logistic

import numpy as np
import tensorflow as tf


def file_len(fname):
    with open(fname) as f:
        for i, l in enumerate(f):
            pass

    return i + 1


def convertData(digit):
    one_hot = np.zeros([10])
    one_hot[int(digit)] = 1

    return one_hot


def getData(filename):
    file_name_queue = tf.train.string_input_producer([filename])
    reader = tf.TextLineReader()
    key, value = reader.read(file_name_queue)
    record_defaults = np.zeros([401, 1]).tolist()
    data = tf.decode_csv(value, record_defaults=record_defaults)
    features = tf.stack(data[:400])
    labels = tf.stack(data[400])

    x = np.array([])
    y = np.array([])
    with tf.Session() as sess:
        coord = tf.train.Coordinator()
        threads = tf.train.start_queue_runners(coord=coord)

        for i in range(file_len(filename)):
            tx, ty = sess.run([features, labels])
            if ty == -1:
                continue
            ty = convertData(ty)
            if len(x) == 0:
                x = np.array([tx])
                y = np.array([ty])
            else:
                x = np.vstack((x, tx))
                y = np.vstack((y, ty))

        coord.request_stop()
        coord.join(threads)

    return (x, y)


def sigmoid(z):
    exp_node = tf.exp(z)
    adder_node = 1 + exp_node
    result_node = 1.0 / adder_node

    return result_node


def logistic():
    W = tf.Variable(tf.zeros([40,40]), tf.float32)
    b = tf.Variable(tf.zeros([40]), tf.float32)
    x = tf.placeholder(tf.float32, [None, 40], name="x")

    linear_model = W * x + b
    hypothesis = sigmoid(linear_model)

    y = tf.placeholder(tf.float32)

    cost = -tf.reduce_mean(y * tf.log(hypothesis) +
                           (1 - y) * tf.log(1 - hypothesis))

    optimizer = tf.train.GradientDescentOptimizer(0.1)
    train = optimizer.minimize(cost)

    x_train = [3, 3.5, 4.1, 4.3, 5.0, 5.2, 5.6, 6.0, 6.1, 6.5, 7.6, 7.9, 9.2, 9.9, 10.2]
    y_train = [0, 0, 0, 0, 0, 0, 1, 0, 1, 1, 0, 1, 1, 1, 1]

    tf.summary.FileWriter("logisticGraph", tf.InteractiveSession().graph)

    init = tf.global_variables_initializer()
    sess = tf.Session()
    sess.run(init)

    for i in range(20000):
        sess.run(train, {x: x_train, y: y_train})

        # evaluate training accuracy
        if i % 1000 == 0:
            curr_W, curr_b, curr_loss = sess.run([W, b, cost], {x: x_train, y: y_train})
            print("W: %s b: %s loss: %s" % (curr_W, curr_b, curr_loss))

    check = sess.run(hypothesis, {x: [1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 6.5, 7.0, 8.0, 9.0, 10.0]})
    print(check)


if __name__ == "__main__":
    a, b = getData("data.csv")
    print(a.shape)
    print(b.shape)