def write_json_sf_top5():
obj_name = 'static_10_sf_city_larger'
sf_data = _load_obj(obj_name)
coordinates, address, retsly_ids, house_img_urls, descriptions = pipeline.query_retsly_v1()
# totall 11
# select top 5 locations
post_nums = []
instagram_top_5_house = dict()
for k,v in enumerate(sf_data):
print k,v # v is id for one listing
idx = retsly_ids.index(v)
print retsly_ids[idx], address[idx], house_img_urls[idx], descriptions[idx]
#urls, text = pipeline.post_retrieval(sf_data[v], query)
instgram_data = sf_data[v]
post_nums.append(len(instgram_data))
if len(instgram_data) >= 196:
instagram_top_5_house[v] = instgram_data
sf_test_top_5_house = []
for k, v in enumerate(instagram_top_5_house):
print k, v
idx = retsly_ids.index(v)
inst_data = instagram_top_5_house[v]
inst_img_urls = ""
inst_captions = ""
count = 0
for k2, v2 in enumerate(inst_data):
inst_one = inst_data[v2]
if inst_one['caption'] == None or inst_one['img'] == None:
continue
inst_captions += inst_one['caption']['text']
inst_captions += '_!@#*()_'
inst_img_urls += inst_one['img']
inst_img_urls += '_!@#*()_'
sf_test_top_5_house.append({'name' : retsly_ids[idx], 'latitude' : float(coordinates[idx][1]),
'longitude' : float(coordinates[idx][0]), 'url' : house_img_urls[idx], 'description' : descriptions[idx],
'address' : address[idx], 'inst_captions' : inst_captions, 'inst_img_urls' : inst_img_urls})
# store house info to json
pdb.set_trace()
with open('sf_test_top_5_house_v2.json', 'w') as fp:
json.dump(sf_test_top_5_house, fp)
pdb.set_trace()
def single_address_full_pipeline(pipeline, query):
# step 1: query retsly to get listing geolocation
coordinates, address, retsly_ids, house_image_urls, descriptions = pipeline.query_retsly_v1()
pdb.set_trace()
sf_data = _load_obj('static_10_sf_city_larger')
max_listing = 3
ids = 0
for k,v in enumerate(sf_data):
print k,v # v is id for one listing
idx = retsly_ids.index(v)
print retsly_ids[idx], address[idx]
pdb.set_trace()
# 0358d2ba3e8f74256e993f5b398cb2b1
urls, text = pipeline.post_retrieval(sf_data[v], query)
pdb.set_trace()
ids += 1
if ids >= max_listing:
break
return urls, text
def sf_batch_full_pipeline(pipeline, query):
# step 1: query retsly to get listing geolocation
coordinates, address, retsly_ids = pipeline.query_retsly_v1()
# get instgram media data
#geolocation_tuple = [-122.4120249,37.73691843]
#current_loc_trulia = [-122.3981040, 37.788900]
# step 2: query instagram to get social media data
try:
sf_data = _load_obj('static_10_sf_city')
except:
pipeline.query_instagram_media(coordinates, retsly_ids, 10)
# step 3: build wordvec similarities and topics from each listing
for k,v in enumerate(sf_data):
print k,v # v is id for one listing
idx = retsly_ids.index(v)
print retsly_ids[idx], address[idx]
urls, text = pipeline.post_retrieval(sf_data[v], query)
pdb.set_trace()
def more_sf_data(save_name): coordinates, address, retsly_ids = pipeline.query_retsly_v1() try: sf_data = _load_obj(save_name) except: pipeline.query_instagram_media(coordinates, retsly_ids, save_name, 10)