Spaces:
Sleeping
Sleeping
added ai request, ai runner
Browse files- app.py +21 -1
- models/ai_request.py +6 -0
- requirements.txt +5 -0
- utils/ai_runner.py +35 -0
app.py
CHANGED
|
@@ -1,7 +1,27 @@
|
|
| 1 |
-
from fastapi import FastAPI
|
|
|
|
| 2 |
|
|
|
|
| 3 |
app = FastAPI()
|
| 4 |
|
| 5 |
@app.get("/")
|
| 6 |
def greet_json():
|
| 7 |
return {"Hello": "World!"}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
from fastapi import FastAPI, HTTPException
|
| 2 |
+
from utils.ai_runner import run_llm
|
| 3 |
|
| 4 |
+
from models.ai_request import AIRequest
|
| 5 |
app = FastAPI()
|
| 6 |
|
| 7 |
@app.get("/")
|
| 8 |
def greet_json():
|
| 9 |
return {"Hello": "World!"}
|
| 10 |
+
|
| 11 |
+
@app.post("/api/v1/ai")
|
| 12 |
+
async def ai_json(request: AIRequest):
|
| 13 |
+
try:
|
| 14 |
+
response = await run_llm(request.text,request.model_name)
|
| 15 |
+
|
| 16 |
+
if not response:
|
| 17 |
+
return HTTPException(status_code=400,detail="Couldn't retrive response on server")
|
| 18 |
+
|
| 19 |
+
return {
|
| 20 |
+
"data":
|
| 21 |
+
{
|
| 22 |
+
"answer": response["answer"]
|
| 23 |
+
}
|
| 24 |
+
}
|
| 25 |
+
except Exception as e:
|
| 26 |
+
print(e)
|
| 27 |
+
return HTTPException(status_code=500,detail="Internal server error occured")
|
models/ai_request.py
ADDED
|
@@ -0,0 +1,6 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
from pydantic import BaseModel
|
| 2 |
+
|
| 3 |
+
|
| 4 |
+
class AIRequest(BaseModel):
|
| 5 |
+
text: str
|
| 6 |
+
model_name: str
|
requirements.txt
CHANGED
|
@@ -1,2 +1,7 @@
|
|
| 1 |
fastapi
|
| 2 |
uvicorn[standard]
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
fastapi
|
| 2 |
uvicorn[standard]
|
| 3 |
+
langchain
|
| 4 |
+
langchain_google_genai
|
| 5 |
+
langchain_groq
|
| 6 |
+
fastembed
|
| 7 |
+
python_dotenv
|
utils/ai_runner.py
ADDED
|
@@ -0,0 +1,35 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
from typing import Any, Dict, List
|
| 2 |
+
from langchain.chains import ConversationalRetrievalChain
|
| 3 |
+
from langchain_google_genai import ChatGoogleGenerativeAI, GoogleGenerativeAIEmbeddings
|
| 4 |
+
from langchain_groq import ChatGroq
|
| 5 |
+
from dotenv import load_dotenv
|
| 6 |
+
|
| 7 |
+
load_dotenv()
|
| 8 |
+
async def run_llm(query: str,model_name: str= "gemini-1.5-flash", chat_history: List[Dict[str, Any]] = [],):
|
| 9 |
+
embeddings = FastEmbedEmbeddings(model_name="BAAI/bge-base-en-v1.5")
|
| 10 |
+
docsearch = PineconeVectorStore(embedding=embeddings, index_name=INDEX_NAME)
|
| 11 |
+
chat = ChatGoogleGenerativeAI(model=model_name,)
|
| 12 |
+
# chat = ChatGroq(model_name="llama-3.3-70b-versatile",
|
| 13 |
+
# )
|
| 14 |
+
# temperature=1,
|
| 15 |
+
|
| 16 |
+
pt = f"""
|
| 17 |
+
تو یک ربات هستی که درمورد سرور مونگو کرفت میدانی.
|
| 18 |
+
مونگوکرفت یک سرور ماینکرفت سروایول است
|
| 19 |
+
|
| 20 |
+
اطلاعات تو تا تاریخ 24/01/2025 به روز شده است.
|
| 21 |
+
|
| 22 |
+
با توجه به سوال کاربر اطلاعات مورد نظر را دریافت و به صورت فرمت شده جواب بده.
|
| 23 |
+
|
| 24 |
+
**تو درمورد چیزهایی غیر از مونگو کرفت، مطالب بد هیچ چیزی نمیدانی و جوابی نمیدهی**
|
| 25 |
+
**به فارسی پاسخ بده**
|
| 26 |
+
|
| 27 |
+
با لحنی دوستانه و ساده بنویس.
|
| 28 |
+
|
| 29 |
+
سوال: {query}
|
| 30 |
+
"""
|
| 31 |
+
|
| 32 |
+
qa = ConversationalRetrievalChain.from_llm(
|
| 33 |
+
llm=chat, retriever=docsearch.as_retriever(), return_source_documents=True,
|
| 34 |
+
)
|
| 35 |
+
return qa.invoke({"question": pt, "chat_history": chat_history})
|