def extract_skills_case_sensitive(resume_text, items_of_interest):
potential_skills_dict = dict()
matched_skills = set()
for skill_input in items_of_interest:
if type(skill_input) is not str and len(skill_input) >= 1:
potential_skills_dict[skill_input[0]] = skill_input
elif type(skill_input) is str:
potential_skills_dict[skill_input] = [skill_input]
else:
pass
#logging.warning('Unknown skill listing type: {}.'.format(skill_input))
for (skill_name, skill_alias_list) in potential_skills_dict.items():
skill_matches = 0
# TODO incorporate word2vec here?
for skill_alias in skill_alias_list:
skill_matches += lib.term_count(resume_text.replace('-', ' ').replace(':', '').replace(',', '').replace('\'', ''), skill_alias.lower()) # add the # of matches for each alias
if skill_matches > 0:
matched_skills.add(skill_name.replace('\x20', ''))
if len(matched_skills) == 0:
matched_skills = ''
return list(matched_skills)
def extract_skills(resume_text, extractor, items_of_interest):
potential_skills_dict = dict()
matched_skills = set()
# TODO This skill input formatting could happen once per run, instead of once per observation.
for skill_input in items_of_interest:
# Format list inputs
if type(skill_input) is list and len(skill_input) >= 1:
potential_skills_dict[skill_input[0]] = skill_input
# Format string inputs
elif type(skill_input) is str:
potential_skills_dict[skill_input] = [skill_input]
else:
logging.warn(
'Unknown skill listing type: {}. Please format as either a single string or a list of strings'
''.format(skill_input))
for (skill_name, skill_alias_list) in potential_skills_dict.items():
skill_matches = 0
# Iterate through aliases
for skill_alias in skill_alias_list:
# Add the number of matches for each alias
skill_matches += lib.term_count(resume_text, skill_alias.lower())
# If at least one alias is found, add skill name to set of skills
if skill_matches > 0:
matched_skills.add(skill_name)
return matched_skills
def extract_skills_case_agnostic(resume_text, items_of_interest):
potential_skills_dict = dict()
matched_skills = set()
for skill_input in items_of_interest:
# Format list of strings inputs
if type(skill_input) is not str and len(skill_input) >= 1:
potential_skills_dict[skill_input[0]] = skill_input
# Format string inputs
if type(skill_input) is str:
potential_skills_dict[skill_input] = [skill_input]
else:
pass
#logging.warning('Unknown skill listing type: {}. Please format as a string or a list of strings'.format(skill_input))
for (skill_name, skill_alias_list) in potential_skills_dict.items():
skill_matches = 0
# iterate through each string in the list of equivalent words (i.e. a line in the yaml file)
# TODO incorporate word2vec here?
for skill_alias in skill_alias_list:
skill_matches += lib.term_count(resume_text.replace('-', ' ').replace(':', '').replace(',', '').replace('\'', ''), skill_alias.lower()) # add the # of matches for each alias
if skill_matches > 0: # if at least one alias is found, add skill name to set of skills
matched_skills.add(skill_name.replace('\x20', ''))
if len(matched_skills) == 0: # so it doesn't save 'set()' in the csv when it's empty
matched_skills = ''
return list(matched_skills)
def extract_universities(resume_text):
# Reference variables
matched_universities = set()
normalized_resume_text = ' '.join(simple_preprocess(resume_text))
# Iterate through possible universities
for university in lib.get_conf('universities'):
university = ' '.join(simple_preprocess(university))
university_count = lib.term_count(normalized_resume_text, university)
if university_count > 0:
matched_universities.add(university)
return matched_universities