def hello_firestore(event, context):
path_parts = context.resource.split('/documents/')[1].split('/')
collection_path = path_parts[0]
document_path = '/'.join(path_parts[1:])
affected_doc = client.collection(collection_path).document(document_path)
affected_doc.update({
u'checking':
"checing",
u'gotUrl':
event["value"]['fields']["url"]['stringValue']
})
file_url = event["value"]['fields']["url"]['stringValue']
fil = urllib.request.urlopen(file_url)
response = requests.get(file_url)
fine_name = os.path.join(tempfile.gettempdir(), "metadata.pdf")
with open(fine_name, 'wb') as f:
f.write(response.content)
text = textract.process(fine_name, method='pdfminer', encoding='ascii')
data = resumeparse.read_file(fine_name)
affected_doc.update({u'resumeData': data})
affected_doc.update(
{u'original': event["value"]['fields']["url"]['stringValue']})
def extract_fields(txt):
# import pdb;pdb.set_trace()
resume_fields = resumeparse.read_file(
'/home/ebabu/Downloads/Sanidhya_CV-converted.docx')
return resume_fields
# def main():
# text = extract_text_from_docx('/home/dell/Downloads/Sanidhya_CV-converted.docx')
# experience = extract_experience(text)
# main()
# import pdb;pdb.set_trace()
# for chunk in nltk.ne_chunk(nltk.pos_tag(nltk.word_tokenize(sent))):
# # chunk = list(chunk)
# if hasattr(chunk, 'label') and (chunk.label() == 'ORGANIZATION' or 'PERSON'):
# # projects_list.append(sent)
# # chunk_leave = chunk.leaves()
# # # import pdb;pdb.set_trace()
# # for word,tag in chunk_leave:
# # # projects_title = []
# # # projects_title.append(word)
# # projects_title = word
# # print("projects_title",projects_title)
# # project_sent = sent
# # projects_dict[projects_title] = project_sent
# # while count<=3:
# # print(chunk.label())
# # pass
# # count += 1
# # print("sent",sent," ")
# import pdb;pdb.set_trace()
def index():
if request.method == 'GET':
return "Hello"
content = request.json()
data = resumeparse.read_file('/content/Anubhab_Cover letter.pdf')
result =data
return jsonify( {'solution_text': result } )
#from pyresparser import ResumeParser
from resume_parser import resumeparse
if __name__ == '__main__':
#data = ResumeParser(r'CV_AliDoggaz.pdf').get_extracted_data()
data = resumeparse.read_file('CV_AliDoggaz.pdf')
print(data)
from resume_parser import resumeparse
import os
import warnings
warnings.simplefilter('ignore')
path = 'C:/Users/91884/Desktop/OCR_Resume/OCR_Resume/'
resumes = os.listdir(path)
c = 0
try:
for i in range(8, 10):
print(resumes[i])
data = resumeparse.read_file(f'{path}{resumes[i]}')
print(data['skills'])
c += 1
except:
pass
print(c)
print(len(resumes) + 1 - 9)
#8
# All Resume Parser Dependencies must be installed # Java must be running from resume_parser import resumeparse data = resumeparse.read_file(r"C:\Users\some_\Downloads\Profile.pdf") # PS: First Run takes a while
def home():
filename = request.args.get("filename")
data = resumeparse.read_file('pdfs/' + filename)
name = data['name'].lower().title()
email = data['email']
phone = data['phone']
try:
school = data['university'][0].title()
except:
school = ""
skills = []
count = 0
for x in data['skills']:
skills.append(x.strip())
if count == 5:
break
else:
count += 1
skills = list(filter(None, skills))
skillstr = ', '.join(skills)
matchings = []
listings = [[]]
best_jobs = []
concentration = []
organization, job_type, job_title, job_description, location = [], [], [], [], []
# Converting pdf to txt file for library
def pdf2txt(PDFfile, TXTfile):
in_file = open(PDFfile, 'rb')
res_mgr = PDFResourceManager()
data = io.StringIO()
TxtConverter = TextConverter(res_mgr, data, laparams=LAParams())
interpreter = PDFPageInterpreter(res_mgr, TxtConverter)
for page in PDFPage.get_pages(in_file):
interpreter.process_page(page)
txt = data.getvalue()
with open(TXTfile, 'w') as f:
f.write(txt)
# Function that searches for best match in jobs dataset
def find_matches(user_resume):
with open(user_resume, 'r') as resume:
with open('small_jobs_dataset.csv', 'r') as job_listings_csv:
# Splitting dataset rows by delimiter
csv_reader = csv.reader(job_listings_csv)
count = 0
# Reading user's resume into variable
resume_var = resume.read()
for row in csv_reader:
str_row = str(row)
job_as_list = pp.commaSeparatedList.parseString(
str_row).asList()
# Storing job description
job_desc = job_as_list[4]
if count > 0:
# Feature extraction on job desc and resume
text = [resume_var, job_desc]
count_vec = CountVectorizer()
count_matrix = count_vec.fit_transform(text)
match = cosine_similarity(count_matrix)[0][1] * 100
matchings.append(tuple((match, count)))
listings.append(job_as_list)
count += 1
# Sorting by jobs with highest match
matchings.sort(reverse=True)
# Storing jobs with highest match and user's concentration
for i in range(5):
match = matchings[i]
job = listings[int(match[1])]
