In [2]:

    

%matplotlib inline
from bigbang.archive import Archive
from bigbang import repo_loader;
from bigbang.git_repo import GitRepo,ALL_ATTRIBUTES
import matplotlib.pyplot as plt
import numpy as np
import pandas as pd


    

In [ ]:

    




    

In [5]:

    

repo = repo_loader.get_repo("bigbang", in_type="name")
full_info = repo.commit_data;
#repo.commit_data


    

    




Checking if cached
Running Entity Resolution on bigbang

In [6]:

    

print(full_info.columns)


    

    




Index([u'Unnamed: 0', u'Commit Message', u'Committer Email', u'Committer Name', u'HEXSHA', u'Parent Commit', u'Time', u'Touched File', u'Person-ID'], dtype='object')

In [7]:

    

import networkx as nx;


    

In [8]:

    

G = nx.Graph()

def add_to_graph(r): # Adds info from row to graph
    author = r["Committer Name"]
    files = r["Touched File"]
    for f in files:
        if G.has_edge(author, f):
            G[author][f]['weight'] += 1
        else:
            G.add_edge(author, f, weight = 1)
            G.node[author]["type"] = 'author'
            G.node[f]["type"] = 'file'
        
a = full_info.apply(add_to_graph, axis=1) # Dummy to make it not print
widths = [G[u][v]['weight']/20.0 for u, v in G.edges()]


    

In [9]:

    

#nx.draw(G)
plt.figure(1,figsize=(15, 15))
pos = nx.spring_layout(G, k = 0.6, iterations = 20)
nx.draw_networkx_nodes(G, pos, nodelist = [node[0] for node in G.nodes(data=True) if node[1]["type"] == 'author'], node_size=750, node_color = 'r')
nx.draw_networkx_nodes(G, pos, nodelist = [node[0] for node in G.nodes(data=True) if node[1]["type"] == 'file'], node_color = 'b')
nx.draw_networkx_edges(G, pos, width=widths, edge_color = 'g', alpha = 0.8)
#nx.draw_networkx_edges(G, pos, width=widths, edge_color = widths)
nx.draw_networkx_labels(G, pos);


    

In [10]:

    

from networkx.readwrite import json_graph
data = json_graph.node_link_data(G)


    

In [11]:

    

import json
with open ('viz/git/graph.json', 'w') as outfile:
    json.dump(data, outfile, indent=4)


    

In [ ]: