Obtiene un listado de objetos Regiones y Comunas con su relacion Region de Chile

get_regiones_comunas.py

# -*- coding: utf-8 -*-
"""
Obtiene un listado de objetos de Regiones y Comunas con su relacion Region de
Chile
 
@author: misalabs
"""
import requests
from bs4 import BeautifulSoup
 
 
class Region(object):
    def __init__(self, id, name):
        self.id = id
        self.name = name.encode("utf-8")
 
 
class Comuna(object):
    def __init__(self, id, name, region_id):
        self.id = id
        self.name = name.encode("utf-8")
        self.region_id = region_id
 
 
# Get regions' list Chile
regions_html = requests.get('http://es.wikipedia.org/wiki/Región_de_Chile')
 
comunas_html = requests.get("http://es.wikipedia.org/wiki/Anexo:Comunas_de_Chile")
 
# Construct Soup
page_regions = BeautifulSoup(regions_html.text)
table_regions = page_regions.findAll('table')[3]
 
page_comuna = BeautifulSoup(comunas_html.text)
table_comunas = page_comuna.findAll('table')[0]
 
# Get Regions
regions = []
for idx, row in enumerate(table_regions.findAll('tr')):
    cells_region = row.findAll('td')[3:4]
    for region in cells_region:
        regions.append(Region(id=idx, name=region.getText()))
 
# Get Comunas and set relations with Region
comunas = []
for idx, row in enumerate(table_comunas.findAll('tr')):
    cell_comuna = row.findAll('td')[1:2]
    cell_region = row.findAll('td')[4:5]
    if cell_comuna and cell_region:
      comuna_n = cell_comuna[:1][0].getText()
      region_n = cell_region[:1][0].getText().encode('utf-8')
      # Set relation with Region
      for region in regions:
        if region.name in region_n:
          comunas.append(Comuna(id=idx, name=comuna_n, region_id=region.id))
 
for region in regions:
    print region.name, region.id
 
for comuna in comunas:
    print comuna.name, comuna.region_id