from fastapi import FastAPI, Request from fastapi.responses import StreamingResponse from fastapi.middleware.cors import CORSMiddleware from typing import List, Dict, Any, Optional from pydantic import BaseModel import asyncio import httpx import random from config import cookies, headers, groqapi from prompts import ChiplingPrompts from groq import Groq import json from fastapi.responses import HTMLResponse from fastapi.templating import Jinja2Templates from pathlib import Path from collections import Counter, defaultdict from utils.logger import log_request from chipsearch.main import search from scrape.main import scrape_to_markdown app = FastAPI() # Add CORS middleware app.add_middleware( CORSMiddleware, allow_origins=["http://localhost:8080", "https://www.chipling.xyz"], allow_credentials=True, allow_methods=["*"], allow_headers=["*"], ) templates = Jinja2Templates(directory="templates") LOG_FILE = Path("logs.json") @app.get("/dashboard", response_class=HTMLResponse) async def dashboard(request: Request, endpoint: str = None): try: with open("logs.json") as f: logs = json.load(f) except FileNotFoundError: logs = [] # Filter logs if endpoint: logs = [log for log in logs if log["endpoint"] == endpoint] # Summary stats total_requests = len(logs) endpoint_counts = Counter(log["endpoint"] for log in logs) query_counts = Counter(log["query"] for log in logs) # Requests per date date_counts = defaultdict(int) for log in logs: date = log["timestamp"].split("T")[0] date_counts[date] += 1 # Sort logs by timestamp (desc) logs_sorted = sorted(logs, key=lambda x: x["timestamp"], reverse=True) return templates.TemplateResponse("dashboard.html", { "request": request, "logs": logs_sorted[:100], # show top 100 "total_requests": total_requests, "endpoint_counts": dict(endpoint_counts), "query_counts": query_counts.most_common(5), "date_counts": dict(date_counts), "filter_endpoint": endpoint or "", }) # Define request model class ChatRequest(BaseModel): message: str messages: List[Dict[Any, Any]] model: Optional[str] = "meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8" client = Groq(api_key=groqapi) async def generate(json_data: Dict[str, Any]): max_retries = 5 for attempt in range(max_retries): async with httpx.AsyncClient(timeout=None) as client: try: request_ctx = client.stream( "POST", "https://api.together.ai/inference", cookies=cookies, headers=headers, json=json_data ) async with request_ctx as response: if response.status_code == 200: async for line in response.aiter_lines(): if line: yield f"{line}\n" return elif response.status_code == 429: if attempt < max_retries - 1: await asyncio.sleep(0.5) continue yield "data: [Rate limited, max retries]\n\n" return else: yield f"data: [Unexpected status code: {response.status_code}]\n\n" return except Exception as e: yield f"data: [Connection error: {str(e)}]\n\n" return yield "data: [Max retries reached]\n\n" def convert_to_groq_schema(messages: List[Dict[str, Any]]) -> List[Dict[str, str]]: converted = [] for message in messages: role = message.get("role", "user") content = message.get("content") if isinstance(content, list): flattened = [] for item in content: if isinstance(item, dict) and item.get("type") == "text": flattened.append(item.get("text", "")) content = "\n".join(flattened) elif not isinstance(content, str): content = str(content) converted.append({"role": role, "content": content}) return converted def conver_to_xai_schema(messages: List[Dict[str, Any]]) -> List[Dict[str, str]]: converted = [] for message in messages: role = message.get("role", "user") content = message.get("content", "") if isinstance(content, list): # Handle content that's already in parts format parts = content text_content = "\n".join([p.get("text", "") for p in content if p.get("type") == "text"]) else: # Create parts format for text content text_content = str(content) parts = [{"type": "text", "text": text_content}] if role == "assistant": parts.insert(0, {"type": "step-start"}) converted.append({ "role": role, "content": text_content, "parts": parts }) return converted async def groqgenerate(json_data: Dict[str, Any]): try: messages = convert_to_groq_schema(json_data["messages"]) chunk_id = "groq-" + "".join(random.choices("0123456789abcdef", k=32)) created = int(asyncio.get_event_loop().time()) # Create streaming response stream = client.chat.completions.create( messages=messages, model=json_data.get("model", "meta-llama/llama-4-scout-17b-16e-instruct"), temperature=json_data.get("temperature", 0.7), max_completion_tokens=json_data.get("max_tokens", 1024), top_p=json_data.get("top_p", 1), stop=json_data.get("stop", None), stream=True, ) total_tokens = 0 # Use normal for-loop since stream is not async for chunk in stream: content = chunk.choices[0].delta.content if content: response = { "id": chunk_id, "object": "chat.completion.chunk", "created": created, "model": json_data.get("model", "meta-llama/llama-4-scout-17b-16e-instruct"), "choices": [{ "index": 0, "text": content, "logprobs": None, "finish_reason": None }], "usage": None } yield f"data: {json.dumps(response)}\n\n" total_tokens += 1 final = { "id": chunk_id, "object": "chat.completion.chunk", "created": created, "model": json_data.get("model", "meta-llama/llama-4-scout-17b-16e-instruct"), "choices": [], "usage": { "prompt_tokens": len(messages), "completion_tokens": total_tokens, "total_tokens": len(messages) + total_tokens, } } yield f"data: {json.dumps(final)}\n\n" yield "data: [DONE]\n\n" except Exception as e: generate(json_data) async def vercelXaigenerate(json_data: Dict[str, Any]): headers = { 'accept': '*/*', 'accept-language': 'en-US,en;q=0.9,ja;q=0.8', 'content-type': 'application/json', 'origin': 'https://ai-sdk-starter-xai.vercel.app', 'referer': 'https://ai-sdk-starter-xai.vercel.app/', 'sec-ch-ua': '"Google Chrome";v="135", "Not-A.Brand";v="8", "Chromium";v="135"', 'sec-ch-ua-mobile': '?0', 'sec-ch-ua-platform': '"macOS"', 'user-agent': 'Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_7) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/135.0.0.0 Safari/537.36' } messages = conver_to_xai_schema(json_data["messages"]) request_data = { "id": "".join(random.choices("0123456789abcdef", k=16)), "messages": messages, "selectedModel": json_data.get("model", "grok-2-1212"), } print(request_data) chunk_id = "xai-" + "".join(random.choices("0123456789abcdef", k=32)) created = int(asyncio.get_event_loop().time()) total_tokens = 0 try: async with httpx.AsyncClient(timeout=None) as client: async with client.stream( "POST", "https://ai-sdk-starter-xai.vercel.app/api/chat", headers=headers, json=request_data ) as request_ctx: if request_ctx.status_code == 200: async for line in request_ctx.aiter_lines(): if line: if line.startswith('0:'): # Clean up the text and properly escape JSON characters text = line[2:].strip() if text.startswith('"') and text.endswith('"'): text = text[1:-1] text = text.replace('\\n', '\n').replace('\\', '') response = { "id": chunk_id, "object": "chat.completion.chunk", "created": created, "model": json_data.get("model", "grok-2-1212"), "choices": [{ "index": 0, "text": text, "logprobs": None, "finish_reason": None }], "usage": None } yield f"data: {json.dumps(response)}\n\n" total_tokens += 1 elif line.startswith('d:'): final = { "id": chunk_id, "object": "chat.completion.chunk", "created": created, "model": json_data.get("model", "grok-2-1212"), "choices": [], "usage": { "prompt_tokens": len(messages), "completion_tokens": total_tokens, "total_tokens": len(messages) + total_tokens } } yield f"data: {json.dumps(final)}\n\n" yield "data: [DONE]\n\n" return else: yield f"data: [Unexpected status code: {request_ctx.status_code}]\n\n" except Exception as e: yield f"data: [Connection error: {str(e)}]\n\n" async def vercelGroqgenerate(json_data: Dict[str, Any]): headers = { 'accept': '*/*', 'accept-language': 'en-US,en;q=0.9,ja;q=0.8', 'content-type': 'application/json', 'origin': 'https://ai-sdk-starter-groq.vercel.app', 'priority': 'u=1, i', 'referer': 'https://ai-sdk-starter-groq.vercel.app/', 'sec-ch-ua': '"Google Chrome";v="135", "Not-A.Brand";v="8", "Chromium";v="135"', 'sec-ch-ua-mobile': '?0', 'sec-ch-ua-platform': '"macOS"', 'sec-fetch-dest': 'empty', 'sec-fetch-mode': 'cors', 'sec-fetch-site': 'same-origin', 'user-agent': 'Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_7) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/135.0.0.0 Safari/537.36', } messages = conver_to_xai_schema(json_data["messages"]) request_data = { "id": "".join(random.choices("0123456789abcdef", k=16)), "messages": messages, "selectedModel": json_data.get("model", "deepseek-r1-distill-llama-70b"), } chunk_id = "vercel-groq-" + "".join(random.choices("0123456789abcdef", k=32)) created = int(asyncio.get_event_loop().time()) total_tokens = 0 try: async with httpx.AsyncClient(timeout=None) as client: async with client.stream( "POST", "https://ai-sdk-starter-groq.vercel.app/api/chat", headers=headers, json=request_data ) as request_ctx: print(request_ctx.status_code) if request_ctx.status_code == 200: async for line in request_ctx.aiter_lines(): if line: if line.startswith('0:'): # Clean up the text and properly escape JSON characters text = line[2:].strip() if text.startswith('"') and text.endswith('"'): text = text[1:-1] text = text.replace('\\n', '\n').replace('\\', '') response = { "id": chunk_id, "object": "chat.completion.chunk", "created": created, "model": json_data.get("model", "deepseek-r1-distill-llama-70b"), "choices": [{ "index": 0, "text": text, "logprobs": None, "finish_reason": None }], "usage": None } yield f"data: {json.dumps(response)}\n\n" total_tokens += 1 elif line.startswith('d:'): final = { "id": chunk_id, "object": "chat.completion.chunk", "created": created, "model": json_data.get("model", "deepseek-r1-distill-llama-70b"), "choices": [], "usage": { "prompt_tokens": len(messages), "completion_tokens": total_tokens, "total_tokens": len(messages) + total_tokens } } yield f"data: {json.dumps(final)}\n\n" yield "data: [DONE]\n\n" return else: yield f"data: [Unexpected status code: {request_ctx.status_code}]\n\n" except Exception as e: yield f"data: [Connection error: {str(e)}]\n\n" @app.get("/") async def index(): return {"status": "ok", "message": "Welcome to the Chipling API!", "version": "2.0", "routes": ["/chat", "/generate-modules", "/generate-topics", "/v1/generate", "/v1/generate-images", "/chipsearch", "/scrape-md"]} @app.post("/chat") async def chat(request: ChatRequest): current_messages = request.messages.copy() # Handle both single text or list content if request.messages and isinstance(request.messages[-1].get('content'), list): current_messages = request.messages else: current_messages.append({ 'content': [{ 'type': 'text', 'text': request.message }], 'role': 'user' }) json_data = { 'model': request.model, 'max_tokens': None, 'temperature': 0.7, 'top_p': 0.7, 'top_k': 50, 'repetition_penalty': 1, 'stream_tokens': True, 'stop': ['<|eot_id|>', '<|eom_id|>'], 'messages': current_messages, 'stream': True, } selected_generator = random.choice([generate, groqgenerate, vercelGroqgenerate, vercelXaigenerate]) log_request("/chat", selected_generator.