llm/api.py

import re
from typing import Dict, Optional
from fastapi import FastAPI
from pydantic import BaseModel
from transformers import BitsAndBytesConfig, AutoModelForCausalLM, AutoTokenizer
from peft import PeftModel
import uvicorn
import torch

base_model_id = "Qwen/Qwen2.5-0.5B-Instruct"
# base_model_id = "qwen2.5-0.5b-instruct-sumobot-merged"
#lora_adapter_or_id = "qwen2.5-0.5b-instruct-sumobot"
lora_adapter_or_id = "adapters/qwen2.5_0.5b_lora_half"

# # Load tokenizer
tokenizer = AutoTokenizer.from_pretrained(
    base_model_id,
    trust_remote_code=True, use_fast=True)

bnb_config = BitsAndBytesConfig(
    load_in_4bit=True,
    bnb_4bit_use_double_quant=False,
    bnb_4bit_quant_type="nf4",
    bnb_4bit_compute_dtype=torch.bfloat16,
)

# Load base model
base_model = AutoModelForCausalLM.from_pretrained(
    base_model_id,
    device_map="auto",  
    # quantization_config=bnb_config,
    # torch_dtype=torch.bfloat16,
    trust_remote_code=True )


# Load LoRA weights
model = PeftModel.from_pretrained(base_model, lora_adapter_or_id)

# Merge LoRA into the base model (optional if you want a standalone model)
model = model.merge_and_unload()
# model.eval()

import re
from typing import Dict, Optional

def parse_action(output: str):
    # Mapping of shorthand to full action names
    action_map = {
        "SK": "Skill",
        "DS": "Dash",
        "FWD": "Accelerate",
        "TL": "TurnLeft",
        "TR": "TurnRight",
    }

    actions: Dict[str, Optional[float]] = {}

    for part in [p.strip() for p in output.split(",")]:
        name = part
        duration = None

        direct_match = re.match(r"^([A-Za-z]+)\s*([\d.]+)$", part)
        if direct_match:
            name = direct_match.group(1).strip()
            duration = float(direct_match.group(2))

        # Normalize shorthand to full name
        for short, full in action_map.items():
            if name.upper().startswith(short):
                name = full
                break

        actions[name] = duration

    print(f"result: {actions}")
    return {"action": actions}


def get_finetuned_action(query_state):
    # Inference with chat template
    messages = [
        {"role": "system", "content": "You are a Sumobot assistant"},
        {"role": "user", "content": f"State: {query_state}"},
    ]

    # Apply the tokenizer's built-in chat template
    chat_prompt = tokenizer.apply_chat_template(
        messages,
        tokenize=False,             
        add_generation_prompt=True   
    )

    inputs = tokenizer(chat_prompt, return_tensors="pt").to(model.device)

    outputs = model.generate(
        **inputs,
        max_new_tokens=15
    )

    # Slice out only the newly generated tokens
    generated_tokens = outputs[0][inputs["input_ids"].shape[1]:]

    decoded = tokenizer.decode(generated_tokens, skip_special_tokens=True).strip()

    parsedResult = parse_action(decoded)
    return parsedResult

# --------- API Setup ----------
app = FastAPI()

class QueryInput(BaseModel):
    state: str

@app.post("/query")
def query(input: QueryInput):
    return get_finetuned_action(input.state)

# Run with: python rag_api.py
if __name__ == "__main__":
    uvicorn.run(app, host="0.0.0.0", port=8000)


# TEST CURL
# curl --location 'http://localhost:8000/query' \
# --header 'Content-Type: application/json' \
# --data '{
#     "state":"AngleToEnemy=7.77, AngleToEnemyScore=0.99, DistanceToEnemyScore=0.76, NearBorderArenaScore=0.81, FacingToArena=-0.99"
# }'