def getPrecision(project_id):
# load collected eids
eids = eids_service.load_eid_list(project_id)
relevance_measure = relevance_measure_service.load_relevance_measure(
project_id)
if relevance_measure is None:
relevance_measure = RelevanceMeasure()
relevance_measure.number_of_search_results = len(eids)
judgement_list = eids_service.load_judgement_file(project_id)
app.logger.info('project {}: loaded judgements'.format(project_id))
relevance_measure.number_sample_entries = len(judgement_list)
relevance_measure.number_positive_sample_entries = 0
for judgement in judgement_list:
if judgement.isRelevant:
relevance_measure.number_positive_sample_entries = \
relevance_measure.number_positive_sample_entries + 1
if relevance_measure.number_sample_entries > 0:
relevance_measure.precision = relevance_measure.number_positive_sample_entries / relevance_measure.number_sample_entries
else:
relevance_measure.precision = 0
relevance_measure_service.save_relevance_measures(project_id,
relevance_measure)
app.logger.info(
'project {}: calculated relevance measure precision'.format(
project_id))
return jsonify(relevance_measure)
def check_test_eids(project_id):
"""
calcluates the Recall by comparing the list of EIDs retrieved from the query against lists of EIDs as obtained from
the survey
:param project_id: the ID of the current project
:return: a JSON formatted relevance measure object.
"""
test_eids = eids_service.load_eid_list(project_id, 'test_')
app.logger.info('project {}: loaded test eids'.format(project_id))
# load collected eids
eids = eids_service.load_eid_list(project_id)
relevance_measure = relevance_measure_service.load_relevance_measure(
project_id)
if relevance_measure is None:
relevance_measure = RelevanceMeasure()
relevance_measure.number_of_search_results = len(eids)
relevance_measure.number_test_entries = len(test_eids)
relevance_measure.number_test_entries_found = 0
for test_eid in test_eids:
if test_eid in eids:
relevance_measure.number_test_entries_found = relevance_measure.number_test_entries_found + 1
if relevance_measure.number_of_search_results > 0:
relevance_measure.recall = relevance_measure.number_test_entries_found / relevance_measure.number_test_entries
else:
relevance_measure.recall = 0
relevance_measure_service.save_relevance_measures(project_id,
relevance_measure)
app.logger.info(
'project {}: calculated relevance measure recall'.format(project_id))
return jsonify(relevance_measure)
def check_sample_eids(project_id):
# load collected eids
eids = eids_service.load_eid_list(project_id)
relevance_measure = relevance_measure_service.load_relevance_measure(
project_id)
if relevance_measure is None:
relevance_measure = RelevanceMeasure()
relevance_measure.number_of_search_results = len(eids)
judgement_list = eids_service.load_judgement_file(project_id)
relevance_measure['number_sample_entries'] = len(judgement_list)
for judgement in judgement_list:
if judgement['isRelevant']:
relevance_measure['number_positive_sample_entries'] = \
relevance_measure['number_positive_sample_entries'] + 1
relevance_measure_service.save_relevance_measures(project_id,
relevance_measure)
return jsonify(relevance_measure.__dict__)
def load_relevance_measure(project_id):
with app.app_context():
location = app.config.get("LIBINTEL_DATA_DIR")
path_to_file = location + '/out/' + project_id + '/relevance_measures.json'
with open(path_to_file) as json_file:
try:
relevance_measure = json.load(json_file)
json_file.close()
return RelevanceMeasure(**relevance_measure)
except FileNotFoundError:
return {}
except JSONDecodeError:
return {}
def check_test_eids(project_id):
# load test eids
test_eids = eids_service.load_eid_list(project_id, 'test_')
# load collected eids
eids = eids_service.load_eid_list(project_id)
relevance_measure = relevance_measure_service.load_relevance_measure(
project_id)
if relevance_measure is None:
relevance_measure = RelevanceMeasure()
relevance_measure.number_of_search_results = len(eids)
relevance_measure.number_test_entries = len(test_eids)
for test_eid in test_eids:
if test_eid in eids:
relevance_measure.number_test_entries_found = relevance_measure.number_test_entries_found + 1
if relevance_measure.number_of_search_results > 0:
relevance_measure.recall = relevance_measure.number_test_entries_found / relevance_measure.number_test_entries
else:
relevance_measure.recall = 0
relevance_measure_service.save_relevance_measures(project_id,
relevance_measure)
return jsonify(relevance_measure.__dict__)
def query_execution(project_id):
"""
executes the defined and saved query in scopus
:param project_id: the ID of the current project
:return: 'finished' with a status of 204 when the query was executed successfully
"""
app.logger.info('project {}: running queries'.format(project_id))
# reads the saved Scopus search string from disk
scopus_queries = query_service.load_scopus_queries(project_id)
# retrieve the project from disk, set the booleans and save the project
project = project_service.load_project(project_id)
project.isEidsCollected = False
project.isEidsCollecting = True
project_service.save_project(project)
# prepares the status file
status = Status("EIDS_COLLECTING")
status_service.save_status(project_id, status)
# prepare EIDs list
eids = []
for index, search_strings in enumerate(scopus_queries.search_strings):
individual_eids = []
for search_string in search_strings:
app.logger.info('project {}: executing search {} - {}'.format(
project_id, index, search_string))
search = scopus.ScopusSearch(search_string,
refresh=True,
field='eid',
view='STANDARD')
if search.results is not None:
app.logger.info(
'project {}: result search {} - {} entries found'.format(
project_id, index, len(search.results)))
for result in search.results:
# add EID if it is not already in the list (from a former search)
eids.append(result.eid)
individual_eids.append(result.eid)
eids_service.save_eid_list(
project_id=project_id,
eids=set(individual_eids),
prefix=(str(scopus_queries.search_ids[index]) + '_'))
# convert to set in order to remove duplicates
eids = set(eids)
# print the results to the command line for logging
app.logger.info('project {}: found {} eids in Scopus'.format(
project_id, len(eids)))
# persist EIDs to file
eids_service.save_eid_list(project_id=project_id, eids=eids)
# set the total number of results to the relevance_measures measure save it to disk
relevance_measure = RelevanceMeasure(
number_of_search_results=eids.__len__())
relevance_measure_service.save_relevance_measures(project_id,
relevance_measure)
# set the total number of results to the status save it to disk
status.total = relevance_measure.number_of_search_results
status_service.save_status(project_id, status)
# set the status and save it to disk
status = Status("EIDS_COLLECTED")
status_service.save_status(project_id, status)
# set the project boolean and save the project
project.isEidslist = True
project.isEidsCollected = True
project.isEidsCollecting = False
project_service.save_project(project)
return Response({"status": "FINISHED"}, status=204)