split_string = job
organization.append(split_string[0].strip('\"'))
job_type.append(split_string[3].strip('\"'))
job_title.append(split_string[7].strip('\"'))
job_description.append(split_string[4].strip('\"'))
location.append(split_string[6].strip('\"'))
if i == 0:
job_industry = split_string[3]
job_industry = job_industry.strip('\"')
concentration.append(job_industry)
# Resume needs to be converted from pdf to txt
PDFfile = 'pdfs/' + filename
TXTfile = 'parsed_resume.txt'
# Converting pdf resume to txt
pdf2txt(PDFfile, TXTfile)
# Calling find_matches function
find_matches('parsed_resume.txt')
# Printing results
company_names = []
company_desc = []
company_match = []
for i in range(3):
try:
company_names.append(re.sub(r'[^A-Za-z0-9 ]+', '',
organization[i]))
company_desc.append(job_title[i].strip())
matchc = round(float(str(matchings[i][0]).strip()))
company_match.append(matchc)
except:
pass
try:
avg = 0
for x in company_match:
if x < 80:
avg += x + 20
else:
avg += x
avg = round(avg / 3)
except:
print('didnt work')
import os
os.remove("parsed_resume.txt")
return render_template("results.html",
name=name,
email=email,
phone=phone,
school=school,
skills=skillstr,
company_desc=company_desc,
company_match=company_match,
company_names=company_names,
avg=avg)
from resume_parser import resumeparse
# nltk.download('popular')
data = resumeparse.read_file(
'/home/ebabu/Downloads/SrishtiJain 2021-converted.docx')
print(data)
# from pyresparser import ResumeParser
# data = ResumeParser('/path/to/resume/file').get_extracted_data()
# resume_fields = ResumeParser('/home/ebabu/Downloads/Sanidhya_CV-converted.docx').get_extracted_data()
# print(resume_fields)
# This file conatins python code to extract skills from uploaded resume using natural language processing
import json
import re
import os
import sys
from resume_parser import resumeparse
from itertools import filterfalse
from nltk.corpus import stopwords
stopwords = set(stopwords.words('english'))
file_name = "./backend-engine/" + sys.argv[1]
data = resumeparse.read_file(file_name)
skills = data["skills"]
# remove all non-alphabet or non-numbers
# remove beginning and trailling white spaces
for i in range(0, len(skills)):
skills[i] = re.sub('[^0-9a-zA-Z\s]+', '', skills[i])
skills[i] = skills[i].strip()
# remove strings with length greater than 2
skills[:] = filterfalse(lambda elm: len(elm.split()) > 2, skills)
# remove any stopwords
for i in range(0, len(skills)):
skills[i] = ' '.join(
filter(lambda w: not w in stopwords, skills[i].split()))
# remove empty string
skills[:] = filterfalse(lambda elm: len(elm) == 0, skills)
def find_cv():
"""
:param: N/A
:return: The name of the file that is the most likely to be the candidate's resume/cv, as well as the candidate email
adress, his phone number, and his name.
"""
#Special case where candidate's folder is empty or contains only 1 file
if len(os.listdir('PDF_Converted_Files')) == 0:
return
if len(os.listdir('PDF_Converted_Files')) == 1:
return os.listdir('PDF_Converted_Files')[0]
# If a file contains the words CV/Resume/etc..., returns directly that file name
for name in os.listdir('PDF_Converted_Files'):
if name.lower().startswith("cv"):
data = resumeparse.read_file('PDF_Converted_Files' + os.sep + name)
return name, data['email'], data['phone'], data['name']
for keyword in [
'cv.', 'resume', 'résumé', 'curriculum vitae',
'curriculumvitae'
]:
if keyword in name.lower():
data = resumeparse.read_file('PDF_Converted_Files' + os.sep +
name)
return name, data['email'], data['phone'], data['name']
# Attribute a score to each file. The score will allow us to estimate the probability of
# that file being the candidate's resume.
# "maxi" will store the highest score reached yet.
maxi = 0
# loop over all files and attribute a "score" to each one.
# If the file's score is >= to maxi, maxi = score. In this case,
# we will also store the email, phone number, and fullname present in the file.
for name in os.listdir('PDF_Converted_Files'):
# Parse the file, looking for relevant info (email, skills, education, etc...)
score = 0
data = resumeparse.read_file('PDF_Converted_Files' + os.sep + name)
# Increase score if we find relevant info in the file
if data['skills']:
score += 5 # If the file contains the candidate's skills, there are very high chances that this
# file is the candidate's resume. So we increase its score by 5.
if data['email']:
score += 1
if data['phone']:
score += 1
if data['degree']:
score += 3
if data['university']:
score += 2
if data['total_exp']:
score += 2
if score >= maxi:
cv_name, email, phone, FullName = name, data['email'], data[
'phone'], data['name']
maxi = score
# Return the file with the highest score (highest chances of being the candidate's resume)
return cv_name, email, phone, FullName