def __init__(self, schema_='metrics'):
self.logger = setup_logging('AdsDataSqlSync', 'INFO')
self.schema = schema_
self.table = models.MetricsTable()
self.table.schema = self.schema
# used to buffer writes
self.upserts = []
self.tmp_update_buffer = []
self.tmp_count = 0
self.config = {}
self.config.update(load_config())
def __init__(self, schema_='metrics'):
# - Use app logger:
#import logging
#logger = logging.getLogger('ads-data')
# - Or individual logger for this file:
proj_home = os.path.realpath(
os.path.join(os.path.dirname(__file__), '../'))
self.config = load_config(proj_home=proj_home)
self.logger = setup_logging(
__name__,
proj_home=proj_home,
level=self.config.get('LOGGING_LEVEL', 'INFO'),
attach_stdout=self.config.get('LOG_STDOUT', False))
self.schema = schema_
self.table = models.MetricsTable()
self.table.schema = self.schema
# used to buffer writes
self.upserts = []
self.tmp_update_buffer = []
self.tmp_count = 0
def create_metrics_table(self, db_engine):
db_engine.execute(CreateSchema(self.schema))
table = models.MetricsTable()
table.__table__.schema = self.schema
table.__table__.create(db_engine)
self.logger.info('metrics.py, metrics table created')
def row_view_to_metrics(self,
passed_row_view,
nonbib_db_conn,
row_view_schema='nonbib',
m=None):
"""convert the passed row view into a complete metrics dictionary"""
if m is None:
m = models.MetricsTable()
# first do easy fields
bibcode = passed_row_view.bibcode
m.bibcode = bibcode
m.refereed = passed_row_view.refereed
m.citations = passed_row_view.citations
m.reads = passed_row_view.reads
m.downloads = passed_row_view.downloads
m.citation_num = len(
passed_row_view.citations) if passed_row_view.citations else 0
m.author_num = max(len(passed_row_view.authors),
1) if passed_row_view.authors else 1
m.reference_num = len(
passed_row_view.reference) if passed_row_view.reference else 0
#metrics_dict['citation_num'] = len(passed_row_view.get('citations', [])
#metrics_dict['author_num'] = max(len(passed_row_view.get('authors'),[]),1)
#metrics_dict['reference_num'] = len(passed_row_view.get('reference'),[])
# next deal with papers that cite the current one
# compute histogram, normalized values of citations
# and create list of refereed citations
citations = passed_row_view.citations
normalized_reference = 0.0
citations_json_records = []
refereed_citations = []
citations_histogram = defaultdict(float)
total_normalized_citations = 0.0
if citations:
q = 'select refereed,array_length(reference,1),bibcode from ' + row_view_schema + \
'.RowViewM where bibcode in (select unnest(citations) from ' + row_view_schema + \
'.RowViewM where bibcode=%s);'
result = nonbib_db_conn.execute(q, bibcode)
for row in result:
citation_refereed = row[0] if row[0] else False
citation_refereed = citation_refereed in (True, 't', 'true')
len_citation_reference = int(row[1]) if row[1] else 0
citation_bibcode = row[2]
citation_normalized_references = 1.0 / float(
max(5, len_citation_reference))
total_normalized_citations += citation_normalized_references
normalized_reference += citation_normalized_references
tmp_json = {
"bibcode": citation_bibcode.encode('utf-8'),
"ref_norm": citation_normalized_references,
"auth_norm": 1.0 / m.author_num,
"pubyear": int(bibcode[:4]),
"cityear": int(citation_bibcode[:4])
}
citations_json_records.append(tmp_json)
if (citation_refereed):
refereed_citations.append(citation_bibcode)
citations_histogram[
citation_bibcode[:4]] += total_normalized_citations
m.refereed_citations = refereed_citations
m.refereed_citation_num = len(refereed_citations)
# annual citations
today = datetime.today()
resource_age = max(1.0, today.year - int(bibcode[:4]) + 1)
m.an_citations = float(m.citation_num) / float(resource_age)
m.an_refereed_citations = float(
m.refereed_citation_num) / float(resource_age)
# normalized info
m.rn_citations = normalized_reference # total_normalized_citations
m.rn_citation_data = citations_json_records
m.rn_citations_hist = dict(citations_histogram)
m.modtime = datetime.now().strftime('%Y-%m-%d %H:%M:%S')
return m