Skip to content

Commit

Permalink
Merge 5b0c9ec into be686d4
Browse files Browse the repository at this point in the history
  • Loading branch information
aguilerapy committed Mar 10, 2021
2 parents be686d4 + 5b0c9ec commit d7df252
Show file tree
Hide file tree
Showing 4 changed files with 61 additions and 6 deletions.
Original file line number Diff line number Diff line change
Expand Up @@ -4,14 +4,13 @@
from kingfisher_scrapy.util import parameters


class MexicoAdministracionPublicaFederal(IndexSpider):
class MexicoAPFAPI(IndexSpider):
"""
Domain
Administración Pública Federal (APF)
Bulk download documentation
https://datos.gob.mx/busca/dataset/concentrado-de-contrataciones-abiertas-de-la-apf
Administración Pública Federal (APF) - Secretaria de la Función Pública (SFP) - Secretaría de Hacienda y Crédito
Público (SHCP)
"""
name = 'mexico_administracion_publica_federal'
name = 'mexico_apf_api'

# BaseSpider
root_path = 'results.item'
Expand Down
23 changes: 23 additions & 0 deletions kingfisher_scrapy/spiders/mexico_apf_bulk.py
Original file line number Diff line number Diff line change
@@ -0,0 +1,23 @@
from kingfisher_scrapy.base_spider import CompressedFileSpider
from kingfisher_scrapy.util import components


class MexicoAPFBulk(CompressedFileSpider):
"""
Domain
Administración Pública Federal (APF) - Secretaria de la Función Pública (SFP) - Secretaría de Hacienda y Crédito
Público (SHCP)
Bulk download documentation
https://datos.gob.mx/busca/dataset/concentrado-de-contrataciones-abiertas-de-la-apf
"""
name = 'mexico_apf_bulk'

# BaseSpider
root_path = 'item'

# CompressedFileSpider
data_type = 'release'

def start_requests(self):
url = 'https://compranetinfo.hacienda.gob.mx/dabiertos/contrataciones_arr.json.zip'
yield self.build_request(url, formatter=components(-1))
34 changes: 34 additions & 0 deletions kingfisher_scrapy/spiders/mexico_apf_shcp_bulk.py
Original file line number Diff line number Diff line change
@@ -0,0 +1,34 @@
import scrapy

from kingfisher_scrapy.base_spider import CompressedFileSpider
from kingfisher_scrapy.util import components, handle_http_error


class MexicoAPFSHCPBulk(CompressedFileSpider):
"""
Domain
Administración Pública Federal (APF) - Secretaría de Hacienda y Crédito Público (SHCP)
Bulk download documentation
https://datos.gob.mx/busca/dataset/concentrado-de-contrataciones-abiertas-de-la-apf-shcp
"""
name = 'mexico_apf_shcp_bulk'

# BaseSpider
root_path = 'item'

# CompressedFileSpider
data_type = 'release'

def start_requests(self):
yield scrapy.Request(
'https://drive.google.com/uc?id=1dHsnijrC_IQyGn0eY4ZzFwDLiJ-8kpS3',
meta={'file_name': 'confirmation.html'},
callback=self.parse_list
)

@handle_http_error
def parse_list(self, response):
url = response.xpath('//a[@id="uc-download-link"]/@href').get()
yield self.build_request(url=f'https://drive.google.com{url}',
formatter=components(-1),
meta={'file_name': 'contrataciones.zip'})
Original file line number Diff line number Diff line change
Expand Up @@ -5,7 +5,6 @@


class MexicoPlataformaDigitalNacional(CompressedFileSpider):

"""
Domain
Plataforma Digital Nacional
Expand Down

0 comments on commit d7df252

Please sign in to comment.