facebook-elasticsearch/fb_weekly.py
François Pelletier 466c3b0de1 first commit
2017-05-10 23:43:29 -04:00

131 lines
4.4 KiB
Python
Executable file

import requests
import facebook
import os
from elasticsearch import Elasticsearch
from datetime import datetime
from datetime import timedelta
app_id=os.environ['FB_APP_ID']
app_secret=os.environ['FB_APP_SECRET']
index_name=''
page_name=''
def get_fb_token(app_id, app_secret):
payload = {'grant_type': 'client_credentials', 'client_id': app_id, 'client_secret': app_secret}
file = requests.post('https://graph.facebook.com/oauth/access_token?', params = payload)
#print file.text #to test what the FB api responded with
result = file.text.split("=")[1]
#print file.text #to test the TOKEN
return result
access_token=get_fb_token(app_id,app_secret)
week_delta = timedelta(days=7)*4
es = Elasticsearch()
es.indices.create(index=index_name, ignore=400)
graph = facebook.GraphAPI(access_token=access_token, version='2.7')
page_obj = graph.get_object(page_name)
def index_post(post,index_name):
es.index(
index=index_name,
doc_type="data",
id=post[u'id'],
timestamp=post[u'created_time'],
body=post)
def index_comments(post,index_name):
es.index(
index=index_name,
doc_type="comment",
id=post[u'id'],
timestamp=post[u'created_time'],
body=post)
def index_like(post,index_name):
es.index(
index=index_name,
doc_type="like",
id=post[u'like_id'],
timestamp=post[u'like_time'],
body=post)
def getfbpostsfrompage(fb_graph,page_id,field_list,time_since,time_until):
all_posts = []
res = fb_graph.get_object('/'+
page_id+
'/posts?fields='+','.join(field_list)+
'&since='+time_since+
'&until='+time_until)
while(True):
try:
for page in res[u'data']:
all_posts.append(page)
res=requests.get(res['paging']['next']).json()
except KeyError:
break
return all_posts
def getpostmetacomplet(fb_graph,post_id,field_list):
post_meta_complet = fb_graph.get_object('/'+
post_id+
'?fields='+','.join(field_list))
return post_meta_complet
def getpostreact(fb_graph,post_id,field_list,react_type,filter_type):
res = fb_graph.get_object('/'+post_id+
'/'+react_type+'/?fields='+','.join(field_list)+
'&filter='+filter_type)
all_comments = []
while(True):
try:
for comment in res[u'data']:
all_comments.append(comment)
res=requests.get(res[u'paging'][u'next']).json()
except KeyError:
break
return all_comments
def dict_update(l,x):
l.update(x)
return l
res = getfbpostsfrompage(graph,
page_obj[u'id'],
['id','created_time'],
(datetime.now().date()-week_delta).isoformat(),
datetime.now().date().isoformat())
for pp in res:
# Post
post_complet = getpostmetacomplet(graph,
pp[u'id'],
['message','created_time','id',
'status_type','shares','link',
'via'])
# Like
all_post_likes = getpostreact(graph,pp[u'id'],
['id','name'],
'likes',
'stream')
like_count = len(all_post_likes)
post_complet.update({u'like_count':like_count})
# Sauvegarde des "post"
index_post(post_complet,index_name)
# Sauvegarde des "like"
like_dicts = [dict_update(x,{u'like_time':pp['created_time'],
u'like_id':pp[u'id']+'_'+x['id']}) for x in all_post_likes]
for l in like_dicts:
index_like(l,index_name)
# Comments
res_comments = getpostreact(graph,pp[u'id'],['id'],'comments','stream')
for cc in res_comments:
comment_complet = getpostmetacomplet(graph,
cc[u'id'],
['id','from','message',
'created_time','comment_count','like_count',
'parent'])
# Sauvegarde des "comments"
index_comments(comment_complet,index_name)
print pp