File size: 2,110 Bytes
d22d262 e0e8487 d22d262 7e70df5 9a6af66 |
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 |
from sentence_transformers import SentenceTransformer
from sklearn.metrics.pairwise import cosine_similarity
import numpy as np
import pandas as pd
import pickle
st = SentenceTransformer('all-mpnet-base-v2')
filename = 'svc.pkl'
with open(filename, 'rb') as file:
model = pickle.load(file)
# role_req-exp 0.341522
# role_pos 0.350747
# major_similarity 0.846268
# skill_similarity 0.774542
# score 0.986356
# cv = {
# "experiences": str(body.cv.experiences),
# "positions": str(positions),
# "userMajors": str(userMajors),
# "skills": str(body.cv.skills),
# "yoe": yoe
# }
# job = {
# "jobDesc": body.job.jobDesc,
# "role": body.job.role,
# "majors": str(body.job.majors),
# "skills": str(body.job.skills),
# "minYoE": body.job.minYoE
# }
def predict(cv, job):
diffYoe = cv['yoe'] - job['minYoE']
results = {}
role_req_exp = cosine_similarity(st.encode(cv['experiences']).reshape(1,-1), st.encode(job['role']+'\n'+job['jobDesc']).reshape(1,-1))[0][0] if cv['experiences'] != '[]' else 0
role_pos = cosine_similarity(st.encode(cv['positions']).reshape(1,-1), st.encode(job['role']).reshape(1,-1))[0][0] if cv['positions'] != '[]' else 0
major_similarity = cosine_similarity(st.encode(cv['userMajors']).reshape(1,-1), st.encode(job['majors']).reshape(1,-1))[0][0] if cv['userMajors'] != '[]' else 0
skill_similarity = cosine_similarity(st.encode(cv['skills']).reshape(1,-1), st.encode(job['skills']).reshape(1,-1))[0][0] if cv['skills'] != '[]' else 0
score_yoe = 0.5 if diffYoe == -1 else (0 if diffYoe < 0 else 1)
score = 0.35 * role_req_exp + 0.1 * role_pos + 0.15 * major_similarity + 0.3* score_yoe + 0.1 * skill_similarity
data = [{
'role_req-exp': role_req_exp,
'role_pos': role_pos,
'major_similarity': major_similarity,
'skill_similarity': skill_similarity,
'score': score
}]
X = pd.DataFrame.from_dict(data)
res = model.predict(X)
results['score'] = model.predict_proba(X)[:, 1]
results['is_accepted'] = res[0]
return results |