import itertools
import networkx as nx
import numpy as np

# https://www.dezyre.com/article/top-10-machine-learning-algorithms/202
algos_list1 = [
    "Naive Bayes",
    "k-Means",
    "SVM",
    "Apriori",
    "Linear Regression",
    "Logistic Regression",
    "ANN",
    "Random Forests",
    "Decision Tree",
    "kNN"
]

# https://www.quora.com/What-are-the-top-10-data-mining-or-machine-learning-algorithms
algos_list2 = [
    "Decision Trees ",
    "k-Means",
    "SVM",
    "Apriori",
    "Expectation Maximization",
    "PageRank",
    "AdaBoost",
    "kNN",
    "Naive Bayes",
]

algos_list3 = [
    "Naive Bayes",
    "k-Means",
    "Kernel PCA",
    "Linear Regression",
    "kNN",
    "NNMF",
    "SVM",
    "Dimensionality Reduction",
    "SVD",
    "Decision Tree",
    "Bootstapped SVM",
    "Decision Tree",
    "Gaussian Processes",
    "Logistic Regression",
    "Logit Boost",
    "Model Tree",
    "PLS",
    "Random Forests",
    "Ridge Regression",
]

algos_list4 = [
    "Linear Regression",
    "Logistic Regression",
    "k-Means",
    "SVM",
    "Random Forests",
    "SVD",
    "Decision Tree",
    "Naive Bayes",
    "ANN",
    "Bayesian Networks",
    "Elastic Nets",
    "LDA",
    "Conditional Random Fields",
    "HDPs"
]

algos_list6 = [
    "SVM",
    "ANN",
    "Logistic Regression",
    "Naive Bayes",
    "KNN",
    "Random Forests",
    "Decision Tree",
    "Bagged Tree"
]

algos_list7 = [
    "Linear Regression",
    "Logistic Regression",
    "SVM",
    "SVD",
    "PCA",
    "Kernel PCA",
    "k-Means",
    "Decision Trees",
    "Random Forests",
    "Neural Networks",
    "Regularization",
    "Boosting",
    "Naive Bayes"
]


algos_list8 = [
    "Convolution Neural Network",
    "RNN",
    "Deep Q-Learning",
    "Deep Neural Network",
    "Random Forests",
    "Probabilistic Graphical Models",
    "Compressed Sensing",
    "Kernel Machines",
    "Counter Factual Regret Minimization",
    "Gaussian Processes",
]

algos_list9 = [
    "k-Means",
    "KNN",
    "Dimensionality Reduction",
    "PCA",
    "Collaborative filtering",
    "Association Rules",
    "Logistic Regression",
    "LDA",
    "Shortest Path",
    "PageRank"
]

algos_list10 = [
    "Linear Regression",
    "Logistic Regression",
    "k-Means",
    "SVM",
    "Random Forests",
    "SVD",
    "Boosted Trees",
    "Naive Bayes",
    "ANN",
    "AdaBoost"
]

algos_list11 = [
    "Apriori",
    "FPGrowth",
    "GSP",
    "PrefixSpan",
    "k-Means",
    "Regression Tree",
    "Decision Tree",
    "SVM",
    "PageRank",
    "Naive Bayes"
]

algos_list12 = [
    "Decision Tree",
    "Apriori",
    "ANN",
    "SVM",
    "SOM",
    "Genetic Algorithms",
    "Naive Bayes",
    "Ant Colony Optimization",
    "Linear Regression"
]

# https://bickson.blogspot.in/2011/06/what-are-most-widely-deployed-machine.html?spref=tw
algos_list5 = [
    "SVD",
    "k-Means",
    "Naive Bayes",
    "Dirichlet clustering",
    "Matrix Factorization",
    "Frequent Pattern Mining",
    "LDA",
    "Expectation Maximization",
    "SVM",
    "Decision Tree",
    "Logistic Regression",
    "Random Forests"
]

all_algos = [
    algos_list1,
    algos_list2,
    algos_list3,
    algos_list4,
    algos_list5,
    algos_list6,
    algos_list7,
    algos_list8,
    algos_list9,
    algos_list10,
    algos_list11,
    algos_list12
]


items_list = list(set(itertools.chain(*all_algos)))
id2items_dict = {i+1: x for i,x in enumerate(items_list)}
items2id_dict = {x: i+1 for i,x in enumerate(items_list)}

N = len(items_list)
rank_counts = np.zeros(shape=(N, N))

for items in all_algos:
    for i, item in enumerate(items):
        rank = i+1
        rank_counts[rank-1, items2id_dict[item]-1] += 1

G = nx.Graph()
for item in items_list:
    for i in range(N):
        rank = i+1
        weight = rank_counts[rank-1, items2id_dict[item]-1]
        if weight > 0.0:
            # print((item, rank, {'weight': weight}))
            G.add_edge(item, rank, attr_dict={'weight': weight})

matchings = nx.max_weight_matching(G, maxcardinality=True)

for i in range(1,20):
    print("Rank %2d: %s" % (i, matchings[i]))