import_representatives.py 13.3 KB
Newer Older
Nicolas Joyard's avatar
Nicolas Joyard committed
1
2
3
# coding: utf-8

import logging
4
import re
Nicolas Joyard's avatar
Nicolas Joyard committed
5
6
7
8
9
10
11
12
13
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
14
from django.utils.text import slugify
Nicolas Joyard's avatar
Nicolas Joyard committed
15

16
from representatives.models import (Country, Mandate, Email, Address, WebSite,
17
18
                                    Representative, Constituency, Phone, Group,
                                    Chamber)
19
20
from variants import FranceDataVariants

Nicolas Joyard's avatar
Nicolas Joyard committed
21
22
23
24
25
26
27
28
29

logger = logging.getLogger(__name__)

representative_pre_import = django.dispatch.Signal(
    providing_args=['representative_data'])


def _get_mdef_item(mdef, item, json, default=None):
    if item in mdef:
30
31
32
33
        try:
            return mdef[item] % json
        except:
            return default
Nicolas Joyard's avatar
Nicolas Joyard committed
34
35
36
37
38
39
40
41
42
43
44
45
46
47

    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()


48
49
def _create_mandate(representative, group, constituency, role='',
                    begin_date=None, end_date=None):
Nicolas Joyard's avatar
Nicolas Joyard committed
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
    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
        '''
85

Nicolas Joyard's avatar
Nicolas Joyard committed
86
87
88
89
90
91
92
93
94
        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)


Nicolas Joyard's avatar
Nicolas Joyard committed
95
96
97
98
99
100
101
102
103
104
105
106
def ensure_chambers():
    """
    Ensures chambers are created
    """
    france = Country.objects.get(name="France")
    for key in ('AN', 'SEN'):
        variant = FranceDataVariants[key]
        Chamber.objects.get_or_create(name=variant['chamber'],
                                      abbreviation=variant['abbreviation'],
                                      country=france)


Nicolas Joyard's avatar
Nicolas Joyard committed
107
class FranceDataImporter(GenericImporter):
108
    url = 'http://francedata.future/data/parlementaires.json'
Nicolas Joyard's avatar
Nicolas Joyard committed
109
110
111
112
113

    def parse_date(self, date):
        return _parse_date(date)

    def __init__(self, variant):
114
        self.france = Country.objects.get(name="France")
Nicolas Joyard's avatar
Nicolas Joyard committed
115
        self.variant = FranceDataVariants[variant]
Nicolas Joyard's avatar
Nicolas Joyard committed
116
        self.chamber = Chamber.objects.get(name=self.variant['chamber'])
117
118
        self.ch_constituency, _ = Constituency.objects.get_or_create(
            name=self.variant['chamber'], country=self.france)
Nicolas Joyard's avatar
Nicolas Joyard committed
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136

    @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)
        '''

        # 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

137
        changed = False
Nicolas Joyard's avatar
Nicolas Joyard committed
138
        slug = slugify('%s-%s' % (
139
            rep_json['nom'] if 'nom' in rep_json
Nicolas Joyard's avatar
Nicolas Joyard committed
140
141
142
            else rep_json['prenom'] + " " + rep_json['nom_de_famille'],
            _parse_date(rep_json["date_naissance"])
        ))
143

Nicolas Joyard's avatar
Nicolas Joyard committed
144
        try:
145
            representative = Representative.objects.get(slug=slug)
Nicolas Joyard's avatar
Nicolas Joyard committed
146
        except Representative.DoesNotExist:
147
148
            representative = Representative(slug=slug)
            changed = True
Nicolas Joyard's avatar
Nicolas Joyard committed
149

150
151
        if rep_json['num_circo'] == 'non disponible':
            rep_json['num_circo'] = 'nd'
Nicolas Joyard's avatar
Nicolas Joyard committed
152

153
154
        # Save representative attributes
        self.import_representative_details(representative, rep_json, changed)
Nicolas Joyard's avatar
Nicolas Joyard committed
155
156
157
158
159
160
161
162
163

        self.add_mandates(representative, rep_json)

        self.add_contacts(representative, rep_json)

        logger.debug('Imported MEP %s', unicode(representative))

        return representative

164
165
    def import_representative_details(self, representative, rep_json, changed):
        active = True
166
        if rep_json.get("ancien_depute", 0) == 1:
167
            active = False
168
        if rep_json.get("ancien_senateur", 0) == 1:
169
170
171
172
            active = False
        if representative.active != active:
            representative.active = active
            changed = True
Nicolas Joyard's avatar
Nicolas Joyard committed
173
174

        if rep_json.get("date_naissance"):
175
176
177
178
            birth_date = _parse_date(rep_json["date_naissance"])
            if representative.birth_date != birth_date:
                representative.birth_date = birth_date
                changed = True
Nicolas Joyard's avatar
Nicolas Joyard committed
179

