output forcenode format for d3.json from lucidworks fusion

README.txt

./forcenode.py  35.233.253.193 celeb-list.txt > amps.json

celeb-list.txt

Ale_BarralesM
Taibo2
RicardoAnayaC
MikelArriolaP
PRI_Nacional
EPN
JoseAMeadeK
Mzavalagc
tatclouthier
AntonioAttolini
JuarezCisneros
Claudiashein
dparamooficial
yeidckol
lopezobrador_

forcenode.py

#!/usr/bin/python
import time
import sys  
import tweepy
import json
import requests

reload(sys)
sys.setdefaultencoding('utf8')


if __name__ == '__main__':
    try:
        if sys.argv[1] and sys.argv[2]:
            f = open(sys.argv[2])

            group = 0
            amps = []
            links = []

            # loop over short list of screen names in file looking for who else is mentioned with them in original tweets
            for screenName in f:
                group = group + 1

                screenName = screenName.rstrip("\r\n")

                url = "http://%s:8764/api/apollo/apps/sockitter/query-pipelines/sockitter/collections/sockitter/select" % sys.argv[1]
                payload = {
                    'debug': 'true',
                    'echoParams': 'all',
                    'f.userMentionScreenName_ss.facet.limit': 100,
                    'facet.field': 'userMentionScreenName_ss',
                    'wt': 'json',
                    'rows': 5,
                    'q': 'tweet_t:%s AND -RT' % screenName
                }
                auth = ('admin', 'f00bar222')
                r = requests.get(url, params=payload, auth=auth)
        
                # list user screennames mentioned from query
                i = 0
                citem = ""
                
                # print "processing co-mentions for %s" % screenName.rstrip("\r\n") 
                for item in r.json()['facet_counts']['facet_fields']['userMentionScreenName_ss']:
                    if (i % 2) == 0: 
                        # facet item 
                        citem = item.rstrip("\r\n")
                        amps.append({"id": citem, "group": group}) 
                    else:
                        # facet count
                        # print citem, item
                        links.append({"source": screenName, "target": citem, "value": item}) 
                        if item < 1:
                            break
                    i = i + 1
    
        else:
            # doesn't happen 
            sys.exit()


        output = {}
        output["nodes"] = amps
        output["links"] = links
        print json.dumps(output)

    except Exception as ex:
        if len(sys.argv) == 2: 
            print "enter an IP"
        elif len(sys.argv) == 1: 
            print "enter an IP and filename"
        else:
            print "something wrong: %s" % ex