matteo/wikimedica-disease-search
matteo
/
wikimedica-disease-search
1
0
Fork 0
Code Issues 2 Pull Requests Releases Activity

pageviews: N’utilise pas les données du jour présent

This commit is contained in:
Mattéo Delabre 2019-12-04 21:05:18 -05:00
parent 1767738167
commit 9734bb401f
Signed by: matteo
GPG Key ID: AE3FBD02DC583ABB
1 changed files with 6 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

9
data/fetch/wikipedia_pageviews.py
View File

@ -1,5 +1,5 @@
import collections import collections
from datetime import datetime from datetime import datetime, timedelta
from .http import session from .http import session
import numpy import numpy
from scipy import stats from scipy import stats
@ -27,6 +27,9 @@ pageviews_date_format = '%Y%m%d'
# Date de première disponibilité des pageviews sur l’API Wikimedia # Date de première disponibilité des pageviews sur l’API Wikimedia
pageviews_first_data = datetime(2015, 7, 1) pageviews_first_data = datetime(2015, 7, 1)
# Date de dernière disponibilité des pageviews sur l’API Wikimedia
pageviews_last_data = datetime.today() - timedelta(days=1)
# Tableau contenant tous les jours de l’année de 1 à 365 # Tableau contenant tous les jours de l’année de 1 à 365
year_all_days = numpy.arange(1, 366) year_all_days = numpy.arange(1, 366)
@ -88,7 +91,7 @@ def get_aggregate(project):
agent='user', agent='user',
granularity='daily', granularity='daily',
start=pageviews_first_data.strftime(pageviews_date_format), start=pageviews_first_data.strftime(pageviews_date_format),
end=datetime.today().strftime(pageviews_date_format) end=pageviews_last_data.strftime(pageviews_date_format)
)) ))
data = res.json() data = res.json()
@ -124,7 +127,7 @@ def get_article(project, article):
agent='user', agent='user',
granularity='daily', granularity='daily',
start=pageviews_first_data.strftime(pageviews_date_format), start=pageviews_first_data.strftime(pageviews_date_format),
end=datetime.today().strftime(pageviews_date_format) end=pageviews_last_data.strftime(pageviews_date_format)
)) ))
data = res.json() data = res.json()