04.Placenta.4p_one_by_one.py

# 这个是把胎盘四个部位的每个部位分别预测，，并加上了PBMC作为参数
# 这个是把胎盘四个部位的每个部位分别预测，，并加上了PBMC作为参数
# 这个是把胎盘四个部位的每个部位分别预测，，并加上了PBMC作为参数
# 这个是把胎盘四个部位的每个部位分别预测，，并加上了PBMC作为参数
# 这个是把胎盘四个部位的每个部位分别预测，，并加上了PBMC作为参数


import tensorflow as tf
import utils
# import functools
import numpy as np
from sklearn import preprocessing
import keras_tuner as kt
from  matplotlib import  pyplot as plt
import pandas as pd
from scipy.stats import spearmanr, pearsonr
import os 

import random
seed=4
seed=7


def scale_x(x):
	# 预处理
	# 先对非0特征取log
	# 然后进行归一化
	min_max_scaler = preprocessing.MinMaxScaler()
	standar_scaler = preprocessing.StandardScaler()
	x[x!=0] = np.log(x[x!=0])
	x = standar_scaler.fit_transform(x)
	x = min_max_scaler.fit_transform(x)
	return x

def load_data(path, col_index):
    # col_index = 264
    # col_index = 265
    # col_index = 266
	raw_data = np.loadtxt(path, dtype=np.float32, delimiter=",", skiprows=1)
	return (raw_data[..., :264]), (raw_data[..., col_index])


def train_one_position(position_index):

    data_dir = "../01.train_data/4p/"
    train_x, train_y = load_data(data_dir + "train.csv", position_index)
    valid_x, valid_y = load_data(data_dir + "valid.csv", position_index)
    test_x, test_y   = load_data(data_dir +  "test.csv", position_index)




    def model_with_seed(seed):
        random.seed(seed)# 为python设置随机种子
        os.environ['PYTHONHASHSEED'] = str(seed)
        np.random.seed(seed)# 为numpy设置随机种子
        tf.compat.v1.set_random_seed(seed)# tf cpu fix seed
        os.environ['TF_DETERMINISTIC_OPS'] = '1'  # tf gpu fix seed


        # 263 -> TPM



        model = tf.keras.Sequential()

        model.add(tf.keras.layers.Dense(32, activation="relu"))

        for _ in range(4):
            model.add(tf.keras.layers.BatchNormalization())
            model.add(tf.keras.layers.Dense(32, activation="relu"
                )
            )
        
        model.add(tf.keras.layers.BatchNormalization())
        model.add(tf.keras.layers.Dense(1))

        model.compile(
            loss = "mse",
            optimizer = "adam"
        )



        stop_early = tf.keras.callbacks.EarlyStopping(monitor='val_loss', patience=15)

        y_pre_batch = []
        class logPredictPer100epoch(tf.keras.callbacks.Callback):
            def on_epoch_end(self, batch, logs=None):
                example_result = self.model.predict(test_x)
                y_pre_batch.append(tf.reshape(example_result,[-1]))


        history = model.fit(x=train_x,
                y=train_y,
                batch_size=512,
                validation_data=(valid_x, valid_y),
                epochs=40,
                callbacks=[stop_early, logPredictPer100epoch()]
                )
                
        loss = history.history['loss']
        val_loss = history.history['val_loss']

        example_batch = test_x
        example_result = model.predict(example_batch)
        
        return (model, history, loss, val_loss, example_batch, example_result, y_pre_batch)




    seed = 946063
    model, history, loss, val_loss, example_batch, example_result, y_pre_batch = model_with_seed(seed)






    res_dir = "res/04.Placenta.4p_one_by_one.py/" + str(position_index) + "/"

    if not os.path.exists(res_dir):
        os.makedirs(res_dir)

    np.savetxt(res_dir + "01.y_pre_history.csv", y_pre_batch, delimiter=",")
    pd.DataFrame({"loss": loss, "val_loss": val_loss}).to_csv(res_dir + "02.loss_history.csv")
    pd.DataFrame({"real_y":test_y, "pred_y": tf.reshape(example_result, [-1])}).to_csv(res_dir + "03.res.csv")



    # plt.scatter(test_y,tf.reshape(example_result, [-1]))
    # plt.savefig("123.pdf")
    # plt.savefig("123.tiff")


    # y_pre = pd.Series(tf.reshape(example_result, [-1]))
    # y = pd.Series(test_y)

    # print(y.corr(y_pre))
    # print(spearmanr(y_pre, y))
    # print(pearsonr(y_pre, y))
    return seed


res_seed = []
for i in [264, 265, 266]:
    res_seed.append(train_one_position(i))

print(res_seed)




# [742271, 987885, 189512]
# [762307, 430117, 144375]