@@ -8,6 +8,7 @@
88# @stable no (HTML can change)
99# @parse url, title, content, suggestion
1010
11+import re
1112from urllib import urlencode
1213from urlparse import urlparse, parse_qsl
1314from lxml import html
@@ -78,15 +79,22 @@ country_to_hostname = {
7879 'TW': 'www.google.com.tw' # Taiwan
7980}
8081
82+# osm
83+url_map = 'https://www.openstreetmap.org/'\
84+ + '?lat={latitude}&lon={longitude}&zoom={zoom}&layers=M'
85+
8186# search-url
8287search_path = '/search'
83−maps_path = '/maps/'
84−redirect_path = '/url'
85−images_path = '/images'
8688search_url = ('https://{hostname}' +
8789 search_path +
8890 '?{query}&start={offset}&gbv=1')
8991
92+# other URLs
93+map_hostname_start = 'maps.google.'
94+maps_path = '/maps'
95+redirect_path = '/url'
96+images_path = '/images'
97+
9098# specific xpath variables
9199results_xpath = '//li[@class="g"]'
92100url_xpath = './/h3/a/@href'
@@ -95,10 +103,32 @@ content_xpath = './/span[@class="st"]'
95103content_misc_xpath = './/div[@class="f slp"]'
96104suggestion_xpath = '//p[@class="_Bmc"]'
97105
106+# map : detail location
107+map_address_xpath = './/div[@class="s"]//table//td[2]/span/text()'
108+map_phone_xpath = './/div[@class="s"]//table//td[2]/span/span'
109+map_website_url_xpath = 'h3[2]/a/@href'
110+map_website_title_xpath = 'h3[2]'
111+
112+# map : near the location
113+map_near = 'table[@class="ts"]//tr'
114+map_near_title = './/h4'
115+map_near_url = './/h4/a/@href'
116+map_near_phone = './/span[@class="nobr"]'
117+
118+# images
98119images_xpath = './/div/a'
99120image_url_xpath = './@href'
100121image_img_src_xpath = './img/@src'
101122
123+# property names
124+# FIXME : no translation
125+property_address = "Address"
126+property_phone = "Phone number"
127+property_location = "Location"
128+property_website = "Web site"
129+property_gplus_website = "Google plus"
130+
131+# cookies
102132pref_cookie = ''
103133nid_cookie = {}
104134
@@ -122,6 +152,11 @@ def get_google_nid_cookie(google_hostname):
122152
123153# remove google-specific tracking-url
124154def parse_url(url_string, google_hostname):
155+ # sanity check
156+ if url_string is None:
157+ return url_string
158+
159+ # normal case
125160 parsed_url = urlparse(url_string)
126161 if (parsed_url.netloc in [google_hostname, '']
127162 and parsed_url.path == redirect_path):
@@ -131,6 +166,19 @@ def parse_url(url_string, google_hostname):
131166 return url_string
132167
133168
169+# URL : get label
170+def url_get_label(url_string):
171+ # sanity check
172+ if url_string is None:
173+ return url_string
174+
175+ # normal case
176+ parsed_url = urlparse(url_string)
177+ if parsed_url.netloc == 'plus.google.com':
178+ return property_gplus_website
179+ return property_website
180+
181+
134182# returns extract_text on the first result selected by the xpath or None
135183def extract_text_from_dom(result, xpath):
136184 r = result.xpath(xpath)
@@ -151,7 +199,7 @@ def request(query, params):
151199 if len(language_array) == 2:
152200 country = language_array[1]
153201 else:
154− country = ' '
202+ country = 'US'
155203 language = language_array[0] + ',' + language_array[0] + '-' + country
156204
157205 if use_locale_domain:
@@ -196,21 +244,32 @@ def response(resp):
196244 try:
197245 url = parse_url(extract_url(result.xpath(url_xpath), google_url), google_hostname)
198246 parsed_url = urlparse(url, google_hostname)
199− if (parsed_url.netloc == google_hostname
200− and (parsed_url.path == search_path
201− or parsed_url.path.startswith(maps_path))):
202− # remove the link to google news and google maps
203− # FIXME : sometimes the URL is https://maps.google.*/maps
204− # no consequence, the result trigger an exception after which is ignored
205− continue
247+
248+ # map result
249+ if ((parsed_url.netloc == google_hostname and parsed_url.path.startswith(maps_path))
250+ or (parsed_url.netloc.startswith(map_hostname_start))):
251+ x = result.xpath(map_near)
252+ if len(x) > 0:
253+ # map : near the location
254+ results = results + parse_map_near(parsed_url, x, google_hostname)
255+ else:
256+ # map : detail about a location
