def query_for_visits(operation, location, time, mongo, city):
"""Return a dict resulting of the call of `operation` with `location` and
`time` arguments."""
if 'find' in str(operation.im_func):
return answer_to_dict(operation(location, {time: 1}))
# $near is not supported in aggregate $match
if 'loc' in location:
ids = mongo.foursquare.venue.find({'city': city,
'loc': location['loc']}, {'_id': 1})
ids = [v['_id'] for v in ids]
location['lid'] = {'$in': ids}
del location['loc']
match = {'$match': location}
project = {'$project': {'time': '$'+time, 'lid': 1, '_id': 0}}
group = {'$group': {'_id': '$lid', 'visits': {'$push': '$time'}}}
query = [match, project, group]
convert = (lambda x: map(int, x)) if time == 'tuid' else None
return answer_to_dict(itertools.chain(operation(query)['result']), convert)
def collect_similars(venues_db, client, city):
"""Find similars venues for 100 location in city, save the result in DB and
return matching venues that were already in DB."""
venues = answer_to_dict(venues_db.find({'city': city}, {'loc': 1}))
chosen = sample(venues.items(), 500)
distances = []
all_match = []
for vid, loc in chosen:
similars = af.similar_venues(vid, client=client)
if similars is None:
continue
else:
print(vid, similars)
venues_db.update({'_id': vid}, {'$set': {'similars': similars}})
matching = answer_to_dict(venues_db.find({'_id': {'$in': similars}},
{'loc': 1}))
all_match.append(matching)
distances.append([geodesic_distance(loc, sloc)
for sloc in matching.itervalues()])
return chosen, distances, all_match
