Spaces:
Sleeping
Sleeping
File size: 4,278 Bytes
0279c66 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 | from fastapi import FastAPI, HTTPException
from pydantic import BaseModel
import uvicorn
import logging
import os
from dotenv import load_dotenv
from param_mem.memory.parametric import ParametricMemory
from param_mem.memory.retrieval import CrossSampleMemory
from param_mem.agent.agent_loop import ParamAgent
from fastapi.middleware.cors import CORSMiddleware
from fastapi.staticfiles import StaticFiles
# Load environment variables
load_dotenv(override=True)
logging.basicConfig(level=logging.INFO)
logger = logging.getLogger(__name__)
app = FastAPI(title="ParamMem Agent API", version="1.0.0")
app.add_middleware(
CORSMiddleware,
allow_origins=["*"],
allow_credentials=True,
allow_methods=["*"],
allow_headers=["*"],
)
# Global variables for modules
memory_module = None
retrieval_module = None
agent = None
class SolveRequest(BaseModel):
problem: str
max_iterations: int = 3
use_param_plus: bool = True
@app.on_event("startup")
async def startup_event():
global memory_module, retrieval_module, agent
logger.info("Initializing ParamMem Engine...")
provider = os.getenv("MODEL_PROVIDER", "local").lower()
api_key = os.getenv("GROQ_API_KEY", "").strip()
model_id = os.getenv("GROQ_MODEL", "openai/gpt-oss-20b")
if api_key:
logger.info(f"GROQ_API_KEY detected (starts with {api_key[:5]}...)")
else:
logger.warning("GROQ_API_KEY not found in environment!")
base_model = "TinyLlama/TinyLlama-1.1B-Chat-v1.0"
lora_path = "./param_mem_lora"
if not os.path.exists(lora_path) or not os.listdir(lora_path):
lora_path = None
try:
if provider == "groq" and api_key:
logger.info("Using Groq API provider...")
memory_module = ParametricMemory(
provider="groq",
api_key=api_key,
model_id=model_id
)
else:
logger.info("Using local HuggingFace provider...")
memory_module = ParametricMemory(
base_model_name=base_model,
lora_path=lora_path,
provider="local"
)
retrieval_module = CrossSampleMemory()
agent = ParamAgent(memory_module=memory_module, retrieval_module=retrieval_module)
logger.info(f"ParamMem Engine ready (Provider: {provider}).")
except Exception as e:
logger.error(f"Failed to initialize models: {e}")
agent = None
@app.post("/solve")
async def solve_problem(req: SolveRequest):
if not agent:
raise HTTPException(status_code=503, detail="Agent is not initialized. Please check server logs.")
try:
# Note: In a production app, we wouldn't re-instantiate the agent per request
# but for this demo it allows toggling retrieval.
current_retrieval = retrieval_module if req.use_param_plus else None
# We can reuse the memory_module but might need a new agent instance if retrieval toggles
# For simplicity, we just use the global agent if it matches req.use_param_plus
# but the solve_task doesn't currently care about toggling retrieval internally easily.
# Let's just pass the requirement to the solve_task if possible,
# but the current agent class has it fixed.
# Temporary fix for demo:
temp_agent = ParamAgent(memory_module=memory_module, retrieval_module=current_retrieval)
result = temp_agent.solve_task(req.problem, max_iterations=req.max_iterations)
return result
except Exception as e:
logger.error(f"Error during solving: {e}")
raise HTTPException(status_code=500, detail=str(e))
@app.get("/health")
async def health_check():
return {
"status": "ok",
"agent_loaded": agent is not None,
"provider": os.getenv("MODEL_PROVIDER", "local")
}
# Mount static files (React frontend)
# Ensure this is after all API routes
if os.path.exists("dist"):
app.mount("/", StaticFiles(directory="dist", html=True), name="static")
else:
logger.warning("Dist directory not found. Frontend will not be served.")
if __name__ == "__main__":
uvicorn.run("src.api.server:app", host="0.0.0.0", port=8000, reload=True)
|