import numpy
from urllib.request import urlopen
import scipy.optimize
import random
from sklearn.decomposition import PCA
from collections import defaultdict

### PCA on beer reviews ###

def parseDataFromURL(fname):
  for l in urlopen(fname):
    yield eval(l)

def parseData(fname):
  for l in open(fname):
    yield eval(l)

print("Reading data...")
data = list(parseData("beer_50000.json"))
print("done")

X = [[x['review/overall'], x['review/taste'], x['review/aroma'], x['review/appearance'], x['review/palate']] for x in data]

pca = PCA(n_components=5)
pca.fit(X)
print(pca.components_)

### Network visualization ###
import networkx as nx
import matplotlib.pyplot as plt

# Karate club
G = nx.karate_club_graph()
nx.draw(G)
plt.show()
plt.clf()

edges = set()
nodes = set()
for edge in urllib.urlopen("data/facebook/egonet.txt", 'r'):
  x,y = edge.split()
  x,y = int(x),int(y)
  edges.add((x,y))
  edges.add((y,x))
  nodes.add(x)
  nodes.add(y)

G = nx.Graph()
for e in edges:
  G.add_edge(e[0],e[1])
nx.draw(G)
plt.show()
plt.clf()

### Find all 3 and 4-cliques in the graph ###
cliques3 = set()
cliques4 = set()
for n1 in nodes:
  for n2 in nodes:
    if not ((n1,n2) in edges): continue
    for n3 in nodes:
      if not ((n1,n3) in edges): continue
      if not ((n2,n3) in edges): continue
      clique = [n1,n2,n3]
      clique.sort()
      cliques3.add(tuple(clique))
      for n4 in nodes:
        if not ((n1,n4) in edges): continue
        if not ((n2,n4) in edges): continue
        if not ((n3,n4) in edges): continue
        clique = [n1,n2,n3,n4]
        clique.sort()
        cliques4.add(tuple(clique))