257+ results = results + parse_map_detail(parsed_url, result, google_hostname)
258+
259+ # google news
260+ elif (parsed_url.netloc == google_hostname
261+ and parsed_url.path == search_path):
262+ # skipping news results
263+ pass
206264
207265 # images result
208− if (parsed_url.netloc == google_hostname
209− and parsed_url.path == images_path):
266+ elif (parsed_url.netloc == google_hostname
267+ and parsed_url.path == images_path):
210268 # only thumbnail image provided,
211269 # so skipping image results
212270 # results = results + parse_images(result, google_hostname)
213271 pass
272+
214273 else:
215274 # normal result
216275 content = extract_text_from_dom(result, content_xpath)
@@ -223,7 +282,7 @@ def response(resp):
223282 results.append({'url': url,
224283 'title': title,
225284 'content': content})
226− except Exception:
285+ except:
227286 continue
228287
229288 # parse suggestion
@@ -249,3 +308,96 @@ def parse_images(result, google_hostname):
249308 'template': 'images.html'})
250309
251310 return results
311+
312+
313+def parse_map_near(parsed_url, x, google_hostname):
314+ results = []
315+
316+ for result in x:
317+ title = extract_text_from_dom(result, map_near_title)
318+ url = parse_url(extract_text_from_dom(result, map_near_url), google_hostname)
319+ phone = extract_text_from_dom(result, map_near_phone)
320+ if phone is not None:
321+ phone = property_phone + ": " + phone
322+ results.append({'url': url,
323+ 'title': title,
324+ 'content': phone})
325+
326+ return results
327+
328+
329+def parse_map_detail(parsed_url, result, google_hostname):
330+ results = []
331+
332+ # try to parse the geoloc
333+ m = re.search('@([0-9\.]+),([0-9\.]+),([0-9]+)', parsed_url.path)
334+ if m is None:
335+ m = re.search('ll\=([0-9\.]+),([0-9\.]+)\&z\=([0-9]+)', parsed_url.query)
336+
337+ if m is not None:
338+ # geoloc found
339+ lon = float(m.group(2))
340+ lat = float(m.group(1))
341+ zoom = int(m.group(3))
342+
343+ # TODO : map zoom to dlon / dlat
344+ dlon = 0.000001
345+ dlat = 0.000001
346+
347+ boundingbox = [round(lat - dlat, 7), round(lat + dlat, 7), round(lon - dlon, 7), round(lon + dlon, 7)]
348+ map_url = url_map\
349+ .replace('{latitude}', str(lat))\
350+ .replace('{longitude}', str(lon))\
351+ .replace('{zoom}', str(zoom+2))
352+
353+ geojson = {u'type': u'Point',
354+ u'coordinates': [lon, lat]
355+ }
356+
357+ # attributes
358+ attributes = []
359+ add_attributes(attributes, property_address, extract_text_from_dom(result, map_address_xpath))
360+ add_attributes(attributes, property_phone, extract_text_from_dom(result, map_phone_xpath))
361+
362+ # title / content / url
363+ website_title = extract_text_from_dom(result, map_website_title_xpath)
364+ content = extract_text_from_dom(result, content_xpath)
365+ website_url = parse_url(extract_text_from_dom(result, map_website_url_xpath), google_hostname)
366+
367+ # add an infobox if there is a website
368+ if website_url is not None:
369+ results.append({'infobox': website_title,
370+ 'id': website_url,
371+ 'content': content,
372+ 'attributes': attributes,
373+ 'urls': [
374+ {'title': url_get_label(website_url), 'url': website_url},
375+ {'title': property_location, 'url': map_url}
376+ ]
377+ })
378+
379+ # usefull because user can see the map directly into searx
380+ results.append({'template': 'map.html',
381+ 'title': website_title,
382+ 'content': (content + '<br />' if content is not None else '')
383+ + attributes_to_html(attributes),
384+ 'longitude': lon,
385+ 'latitude': lat,
386+ 'boundingbox': boundingbox,
387+ 'geojson': geojson,
388+ 'url': website_url if website_url is not None else map_url
389+ })
390+ return results
391+
392+
393+def add_attributes(attributes, name, value):
394+ if value is not None and len(value) > 0:
395+ attributes.append({'label': name, 'value': value})
396+
397+
398+def attributes_to_html(attributes):
399+ retval = '<table class="table table-striped">'
400+ for a in attributes:
401+ retval = retval + '<tr><th>' + a.get('label') + '</th><td>' + a.get('value') + '</td></tr>'
402+ retval = retval + '</table>'
403+ return retval