IEMS5720 Blogosphere

gistfile1.py

from apiclient.discovery import build
import httplib2
import json

API_KEY = '...'

vertices = {}
edges = []

def update_user(user):
    name = user['displayName']
    uid = user['id']
    vertices[uid] = name

urls = []
with open('urls.txt', 'r') as f:
    for line in f:
        line = line.strip()
        if len(line) == 0:
            continue
        urls.append(line.replace('blogspot.hk', 'blogspot.com'))

service = build('blogger', 'v3', developerKey=API_KEY)

for url in urls:
    blog = service.blogs().getByUrl(url=url).execute()
    posts = service.posts().list(blogId=blog['id']).execute()
    for post in posts['items']:
        author = post['author']
        update_user(author)
        comments = service.comments().list(blogId=blog['id'], postId=post['id']).execute()
        if not comments.has_key('items'):
            continue
        for comment in comments['items']:
            commenter = comment['author']
            update_user(commenter)
            edges.append((commenter['id'], author['id']))

id_map = {}

def get_int_id(uid):
    if not uid in id_map:
        id_map[uid] = len(id_map)
    return id_map[uid]

nodes = []
weights = {}
links = []

for edge in edges:
    source = get_int_id(edge[0])
    target = get_int_id(edge[1])
    if source == target:
        continue
    if not (source, target) in weights:
        weights[(source, target)] = 1
    else:
        weights[(source, target)] += 1

for vertex in sorted(vertices.iteritems(), key=lambda vertex:get_int_id(vertex[0])):
    nodes.append({
        "index": len(nodes),
        "name": vertex[1],
        "weight": 0,
        "in": 0,
        "out": 0,
    })

for key in weights:
    value = weights[key]
    source, target = key
    nodes[source]["weight"] += value
    nodes[source]["out"] += value
    nodes[target]["weight"] += value
    nodes[target]["in"] += value
    links.append({
        'source': source,
        'target': target,
        'value': value
    })


print json.dumps({
    'nodes': nodes,
    'links': links
}, indent=2)