import tensorflow as tf
import numpy as np
import boto3
import datetime
import os
import json

SAVER_FOLDER = "./saver"
BUCKET = 'BUCKET NAME'
TRAIN_DATA = "data-04-zoo.csv"
RESULT_FILE = 'result.json'

for file in os.listdir(SAVER_FOLDER):
    os.remove(SAVER_FOLDER + "/" + file);

s3_client = boto3.client('s3')
s3_client.download_file(BUCKET, TRAIN_DATA, TRAIN_DATA)

xy = np.loadtxt(TRAIN_DATA, delimiter=',', dtype=np.float32)
x_data = xy[:,0:-1]
y_data = xy[:,[-1]]
nb_classes = 7

X = tf.placeholder(tf.float32, [None, 16])
Y = tf.placeholder(tf.int32, [None, 1])

Y_one_hot = tf.one_hot(Y, nb_classes)
Y_one_hot = tf.reshape(Y_one_hot, [-1, nb_classes])

W = tf.Variable(tf.random_normal([16, nb_classes]), name='weight')
b = tf.Variable(tf.random_normal([nb_classes]), name='bias')

logits = tf.matmul(X,W) + b
hypothesis = tf.nn.softmax(logits)
cost_i = tf.nn.softmax_cross_entropy_with_logits(logits=logits, labels=Y_one_hot)
cost = tf.reduce_mean(cost_i)
optimizer = tf.train.GradientDescentOptimizer(learning_rate=0.1).minimize(cost)


prediction = tf.argmax(hypothesis, 1)
correct_prediction = tf.equal(prediction, tf.argmax(Y_one_hot, 1))
accuracy = tf.reduce_mean(tf.cast(correct_prediction, tf.float32))

saver = tf.train.Saver()

with tf.Session() as sess:
    sess.run(tf.global_variables_initializer())
    
    for step in range(2001):
        sess.run(optimizer, feed_dict={X: x_data, Y: y_data})
        if step % 100 == 0:
            print(step, sess.run([cost, accuracy], feed_dict={X: x_data, Y: y_data}))

    saver.save(sess,"./saver/save.{}.ckpt".format(datetime.datetime.now().strftime("%Y-%m-%d %H:%M:%S")))
    saver.save(sess,"./saver/save.last.ckpt")

    pred = sess.run(prediction, feed_dict={X: x_data})

    pw, pb = sess.run([W, b])
    result = {'W': pw.tolist(), 'b': pb.tolist()}
    print(result)

    with open(RESULT_FILE, 'w') as outfile:
        json.dump(result, outfile)

s3_client.upload_file(RESULT_FILE, 'dev-tensorflow-savedata', RESULT_FILE)

for file in os.listdir(SAVER_FOLDER):
    print(file)
    s3_client.upload_file(SAVER_FOLDER + "/" + file, 'dev-tensorflow-savedata', file)

import boto3
import json
import sys

import functional as F
import linear as L

BUCKET_NAME = 'BUCKET NAME'
KEY = 'result.json'
RESULT_FILE = '/tmp/' + KEY
S3 = boto3.resource('s3')

DEBUG = True
def debug_print(obj):
    if DEBUG:
        print(obj)
    

def get_predict_data():
    try:
        S3.Bucket(BUCKET_NAME).download_file(KEY, RESULT_FILE)
        with open(RESULT_FILE) as data_file:    
            return json.load(data_file)
    except:
        print(sys.exc_info()[0])
        return None
    return None

def get_x(event):
    get_querystring= F.curryr(F.get)('queryStringParameters')
    get_X = F.curryr(F.get)('X')

    return F.go(event, 
        get_querystring, 
        get_X,
        lambda x: json.loads(x),
        lambda x: [x])

def handler(event, context):
    debug_print(event)

    predict_data = get_predict_data()
    debug_print(predict_data)
    if predict_data is None:
        return {'statusCode': 500}

    X = get_x(event)
    if X is None:
        return {'statusCode': 404}

    W = predict_data['W']
    b = predict_data['b']

    H = L.matadd(L.matmul(X,W)[0],b)
    S = [L.sigmoid(x) for x in H]
    M = L.softmax(S)
    answer = L.argmax(M)

    debug_print([answer, M])
    debug_print(sum(M))

    body = { 'answer': answer, 'rating': M }

    response = {
        'statusCode': 200,
        'body': json.dumps(body)
    }

    return response

if __name__ == "__main__":
    event = {"queryStringParameters": {"X": "[0,0,1,0,0,1,1,1,1,0,0,1,0,1,0,0]"}}
    print(handler(event,None))