Spaces:
Building
Building
File size: 1,928 Bytes
092c3de |
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 |
import os
import requests
from utils import log
class LLMConnector:
def __init__(self, service_config):
self.service_config = service_config
def call_spark(self, project_name, prompt, chat_history):
project = self.service_config.get_project(project_name)
if not project:
raise Exception(f"Project not found: {project_name}")
llm_config = project.get("llm", {})
if not llm_config:
raise Exception(f"LLM config missing for project: {project_name}")
spark_url = self.service_config.llm_inference_service_url
work_mode = self.service_config.work_mode
cloud_token = self.service_config.get_auth_token()
headers = {
"Authorization": f"Bearer {cloud_token}",
"Content-Type": "application/json"
}
payload = {
"work_mode": work_mode,
"cloud_token": cloud_token,
"project_name": project_name,
"project_version": project.get("project_version"),
"repo_id": llm_config.get("repo_id"),
"generation_config": llm_config.get("generation_config"),
"use_fine_tune": llm_config.get("use_fine_tune"),
"fine_tune_zip": llm_config.get("fine_tune_zip"),
"user_input": chat_history[-1]["content"] if chat_history else "",
"system_prompt": prompt,
"context": chat_history
}
try:
log(f"🚀 Sending request to Spark for project: {project_name}")
response = requests.post(f"{spark_url}/generate", json=payload, headers=headers, timeout=60)
response.raise_for_status()
result = response.json()
log("✅ Spark response received successfully.")
return result.get("model_answer")
except Exception as e:
log(f"❌ Spark microservice error: {e}")
return None
|