180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
        if rep_json.get("lieu_naissance"):
            birth_place = rep_json["lieu_naissance"]
            if representative.birth_place != birth_place:
                representative.birth_place = birth_place
                changed = True

        photo = rep_json['photo_url']
        if representative.photo != photo:
            representative.photo = photo
            changed = True

        first_name = rep_json['prenom']
        if representative.first_name != first_name:
            representative.first_name = first_name
            changed = True

        last_name = rep_json['nom_de_famille']
        if representative.last_name != last_name:
            representative.last_name = last_name
            changed = True

        full_name = rep_json["nom"]
        if representative.full_name != full_name:
            representative.full_name = full_name
            changed = True

        gender_convertion_dict = {u"F": 1, u"H": 2}
Nicolas Joyard's avatar
Nicolas Joyard committed
207
        if 'sexe' in rep_json:
208
            gender = gender_convertion_dict.get(rep_json['sexe'], 0)
Nicolas Joyard's avatar
Nicolas Joyard committed
209
        else:
210
211
212
213
            gender = 0
        if representative.gender != gender:
            representative.gender = gender
            changed = True
Nicolas Joyard's avatar
Nicolas Joyard committed
214

215
216
        if changed:
            representative.save()
Nicolas Joyard's avatar
Nicolas Joyard committed
217
218
219
220
221
222

    def add_mandates(self, representative, rep_json):
        '''
        Create mandates from rep data based on variant configuration
        '''

223
224
225
226
227
228
229
230
231
232
233
234
235
        # 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
Nicolas Joyard's avatar
Nicolas Joyard committed
236
        for mdef in self.variant['mandates']:
237
238
239
240
241
            if mdef.get('chamber', False):
                chamber = self.chamber
            else:
                chamber = None

Nicolas Joyard's avatar
Nicolas Joyard committed
242
243
244
245
246
247
248
249
250
251
252
253
            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'],
254
                                            chamber=chamber,
Nicolas Joyard's avatar
Nicolas Joyard committed
255
256
257
258
259
260
261
262
263
264
                                            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)

265
266
                _create_mandate(representative, group, self.ch_constituency,
                                role, start, end)
Nicolas Joyard's avatar
Nicolas Joyard committed
267
268
269
270
271
272

                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):
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
        # Chamber page
        changed = False
        try:
            site = WebSite.objects.get(kind=self.variant['abbreviation'],
                                       representative=representative)
        except WebSite.DoesNotExist:
            site = WebSite(kind=self.variant['abbreviation'],
                           representative=representative)
            changed = True

        if site.url != rep_json[self.variant['chamber_url_field']]:
            site.url = rep_json[self.variant['chamber_url_field']]
            changed = True

        if changed:
            site.save()

Nicolas Joyard's avatar
Nicolas Joyard committed
290
291
292
        # Websites
        websites = rep_json.get('sites_web', [])
        for site in websites:
293
294
295
296
297
298
299
            if re.search(r'facebook\.com', site['site']):
                self.touch_model(model=WebSite,
                                 url=site['site'],
                                 kind='facebook',
                                 representative=representative
                                 )
            elif not re.search(r'twitter\.com', site['site']):
Nicolas Joyard's avatar
Nicolas Joyard committed
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
                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)

325
326
327
328
329
330
331
332
333
334
335
336
337
        # 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
                                            )

Nicolas Joyard's avatar
Nicolas Joyard committed
338
339
        # Addresses & phone numbers
        addresses = rep_json.get('adresses', [])
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
        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,
Nicolas Joyard's avatar
Nicolas Joyard committed
364
                                 representative=representative,
365
                                 kind='', number=item['tel']
Nicolas Joyard's avatar
Nicolas Joyard committed
366
367
368
369
370
371
372
                                 )


def main(stream=None):
    if not apps.ready:
        django.setup()

Nicolas Joyard's avatar
Nicolas Joyard committed
373
374
    ensure_chambers()

375
376
377
378
379
    an_importer = FranceDataImporter('AN')
    GenericImporter.pre_import(an_importer)

    sen_importer = FranceDataImporter('SEN')
    GenericImporter.pre_import(sen_importer)
Nicolas Joyard's avatar
Nicolas Joyard committed
380
381
382

    for data in ijson.items(stream or sys.stdin, ''):
        for rep in data:
383
384
385
386
387
            try:
                if rep['chambre'] == 'AN':
                    an_importer.manage_rep(rep)
                elif rep['chambre'] == 'SEN':
                    sen_importer.manage_rep(rep)
388
389
            except Exception:
                logger.exception('error trying to import rep %s', str(rep))