def __init__(self,year,activation=5,overall=False):
self.year = year
'''The year the cohort started.
'''
self.activation = activation
'''Minimum number of edits per month to be included in the cohort
'''
self.overall = overall
ts,ts_i = utils.create_time_stamps_month(fromym='%s01'%self.year,toym='201012')
self.time_stamps = ts
self.time_stamps_index = ts_i
'''Only take time_stamps starting with self.year
'''
self.cohorts = [self.year,'others']
'''Cohort definition
'''
self.cohort_labels = ['%s cohort'% i for i in self.cohorts]
'''Cohort labels
'''
if self.overall:
self.sqlQuery = "SELECT * FROM fabian WHERE user_id IN (SELECT user_id FROM fabian WHERE first_edit_year=%s GROUP BY user_id HAVING SUM(add_edits)>%s);"%(self.year,self.activation)
else:
self.sqlQuery = 'SELECT * FROM fabian WHERE add_edits > %s AND first_edit_year=%s'%(self.activation,self.year)
Cohort.__init__(self)
def readConfig(configfile):
'''
Reads the configuration from `ConfigParser` instance into the runtime settings.
:arg configfile: A file that can be read by a `ConfigParser` instance
'''
global language, filterbots, time_stamps, time_stamps_index, botfile, basedirectory, datadirectory, userlistdirectory, reportdirectory, wikipridedirectory, sqlhost, sqlwikidb, sqluserdb, sqlconfigfile, sqldroptables
import os
import ConfigParser
config = ConfigParser.ConfigParser()
config.read(configfile)
# general settings
language = config.get('General', 'language')
if len(language) != 2:
raise Exception('Language code should be two characters (%s)' %
language)
filterbots = config.get('General', 'filterbots')
startYM = config.get('General', 'startYM')
endYM = config.get('General', 'endYM')
(time_stamps,
time_stamps_index) = utils.create_time_stamps_month(fromym=startYM,
toym=endYM)
# directories
basedirectory = os.path.expanduser(
config.get('Directories', 'basedirectory'))
datadirectory = os.path.expanduser(
config.get('Directories', 'datadirectory'))
userlistdirectory = os.path.expanduser(
config.get('Directories', 'userlistdirectory'))
reportdirectory = os.path.expanduser(
config.get('Directories', 'reportdirectory'))
wikipridedirectory = os.path.expanduser(
config.get('Directories', 'wikipridedirectory'))
# if not os.path.isdir(basedirectory):
# import errno
# try:
# os.makedirs(p)
# except OSError as exc: # Python >2.5
# if exc.errno == errno.EEXIST:
# pass
# else: raise
# sql
sqlhost = config.get('MySQL', 'sqlhost')
sqlwikidb = config.get('MySQL', 'sqlwikidb')
sqluserdb = config.get('MySQL', 'sqluserdb')
sqlconfigfile = config.get('MySQL', 'sqlconfigfile')
sqldroptables = config.getboolean('MySQL', 'sqldroptables')
sqlconfigfile = os.path.expanduser(sqlconfigfile)
if not os.path.isfile(sqlconfigfile):
logger.warning("%s sql config file not found" % sqlconfigfile)
def __init__(self):
ts_month,ts_month_index = utils.create_time_stamps_month()
self.cohorts = [int(i) for i in range(0,len(ts_month))]
'''Cohort definition
'''
self.cohort_labels = ['%s days old'% i for i in self.cohorts]
'''Cohort labels
'''
Age.__init__(self)
def __init__(self):
ts_month, ts_month_index = utils.create_time_stamps_month()
self.cohorts = [int(i) for i in range(0, len(ts_month))]
'''Cohort definition
'''
self.cohort_labels = ['%s days old' % i for i in self.cohorts]
'''Cohort labels
'''
Age.__init__(self)
def mobile_pageviews_target(min_ts, max_ts):
return
# input files
metricfile = "../datafiles/rc_page_requests_mobile.csv"
targetfile = "../misc/targets/mobile_target.csv"
# time range
tss, tss_i = utils.create_time_stamps_month(fromym=min_ts, toym=max_ts)
target = extractData(targetfile, "\t", "MMM-YY", [("Target", 3)])
actual = extractData(metricfile, ",", "YYYY/MM/DD", [("Total", 1)])
def readConfig(configfile):
'''
Reads the configuration from `ConfigParser` instance into the runtime settings.
