Spaces:
Runtime error
Runtime error
| import json | |
| import random | |
| import os | |
| from fastapi import FastAPI, HTTPException, Query | |
| from fastapi.responses import JSONResponse, StreamingResponse | |
| from pydantic import BaseModel | |
| from typing import Optional, List, Dict, Any | |
| import uvicorn | |
| import io | |
| app = FastAPI(title="HitekFresh User Data API", description="Search and stream user data from 83GB JSON", version="1.0") | |
| JSON_FILE_PATH = "/app/users_data.json" # Docker mein path | |
| SAMPLE_SIZE = 100 # Random sample size | |
| # ----- Helper: Load JSON (streaming for large files) ----- | |
| def stream_json(file_path: str): | |
| """Generator to stream JSON array from file.""" | |
| with open(file_path, 'r', encoding='utf-8') as f: | |
| # Assume JSON is an array of objects | |
| data = json.load(f) | |
| for item in data: | |
| yield item | |
| # ----- In-memory cache for random data (first 10k records) ----- | |
| _cache = None | |
| def load_cache(): | |
| global _cache | |
| if _cache is None: | |
| try: | |
| with open(JSON_FILE_PATH, 'r', encoding='utf-8') as f: | |
| _cache = json.load(f)[:10000] # Load first 10k for speed | |
| except Exception as e: | |
| _cache = [] | |
| return _cache | |
| # ----- Models ----- | |
| class UserData(BaseModel): | |
| mobile: str | |
| name: str | |
| fname: Optional[str] = None | |
| address: Optional[str] = None | |
| alt: Optional[str] = None | |
| circle: Optional[str] = None | |
| id: Optional[str] = None | |
| email: Optional[str] = None | |
| # ----- Endpoints ----- | |
| async def root(): | |
| return {"message": "HitekFresh API is live!", "docs": "/docs"} | |
| async def search_users( | |
| mobile: Optional[str] = Query(None, description="Search by mobile number"), | |
| name: Optional[str] = Query(None, description="Search by name (partial)"), | |
| email: Optional[str] = Query(None, description="Search by email (partial)"), | |
| id: Optional[str] = Query(None, description="Search by Aadhar ID"), | |
| limit: int = Query(10, ge=1, le=100, description="Max results") | |
| ): | |
| """Search users by mobile, name, email, or Aadhar ID.""" | |
| if not any([mobile, name, email, id]): | |
| raise HTTPException(status_code=400, detail="Provide at least one search parameter (mobile, name, email, id)") | |
| results = [] | |
| for user in load_cache(): # Using cache for speed | |
| if mobile and user.get("mobile") != mobile: | |
| continue | |
| if name and name.lower() not in user.get("name", "").lower(): | |
| continue | |
| if email and email.lower() not in user.get("email", "").lower(): | |
| continue | |
| if id and user.get("id") != id: | |
| continue | |
| results.append(user) | |
| if len(results) >= limit: | |
| break | |
| if not results: | |
| return JSONResponse({"message": "No users found", "results": []}) | |
| return {"count": len(results), "results": results} | |
| async def get_by_mobile(mobile: str): | |
| """Get user by exact mobile number.""" | |
| for user in load_cache(): | |
| if user.get("mobile") == mobile: | |
| return user | |
| raise HTTPException(status_code=404, detail="User not found") | |
| async def get_by_aadhar(aadhar_id: str): | |
| """Get user by Aadhar ID.""" | |
| for user in load_cache(): | |
| if user.get("id") == aadhar_id: | |
| return user | |
| raise HTTPException(status_code=404, detail="User not found") | |
| async def get_random_users(count: int = Query(5, ge=1, le=50)): | |
| """Get random users from the dataset.""" | |
| cache = load_cache() | |
| if not cache: | |
| raise HTTPException(status_code=500, detail="Data not loaded") | |
| sample = random.sample(cache, min(count, len(cache))) | |
| return {"count": len(sample), "results": sample} | |
| async def stream_all_users(): | |
| """Stream all users as JSON lines (NDJSON) for large downloads.""" | |
| def generate(): | |
| with open(JSON_FILE_PATH, 'r', encoding='utf-8') as f: | |
| data = json.load(f) | |
| for user in data: | |
| yield json.dumps(user) + "\n" | |
| return StreamingResponse(generate(), media_type="application/x-ndjson") | |
| async def get_stats(): | |
| """Get basic stats about the dataset.""" | |
| cache = load_cache() | |
| return { | |
| "total_records_loaded": len(cache), | |
| "full_file_size_gb": 83.4, | |
| "fields": list(cache[0].keys()) if cache else [] | |
| } | |
| # ----- Run ----- | |
| if __name__ == "__main__": | |
| uvicorn.run(app, host="0.0.0.0", port=8000) |