File size: 2,324 Bytes
c5b826e 15a5228 c5b826e 15a5228 c5b826e | 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 | import sys
import os
import copy
import uvicorn
import socket
import logging
import datetime
from models.quotes_search_engine import QuoteSearchEngine
from models.data_reader import load_quotes_from_csv
from models.Query import Query, Query_Multiple, SearchResponse, SimilarQuote, QuoteVector, VectorResponse
from decouple import config
from fastapi import FastAPI, HTTPException, Depends, Body
from sentence_transformers import SentenceTransformer
# quotes_path = r"C:\Users\jov2bg\Desktop\QuotesSearch\search_engine\data\quotes.csv"
quotes_path = "hf://datasets/jstet/quotes-500k/quotes.csv" # path on hugging face
app = FastAPI(title="Search Prompt Engine", description="API for prompt search", version="1.0")
quotes = load_quotes_from_csv(quotes_path)
search_engine = QuoteSearchEngine()
search_engine.add_quotes_to_vector_database(quotes[:10000])
@app.get("/")
def read_root():
return {"message": "Quote Search Engine is running!"}
@app.post("/search/")
async def search_prompts(query: Query, k: int = 3):
print(f'Prompt: {query}')
similar_quotes, distances = search_engine.most_similar(query.quote, top_k=k)
print(f'Similar Quotes: {similar_quotes}')
print(f'Cosine similarity: {distances}')
print(40*'****')
# Format the response
response = [
SimilarQuote(prompt=prompt, distance=float(distance))
for prompt, distance in zip(similar_quotes, distances)
]
return SearchResponse(results=response)
@app.post("/all_vectors_similarities/")
async def all_vectors(query: Query):
query_embedding = search_engine.model.encode([query.quote]) # Encode the prompt to a vector
all_similarities = search_engine.cosine_similarity(query_embedding, search_engine.index)
print(f'Quote or phrase: {query}')
print(f'All Vector Similarities: {all_similarities}')
print(40*'****')
response = [
QuoteVector(vector=index, distance=float(distance))
for index, distance in enumerate(all_similarities)
]
return VectorResponse(results=response)
if __name__ == "__main__":
# Server Config
# SERVER_HOST_IP = socket.gethostbyname(socket.gethostname())
SERVER_HOST_IP = socket.gethostbyname("localhost") # for local deployment
SERVER_PORT = int(8084)
uvicorn.run(app, host=SERVER_HOST_IP, port=SERVER_PORT) |