Newer
Older
# coding: utf-8
import logging
import sys
from datetime import datetime
import django.dispatch
import ijson
import django
from django.apps import apps
from django.db import transaction
from django.utils import timezone
from representatives.models import (Country, Mandate, Email, Address, WebSite,
Representative, Constituency, Phone, Group,
Chamber)
logger = logging.getLogger(__name__)
representative_pre_import = django.dispatch.Signal(
providing_args=['representative_data'])
def _get_rep_district_name(json):
num = json.get('num_circo')
nom = json.get('nom_circo')
if num == 'nd':
return nom
else:
ordinal = u'ère' if num == 1 else u'ème'
return '%s (%d%s circonscription)' % (nom, num, ordinal)
def _get_rep_parl_groups(json):
return [{
'name': g['responsabilite']['organisme'],
'role': g['responsabilite']['fonction'],
'start': json['mandat_debut']
} for g in json['groupes_parlementaires']]
def _get_rep_comittees(json):
return [{
'name': g['responsabilite']['organisme'],
'role': g['responsabilite']['fonction'],
'start': json['mandat_debut']
} for g in json['responsabilites']
if g['responsabilite']['organisme'].startswith('Commission')]
#
# Variant configuration
# - mail_domain is used to distinguish official vs personal emails
# - off_* fields are used for the official address of meps
# - mandates defines how mandates are created from the rep json
#
# Mandates are defined as follows
# - 'kind' indicates the group kind, a constant string
# - 'chamber' tells whether the group belongs to the chamber
# - 'from', if present, must be a function that takes the rep json and returns
# an array of dicts; one group will be created from each item in the dict.
# When 'from' is not present, only one group wil be created using the rep
# json (IOW, 'from' defaults to lambda repjson: [repjson])
# - 'name', 'abbr', 'role', 'start', 'end' are strings that are interpolated
# against the rep json or items returned by 'from'.
# - 'name_path', 'abbr_path', etc. can replace 'name', 'abbr'... by specifying
# a slash-separated dictionnary path where the value is to be found in the
# rep json or item returned by 'from'
# - 'name_fn', 'abbr_fn', etc. can also replace 'name', 'abbr'... by a
# function that takes the input item (rep json or item returned by 'from')
# and returns the value
#
FranceDataVariants = {
"AN": {
"chamber": u"Assemblée nationale",
"off_city": "Paris",
"off_street": u"Rue de l'Université",
"off_number": "126",
"off_code": "75355",
"off_name": u"Assemblée nationale",
"abbr": "%(groupe_sigle)s",
"name_path": "groupe/organisme",
"role_path": "groupe/fonction",
"start": "%(mandat_debut)s"
},
{
"kind": "department",
"abbr": "%(num_deptmt)s",
"name": "%(nom_circo)s",
"start": "%(mandat_debut)s"
},
{
"kind": "district",
"abbr": "%(num_deptmt)s-%(num_circo)s",
"name_fn": _get_rep_district_name,
"start": "%(mandat_debut)s"
},
{
"kind": "parl-group",
"chamber": True,
"from": _get_rep_parl_groups,
"abbr": "%(name)s",
"name": "%(name)s",
"role": "%(role)s",
"start": "%(start)s"
},
{
"kind": "comittee",
"from": _get_rep_comittees,
"abbr": "%(name)s",
"name": "%(name)s",
"role": "%(role)s",
"start": "%(start)s"
}
]
},
"SEN": {
"chamber": u"Sénat",
"mail_domain": "@senat.fr",
"off_city": "Paris",
"off_street": u"Rue de Vaugirard",