:arg configfile: A file that can be read by a `ConfigParser` instance
'''
global language,filterbots,time_stamps,time_stamps_index,botfile,basedirectory,datadirectory,userlistdirectory,reportdirectory,wikipridedirectory,sqlhost,sqlwikidb,sqluserdb,sqlconfigfile,sqldroptables
import os
import ConfigParser
config = ConfigParser.ConfigParser()
config.read(configfile)
# general settings
language = config.get('General','language')
if len(language) != 2:
raise Exception('Language code should be two characters (%s)'%language)
filterbots = config.get('General','filterbots')
startYM = config.get('General','startYM')
endYM = config.get('General','endYM')
(time_stamps,time_stamps_index) = utils.create_time_stamps_month(fromym=startYM,toym=endYM)
# directories
basedirectory = os.path.expanduser(config.get('Directories','basedirectory'))
datadirectory = os.path.expanduser(config.get('Directories','datadirectory'))
userlistdirectory = os.path.expanduser(config.get('Directories','userlistdirectory'))
reportdirectory = os.path.expanduser(config.get('Directories','reportdirectory'))
wikipridedirectory = os.path.expanduser(config.get('Directories','wikipridedirectory'))
# if not os.path.isdir(basedirectory):
# import errno
# try:
# os.makedirs(p)
# except OSError as exc: # Python >2.5
# if exc.errno == errno.EEXIST:
# pass
# else: raise
# sql
sqlhost = config.get('MySQL','sqlhost')
sqlwikidb = config.get('MySQL','sqlwikidb')
sqluserdb = config.get('MySQL','sqluserdb')
sqlconfigfile = config.get('MySQL','sqlconfigfile')
sqldroptables = config.getboolean('MySQL','sqldroptables')
sqlconfigfile = os.path.expanduser(sqlconfigfile)
if not os.path.isfile(sqlconfigfile):
logger.warning("%s sql config file not found"%sqlconfigfile)
def active_editors_target(min_ts, max_ts):
return
# input files
metricfile = os.path.join(old_to_new.output_folder, "rc_active_editors_count.csv")
# metricfile = '../data/datafiles/rc/rc_active_editors_count.csv'
targetfile = "../misc/targets/editor_target.csv"
# time range
tss, tss_i = utils.create_time_stamps_month(fromym=min_ts, toym=max_ts)
target = extractData(targetfile, "\t", "MMM-YY", [("Projected", 1), ("Target", 2)])
actual = extractData(metricfile, ",", "YYYY/MM/DD", [("Total", 1)])
def __init__(self,activation=5):
self.activation = activation
self.NS = ( '4', '5' )
ts,ts_i = utils.create_time_stamps_month(fromym='200401',toym='201012')
self.time_stamps = ts
self.time_stamps_index = ts_i
'''Only take time_stamps starting with self.year
'''
self.cohorts = range(2004,2011)
'''Cohort definition
'''
self.cohort_labels = ['%s cohort'% i for i in self.cohorts]
'''Cohort labels
'''
# self.sqlQuery = 'SELECT * FROM fabian WHERE namespace IN (4,5) AND add_edits > %s AND first_edit_year in (%s)'%(self.activation,','.join([str(c) for c in self.cohorts]))
Cohort.__init__(self)
def setTimeStamps(startYM, endYM):
global time_stamps, time_stamps_index
(time_stamps,
time_stamps_index) = utils.create_time_stamps_month(fromym=startYM,
toym=endYM)
def setTimeStamps(startYM,endYM): global time_stamps,time_stamps_index (time_stamps,time_stamps_index) = utils.create_time_stamps_month(fromym=startYM,toym=endYM)
