Spaces:
Sleeping
Sleeping
File size: 1,904 Bytes
8f0183f |
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 66 67 68 69 70 71 72 73 74 |
from fastapi import FastAPI
from pydantic import BaseModel
from typing import List
from sentence_transformers import SentenceTransformer
from sklearn.metrics.pairwise import cosine_similarity
import numpy as np
# -----------------------
# Load model ONCE (important for performance)
# -----------------------
model = SentenceTransformer("all-MiniLM-L6-v2")
app = FastAPI(
title="CV Matching API",
description="Rank CVs against a Job Description using BERT embeddings",
version="1.0"
)
# -----------------------
# Request schema
# -----------------------
class MatchRequest(BaseModel):
job_description: str
cvs: List[str]
# -----------------------
# Response schema
# -----------------------
class CVScore(BaseModel):
cv_text: str
relevance_score: float
class MatchResponse(BaseModel):
results: List[CVScore]
# -----------------------
# Utility: text cleaning (optional but recommended)
# -----------------------
def clean_text(text: str) -> str:
return text.replace("\n", " ").strip().lower()
# -----------------------
# API endpoint
# -----------------------
@app.post("/match", response_model=MatchResponse)
def match_cvs(request: MatchRequest):
# Clean input
jd = clean_text(request.job_description)
cvs = [clean_text(cv) for cv in request.cvs]
# Embed job description
jd_embedding = model.encode([jd])
# Embed CVs
cv_embeddings = model.encode(cvs)
# Compute cosine similarity
scores = cosine_similarity(jd_embedding, cv_embeddings)[0]
# Build response
results = []
for cv_text, score in zip(request.cvs, scores):
results.append(
CVScore(
cv_text=cv_text,
relevance_score=float(score)
)
)
# Sort by relevance (descending)
results.sort(key=lambda x: x.relevance_score, reverse=True)
return MatchResponse(results=results)
|