"off_number": "15",
"off_code": "75291",
"off_name": u"Palais du Luxembourg",
"mandates": [
{
"kind": "group",
"chamber": True,
"abbr": "%(groupe_sigle)s",
"name_path": "groupe/organisme",
"role_path": "groupe/fonction",
"start": "%(mandat_debut)s"
},
{
"kind": "department",
"abbr": "%(num_deptmt)s",
"name": "%(nom_circo)s",
"start": "%(mandat_debut)s"
},
{
"kind": "district",
"abbr": "%(num_deptmt)s-%(num_circo)s",
"name_fn": _get_rep_district_name,
"start": "%(mandat_debut)s"
},
{
"kind": "parl-group",
"from": _get_rep_parl_groups,
"abbr": "%(name)s",
"name": "%(name)s",
"role": "%(role)s",
"start": "%(start)s"
},
{
"kind": "comittee",
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
"from": _get_rep_comittees,
"abbr": "%(name)s",
"name": "%(name)s",
"role": "%(role)s",
"start": "%(start)s"
}
]
}
}
def _get_mdef_item(mdef, item, json, default=None):
if item in mdef:
return mdef[item] % json
if '%s_path' % item in mdef:
return _get_path(json, mdef['%s_path' % item])
if '%s_fn' % item in mdef:
return mdef['%s_fn' % item](json)
return default
def _parse_date(date):
return datetime.strptime(date, "%Y-%m-%d").date()
def _create_mandate(representative, group, constituency, role='',
begin_date=None, end_date=None):
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
mandate, _ = Mandate.objects.get_or_create(
representative=representative,
group=group,
constituency=constituency,
role=role,
begin_date=begin_date,
end_date=end_date
)
if _:
logger.debug('Created mandate %s', mandate.pk)
def _get_path(dict_, path):
'''
Get value at specific path in dictionary. Path is specified by slash-
separated string, eg _get_path(foo, 'bar/baz') returns foo['bar']['baz']
'''
cur = dict_
for part in path.split('/'):
cur = cur[part]
return cur
class GenericImporter(object):
def pre_import(self):
self.import_start_datetime = timezone.now()
def touch_model(self, model, **data):
'''
This method create or look up a model with the given data
it saves the given model if it exists, updating its
updated field
'''
instance, created = model.objects.get_or_create(**data)
if not created:
if instance.updated < self.import_start_datetime:
instance.save() # Updates updated field
return (instance, created)
class FranceDataImporter(GenericImporter):
url = 'http://francedata.future/data/parlementaires.json'
def parse_date(self, date):
return _parse_date(date)
def __init__(self, variant):
self.france = Country.objects.get(name="France")
self.chamber, _ = Chamber.objects.get_or_create(
name=self.variant['chamber'], country=self.france)
self.ch_constituency, _ = Constituency.objects.get_or_create(
name=self.variant['chamber'], country=self.france)
self.ch_group, _ = Group.objects.get_or_create(
name=self.variant['chamber'], kind='chamber', abbreviation=variant,
chamber=self.chamber)
@transaction.atomic
def manage_rep(self, rep_json):
'''
Import a rep as a representative from the json dict fetched from
FranceData (which comes from nosdeputes.fr)
'''
remote_id = rep_json[self.variant['remote_id_field']]
if rep_json['num_circo'] == 'non disponible':
rep_json['num_circo'] = 'nd'
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
if not remote_id:
logger.warning('Skipping MEP without UID %s %s',
rep_json['nom'],
rep_json[self.variant['remote_id_field']])
