+    scraped_text = await search_web_logic(
+        request.query, request.serper_api_key, request.search_type, request.num_results
     )
-    return {"result": result}
 
-# --- Gradio App ---
-def create_gradio_app():
-    with gr.Blocks(title="Web Search & Summarize UI") as demo:
-        gr.Markdown("# 🔍 AI Search & Summarize")
-        with gr.Tabs():
-            with gr.Tab("App"):
-                gr.Markdown("### Step 1: Web Search")
-                query_input = gr.Textbox(label="Search Query")
-                serper_api_key_input = gr.Textbox(label="Your Serper API Key", type="password")
-                with gr.Row():
-                    search_type_input = gr.Radio(["search", "news"], value="search", label="Search Type")
-                    num_results_input = gr.Slider(1, 20, value=4, step=1, label="Number of Results")
-                
-                gr.Markdown("### Step 2: AI Summarization")
-                # <<< NEW: Research mode ke liye radio buttons >>>
-                research_mode_input = gr.Radio(["normal", "deep"], value="normal", label="Research Mode", info="Normal for fast summary, Deep for detailed report.")
-                gemini_api_key_input = gr.Textbox(label="Your Gemini API Key", type="password", placeholder="Leave empty to skip summarization")
-                gemini_model_input = gr.Textbox(label="Gemini Model", value="gemini-1.5-flash-latest")
-
-                search_button = gr.Button("Search & Summarize", variant="primary")
-                output = gr.Textbox(label="Result", lines=25, max_lines=40)
-                
-                # <<< CHANGE: Naya research_mode_input, inputs list me add kiya gaya >>>
-                search_button.click(
-                    fn=search_and_summarize,
-                    inputs=[query_input, serper_api_key_input, search_type_input, num_results_input, gemini_api_key_input, gemini_model_input, research_mode_input],
-                    outputs=output
-                )
-            with gr.Tab("Analytics"):
-                # Analytics tab remains unchanged
-                requests_plot = gr.BarPlot(x="date", y="count", title="Daily Requests")
-                avg_time_plot = gr.BarPlot(x="date", y="avg_time", title="Avg. Response Time (s)")
-                def update_analytics(): return last_n_days_df(14), last_n_days_avg_time_df(14)
-                demo.load(update_analytics, [], [requests_plot, avg_time_plot])
+    # Step 2: (Optional) Summarize with Gemini
+    if request.gemini_api_key and "Error:" not in scraped_text:
+        summarized_text = await summarize_with_gemini(
+            scraped_text, request.query, request.gemini_api_key, request.gemini_model
+        )
+        # Agar summarization fail hota hai, to original text wapas bhej dein
+        if "⚠️ Gemini Summarization Failed" in summarized_text:
+             return {"result": scraped_text + summarized_text}
+        else:
+             return {"result": summarized_text}
 
-    return demo
+    return {"result": scraped_text}
 
-# --- Mount and Startup ---
-gradio_ui = create_gradio_app()
-app = gr.mount_gradio_app(app, gradio_ui, path="/")
+# --- Gradio App (ab iski zaroorat nahi, lekin rakha ja sakta hai) ---
+# ... (Gradio UI ka code abhi ke liye comment out kar sakte hain ya hata sakte hain)
+# ... agar aapko UI bhi chahiye to hum use bhi update kar sakte hain
 
+# --- Server Startup ---
 if __name__ == "__main__":
     import uvicorn
     uvicorn.run(app, host="0.0.0.0", port=7860)