#!/usr/bin/env python
# coding: utf-8

# In[1]:


import json

# Load our preprocessed abstract data
with open('visvssrelationships_data_2016.json', 'r') as f:
    abstracts = json.load(f)


# In[ ]:


import difflib

list_of_lists = [abstracts[a]['author'] for a in abstracts]
originalList = set([val for sublist in list_of_lists for val in sublist])

authorList = set()
while originalList:
    authorA = originalList.pop()
    deleteSet = set()
    for authorB in originalList:
        ratio = difflib.SequenceMatcher(None, authorA, authorB).ratio()
        if ((ratio > 0.9) & (ratio != 1.0)):
            print('authorA:', authorA)
            print('authorB:', authorB)
            print(' Ratio:', ratio)
            deleteSet.add(authorB)
    originalList.difference_update(deleteSet)
    authorList.add(authorA)
    print(len(authorList),'+',len(originalList),'=',len(authorList)+len(originalList))


# In[ ]:


import networkx as nx
import json
from networkx.readwrite import json_graph
from tqdm import tqdm

# Use this to find the closest authors (assuming the authorList is pre-pruned)
def findClosestAuthor(name,authorList):
    maxRatio = 0.0
    outAuthor = None
    for author in authorList:
        ratio = difflib.SequenceMatcher(None, author, name).ratio()
        if (ratio > maxRatio):
            outAuthor = author
            maxRatio = ratio
    return outAuthor

# Loop through each abstract and add edges between the authors' names
# and the title of the abstract.
G = nx.Graph()
for a in tqdm(abstracts):
    title = abstracts[a]['title']
    
    for name in abstracts[a]['author']:
        author = findClosestAuthor(name, authorList)
        G.add_edge(author, title)
        if not 'group' in G.node[author]:
            G.node[author]['group'] = 1
        
    firstAuthor = findClosestAuthor(abstracts[a]['author'][0], authorList)
    G.node[firstAuthor]['group'] = 2
    G.node[title]['group'] = 3

# Loop through all the nodes and add the name property for D3.js
for n in G:
    G.node[n]['name'] = n
    
# Remove parallel edges
G = nx.Graph(G)
# Remove self loops
G.remove_edges_from(G.selfloop_edges())

# Export for D3.js
d = json_graph.node_link_data(G) # node-link format to serialize
json.dump(d, open('html/force.json','w'))

# Use NetworkX to plot the data (I've had limited success)
#import matplotlib.pyplot as plt
#pos=nx.spring_layout(G) # positions for all nodes
#nx.draw_networkx_nodes(G,pos,node_size=2,alpha=0.5)
#nx.draw_networkx_edges(G,pos,alpha=0.25)
#nx.draw_networkx_labels(G,pos,font_size=7,font_family='sans-serif')
#plt.axis('off')
#plt.show() # display


# In[ ]: