import matplotlib.pyplot as plt
import numpy as np
import cascade_creation
import convex_optimization
import algorithms
import rip_condition


def compare_greedy_and_lagrange_cs284r():
    """
    Compares the performance of the greedy algorithm on the
    lagrangian sparse recovery objective on the Facebook dataset
    for the CS284r project
    """
    G = cascade_creation.InfluenceGraph(max_proba = .8)
    G.import_from_file("../datasets/subset_facebook_SNAPnormalize.txt")
    A = cascade_creation.generate_cascades(G, p_init=.05, n_cascades=100)

    #Greedy
    G_hat = algorithms.greedy_prediction(G, A)
    algorithms.correctness_measure(G, G_hat, print_values=True)

    #Lagrange Objective
    G_hat = algorithms.recovery_l1obj_l2constraint(G, A,
            passed_function=convex_optimization.type_lasso,
            floor_cstt=.05, lbda=10)
    algorithms.correctness_measure(G, G_hat, print_values=True)


def compute_graph(graph_name, n_cascades, lbda, passed_function):
    """
    Test running time on different algorithms
    """
    G = cascade_creation.InfluenceGraph(max_proba=.7, min_proba=.2)
    G.import_from_file(graph_name)
    A = cascade_creation.generate_cascades(G, p_init=.05, n_cascades=n_cascades)

    if passed_function==algorithms.greedy_prediction:
        G_hat = algorithms.greedy_prediction(G, A)
    else:
        G_hat = algorithms.recovery_passed_function(G, A,
                passed_function=passed_function,
                floor_cstt=.1, lbda=lbda, n_cascades=n_cascades)
    algorithms.correctness_measure(G, G_hat, print_values=True)


def plot_watts_strogatz_graph():
    """
    plot information in a pretty way
    """
    plt.clf()
    fig = plt.figure(1)
    labels = [50, 100, 500, 1000, 2000, 5000]
    x = [np.log(50), np.log(100), np.log(500), np.log(1000), np.log(2000), np.log(5000)]
    sparse_recov = [.25, .32, .7, .82, .89, .92]
    max_likel = [.21, .29, .67, .8, .87, .9]
    lasso = [.07, .30, .46, .65, .86, .89]
    greedy = [.09, .15, .4, .63, .82, .92]

    fig, ax = plt.subplots()

    plt.axis((np.log(45), np.log(5500), 0, 1))
    plt.xlabel("Number of Cascades")
    plt.ylabel("F1 score")
    plt.grid(color="lightgrey")
    ax.plot(x, greedy, 'ko-', color="lightseagreen", label='Greedy')
    ax.plot(x, lasso, 'ko-', color="orange", label="Lasso")
    ax.plot(x, max_likel, 'ko-', color="cornflowerblue", label="MLE")
    ax.plot(x, sparse_recov, 'ko-', color="k", label="Our Method")
    plt.legend(loc="lower right")
    ax.set_xticks(x)
    ax.set_xticklabels(tuple(labels))
    plt.savefig("../paper/figures/"+"watts_strogatz.pdf")


def plot_ROC_curve(figure_name):
    """
    plot information in a pretty way
    """
    plt.clf()
    fig = plt.figure(1)
    #labels = [0, .00002, .002, .02, .2, .5]
    x_sparse = [.57, .6, .61, .76, .9]
    y_sparse = [.41, .4, .37, .16, .03]
    
    x_lasso = [.55, .56, .66]
    y_lasso = [.5, .43, .25]

    fig, ax = plt.subplots()

    plt.axis((np.log(45), np.log(5500), 0, 1))
    plt.xlabel("Number of Cascades")
    plt.ylabel("F1 score")
    plt.grid(color="lightgrey")
    ax.plot(x_lasso, y_lasso, 'ko-', color="orange", label="Lasso")
    ax.plot(x_sparse, y_sparse, 'ko-', color="k", label="Our Method")
    plt.legend(loc="lower right")
    ax.set_xticks(x)
    ax.set_xticklabels(tuple(labels))
    plt.savefig("../paper/figures/"+figure_name)


if __name__=="__main__":
    if 0:
        compute_graph("../datasets/watts_strogatz_300_30_point3.txt",
                    n_cascades=100, lbda=.01, passed_function=
                    #convex_optimization.sparse_recovery)
                    #algorithms.greedy_prediction)
                    convex_optimization.sparse_recovery)
    if 0:
        compute_graph("../datasets/powerlaw_200_30_point3.txt",
            n_cascades=300, lbda=.002, passed_function=
            convex_optimization.sparse_recovery)
            #algorithms.greedy_prediction)
            #convex_optimization.type_lasso)
    if 1:
        compute_graph("../datasets/barabasi_albert_300_30.txt",
            n_cascades=100, lbda=.002, passed_function=
            convex_optimization.sparse_recovery)
            #algorithms.greedy_prediction)
            #convex_optimization.type_lasso)