def build_model():
method = request.args.get('method')
if method == 'build':
connection = MongoClient('mongodb+srv://hrlanes-mongodb-reader:hrlanes%[email protected]', 27017)
db = connection['hrlanes-web-db']
data = db['users']
ex = data.find({"$and": [{'ProfileSummaryInfo': {"$exists": True}}, {'recommenderProcessed': {"$exists": True}}, {'recommenderProcessed': True }]})
helper = Helper()
d = helper.createDictionary(ex)
resumeList = []
for key in d:
if len(d[key])>0: # check if resumes/details exist
doc_included = []
for x in d[key]:
resumeList.append(x[1])
doc_included.append(x[0])
documents = []
for f in resumeList:
documents.append(helper.cleanTextAndTokenize(f))
helper.create_tfidf(str(key), documents, doc_included)
return " "
elif method =='recommend':
exp = request.args.get('e')
farea = request.args.get('f')
jd = request.args.get('jd')
if exp and farea and jd:
helper = Helper()
'''with open (jd, 'r') as f:
jobd = f.read()
#for local storage path
jobd = helper.extract_text_from_url(jd)
#for extracting text from pdf url -> from blob storage
'''
jobd = str(jd)
preprocessed = helper.cleanTextAndTokenize(jobd) #tokenizing text
sim_scores = helper.recommend(exp, farea, preprocessed) #returning candidate IDs
if len(sim_scores)==0:
return ("Sorry! No matching candidates!")
response = app.response_class(
response=json.dumps(str(dict(sim_scores))),
status=200,
mimetype='application/json')
return response
else:
return "Please enter exp, f area and jd in the request body!"
else:
return "Please enter the method in request body: build or recommend!"
def build_model():
method = request.args.get('method')
if method == 'build':
helper = Helper()
if not os.path.exists('active'):
os.makedirs('active')
else:
helper.create_backup()
connection = MongoClient(
'mongodb+srv://hrlanes-mongodb-reader:hrlanes%[email protected]',
27017)
db = connection['hrlanes-web-db']
data = db['users']
ex = data.find({
"$and": [{
'ProfileSummaryInfo': {
"$exists": True
}
}, {
'recommenderProcessed': {
"$exists": True
}
}, {
'recommenderProcessed': True
}]
})
d = helper.createDictionary(ex)
path = os.getcwd() + "\\active\\"
with open(path + "dictionary.pkl", "wb") as output:
pickle.dump(d, output)
resumeList = []
for key in d:
if len(d[key]) > 0: # check if resumes/details exist
doc_included = []
for x in d[key]:
resumeList.append(x[1])
doc_included.append(x[0])
documents = []
for f in resumeList:
documents.append(f)
helper.create_tfidf(str(key), documents, doc_included)
#reset recommenderProcessed to false
'''filter = {"$and": [{'ProfileSummaryInfo': {"$exists": True}}, {'recommenderProcessed': {"$exists": True}}, {'recommenderProcessed': True }]}
data.update_many(filter, {"$set": { "recommenderProcessed": False }})
'''
return 'okay'
elif method == 'recommend':
exp = request.args.get('e')
farea = request.args.get('f')
jd = request.args.get('jd')
if exp and farea and jd:
helper = Helper()
jobd = helper.extract_text_from_url(
jd) #for extracting text from pdf url -> from blob storage
jobd = str(jd)
preprocessed = helper.cleanTextAndTokenize(jobd) #tokenizing text
sim_scores = helper.recommend(
exp, farea, preprocessed) #returning candidate IDs
if len(sim_scores) == 0:
return ("Sorry! No matching candidates!")
response = app.response_class(response=json.dumps(
str(dict(sim_scores))),
status=200,
mimetype='application/json')
return response
else:
return "Please enter exp, f area and jd in the request body!"
else:
return "Please enter the method in request body: build or recommend!"