__name__) return StreamingResponse(selected_generator(json_data), media_type='text/event-stream') @app.post("/chipsearch") async def chipsearch(request: Request): data = search( term=request.query_params.get("term"), num_results=int(request.query_params.get("num_results", 10)), advanced=bool(request.query_params.get("advanced", False)), unique=bool(request.query_params.get("unique", False)) ) return data @app.post("/scrape-md") async def scrape_md(request: Request): data = await request.json() url = data.get("url") if not url: return {"error": "URL is required"} data = scrape_to_markdown(url) return {"markdown": data} @app.post("/v1/generate") async def api_generate(request: Request): data = await request.json() messages = data["messages"] model = data["model"] if not messages: return {"error": "messages is required"} elif not model: return {"error": "Model is required"} try: json_data = { 'model': model, 'max_tokens': None, 'temperature': 0.7, 'top_p': 0.7, 'top_k': 50, 'repetition_penalty': 1, 'stream_tokens': True, 'stop': ['<|eot_id|>', '<|eom_id|>'], 'messages': messages, 'stream': True, } # Define model lists for each provider xai_models = ["grok-3-mini", "grok-2-1212", "grok-3", "grok-3-fast", "grok-3-mini-fast"] together_models = ['meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8', 'meta-llama/Llama-4-Scout-17B-16E-Instruct', 'deepseek-ai/DeepSeek-R1', 'deepseek-ai/DeepSeek-V3', 'Qwen/Qwen2.5-VL-72B-Instruct', 'google/gemma-2-27b-it'] groq_models = ['qwen-qwq-32b', 'gemma2-9b-it', 'meta-llama/llama-4-maverick-17b-128e-instruct', 'meta-llama/llama-4-scout-17b-16e-instruct'] vercel_groq_models = ['meta-llama/llama-4-scout-17b-16e-instruct', 'llama-3.1-8b-instant', 'llama-3.3-70b-versatile', 'deepseek-r1-distill-llama-70b'] # Create a list of available generators for the requested model available_generators = [] if model in xai_models: available_generators.append(vercelXaigenerate) if model in together_models: available_generators.append(generate) if model in groq_models: available_generators.append(groqgenerate) if model in vercel_groq_models: available_generators.append(vercelGroqgenerate) if not available_generators: return {"error": "No suitable generator found for the specified model"} # Randomly select one generator if multiple are available selected_generator = random.choice(available_generators) log_request("/v1/generate", selected_generator.__name__) return StreamingResponse(selected_generator(json_data), media_type='text/event-stream') except Exception as e: return {"error": f"Generation failed: {str(e)}"} @app.post("/v1/generate-images") async def generate_images(request: Request): data = await request.json() prompt = data.get("prompt") provider = data.get("provider") modelId = data.get("modelId") if not prompt: return {"error": "Prompt is required"} if not provider: return {"error": "Provider is required"} if not modelId: return {"error": "Model ID is required"} headers = { 'accept': '*/*', 'accept-language': 'en-US,en;q=0.9,ja;q=0.8', 'content-type': 'application/json', 'origin': 'https://fal-image-generator.vercel.app', 'priority': 'u=1, i', 'referer': 'https://fal-image-generator.vercel.app/', 'sec-ch-ua': '"Google Chrome";v="135", "Not-A.Brand";v="8", "Chromium";v="135"', 'sec-ch-ua-mobile': '?0', 'sec-ch-ua-platform': '"macOS"', 'sec-fetch-dest': 'empty', 'sec-fetch-mode': 'cors', 'sec-fetch-site': 'same-origin', 'user-agent': 'Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_7) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/135.0.0.0 Safari/537.36', } json_data = { 'prompt': prompt, 'provider': 'fal', 'modelId': 'fal-ai/fast-sdxl', } async with httpx.AsyncClient() as client: response = await client.post( 'https://fal-image-generator.vercel.app/api/generate-images', headers=headers, json=json_data ) return response.json()