File size: 1,222 Bytes
1c4cf5d
2f13f99
 
1c4cf5d
991437e
f34ea4b
a1b5009
ab4a23f
1c4cf5d
a1b5009
2f13f99
 
1c4cf5d
4ed3db0
 
 
 
f0f3fcb
4ed3db0
02b62d9
4ed3db0
f0f3fcb
4ed3db0
 
 
a1b5009
1c4cf5d
2f13f99
 
 
1c4cf5d
a1b5009
f34ea4b
 
 
1c4cf5d
f0f3fcb
2f13f99
 
1c4cf5d
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
import os
import json

from tools import TOOLS
from metadata import load_metadata
from mistral_hf_wrapper import MistralInference

API_URL = os.getenv("HF_MISTRAL_ENDPOINT")
API_TOKEN = os.getenv("HF_TOKEN")

mistral = MistralInference(api_url=API_URL, api_token=API_TOKEN)

def load_tasks():
    tasks = []
    with open("metadata.jsonl", "r", encoding="utf-8") as f:
        for line in f:
            item = json.loads(line)
            task_id = item.get("task_id")
            question = item.get("Question") or item.get("question")
            if task_id and question:
                tasks.append({
                    "task_id": task_id,
                    "question": question
                })
    return tasks

def solve_task(task, tools=TOOLS):
    system_prompt = (
        "You are a helpful agent. Use reasoning, tools if needed, and return the answer only."
    )
    user_prompt = task["question"]

    prompt = f"<s>[INST] <<SYS>>\n{system_prompt}\n<</SYS>>\n{user_prompt} [/INST]"
    response = mistral.run(prompt)

    return {
        "task_id": task["task_id"],
        "question": task["question"],
        "submitted_answer": response.strip() if response else "ERROR: Empty model response"
    }