return
# Some versions of memopol will connect to this and skip inactive reps.
responses = representative_pre_import.send(sender=self,
representative_data=rep_json)
for receiver, response in responses:
if response is False:
logger.debug(
'Skipping MEP %s', rep_json['nom'])
return
try:
representative = Representative.objects.get(remote_id=remote_id)
except Representative.DoesNotExist:
representative = Representative(remote_id=remote_id)
# Save representative attributes
self.import_representative_details(representative, rep_json)
representative.save()
self.add_mandates(representative, rep_json)
self.add_contacts(representative, rep_json)
logger.debug('Imported MEP %s', unicode(representative))
return representative
def import_representative_details(self, representative, rep_json):
representative.active = True
if rep_json.get("date_naissance"):
representative.birth_date = _parse_date(rep_json["date_naissance"])
if rep_json.get("lieu_naissance"):
representative.birth_place = rep_json["lieu_naissance"]
representative.photo = rep_json['photo_url']
representative.full_name = rep_json["nom"]
gender_convertion_dict = {
u"F": 1,
u"H": 2
}
if 'sexe' in rep_json:
representative.gender = gender_convertion_dict.get(rep_json[
'sexe'], 0)
else:
representative.gender = 0
representative.slug = rep_json['slug']
def add_mandates(self, representative, rep_json):
'''
Create mandates from rep data based on variant configuration
'''
# Mandate in country group for party constituency
if rep_json.get('parti_ratt_financier'):
constituency, _ = Constituency.objects.get_or_create(
name=rep_json.get('parti_ratt_financier'), country=self.france)
group, _ = self.touch_model(model=Group,
abbreviation=self.france.code,
kind='country',
name=self.france.name)
_create_mandate(representative, group, constituency, 'membre')
# Configurable mandates
if mdef.get('chamber', False):
chamber = self.chamber
else:
chamber = None
if 'from' in mdef:
elems = mdef['from'](rep_json)
else:
elems = [rep_json]
for elem in elems:
name = _get_mdef_item(mdef, 'name', elem, '')
abbr = _get_mdef_item(mdef, 'abbr', elem, '')
group, _ = self.touch_model(model=Group,
abbreviation=abbr,
kind=mdef['kind'],
name=name)
role = _get_mdef_item(mdef, 'role', elem, 'membre')
start = _get_mdef_item(mdef, 'start', elem, None)
if start is not None:
start = _parse_date(start)
end = _get_mdef_item(mdef, 'end', elem, None)
if end is not None:
end = _parse_date(end)
_create_mandate(representative, group, self.ch_constituency,
role, start, end)
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
logger.debug(
'%s => %s: %s of "%s" (%s) %s-%s' % (rep_json['slug'],
mdef['kind'], role, name, abbr, start, end))
def add_contacts(self, representative, rep_json):
# Websites
websites = rep_json.get('sites_web', [])
for site in websites:
if not site['site'].startswith('http://twitter.com/'):
self.touch_model(model=WebSite,
url=site['site'],
representative=representative
)
# Twitter
if rep_json.get('twitter'):
tid = rep_json.get('twitter')
self.touch_model(model=WebSite,
representative=representative,
kind='twitter',
url='http://twitter.com/%s' % tid
)
# E-mails
emails = rep_json.get('emails', [])
for email in emails:
mail = email['email']
self.touch_model(
model=Email,
representative=representative,
kind=('official' if mail.endswith(self.variant['mail_domain'])
else 'other'),
email=mail)
# Official address
off_name = self.variant['off_name']
official_addr, _ = self.touch_model(model=Address,
representative=representative,
country=self.france,
city=self.variant['off_city'],
street=self.variant['off_street'],
number=self.variant['off_number'],
postcode=self.variant['off_code'],
kind='official',
name=off_name
)
# Addresses & phone numbers
addresses = rep_json.get('adresses', [])
for item in addresses:
addr = None
if 'geo' in item:
props = item['geo'].get('properties', {})
name = ''
if item['adresse'].lower().startswith('permanence'):
name = 'Permanence'
addr, _ = self.touch_model(model=Address,
representative=representative,
country=self.france,
city=props.get('city', ''),
street=props.get('street', ''),
number=props.get('housenumber', ''),
postcode=props.get('postcode', ''),
kind='',
name=name
)
elif item['adresse'].lower().startswith(off_name.lower()):
addr = official_addr
if 'tel' in item:
self.touch_model(model=Phone, address=addr,
)
def main(stream=None):
if not apps.ready:
django.setup()
an_importer = FranceDataImporter('AN')
GenericImporter.pre_import(an_importer)
sen_importer = FranceDataImporter('SEN')
GenericImporter.pre_import(sen_importer)
for data in ijson.items(stream or sys.stdin, ''):
for rep in data:
if rep['chambre'] == 'AN':
an_importer.manage_rep(rep)
elif rep['chambre'] == 'SEN':
sen_importer.manage_rep(rep)