import os 
import pandas as pd
import numpy as np
import seaborn as sns

targets_path=r"C:\Users\jeuux\Desktop\Carrera\MoAI\TFM\AnnotatedData\Accelerometer_Data\Datasets\HAR_Dataset\targets.npy"
dataset_path=r"C:\Users\jeuux\Desktop\Carrera\MoAI\TFM\AnnotatedData\Accelerometer_Data\Datasets\HAR_Dataset\data.npy"

from sklearn.model_selection import train_test_split
#Train/val/test
def split_data(data,targets,test_size):
    train_idx,test_idx=train_test_split(range(len(data)),test_size=test_size,random_state=42)
    samples_train=data[train_idx]
    targets_train=targets[train_idx]

    samples_test=data[test_idx]
    targets_test=targets[test_idx]
    
    return samples_train,targets_train,samples_test,targets_test

def save_data(folder,data,targets):
    np.save(os.path.join(folder,"data"),data)
    np.save(os.path.join(folder,"targets"),targets)

samples_train,targets_train,samples_test,targets_test=split_data(data,targets,0.2)
samples_train,targets_train,samples_val,targets_val=split_data(samples_train,targets_train,0.1)

train_folder=r"C:\Users\jeuux\Desktop\Carrera\MoAI\TFM\AnnotatedData\Accelerometer_Data\Datasets\HAR_Dataset_raw\Train"
test_folder=r"C:\Users\jeuux\Desktop\Carrera\MoAI\TFM\AnnotatedData\Accelerometer_Data\Datasets\HAR_Dataset_raw\Test"
val_folder=r"C:\Users\jeuux\Desktop\Carrera\MoAI\TFM\AnnotatedData\Accelerometer_Data\Datasets\HAR_Dataset_raw\Val"

save_data(train_folder,samples_train,targets_train)
save_data(test_folder,samples_test,targets_test)
save_data(val_folder,samples_val,targets_val)

participants_dir = r"C:\Users\jeuux\Desktop\Carrera\MoAI\TFM\AnnotatedData\Accelerometer_Data\Participants"

participant_folders = [os.path.join(participants_dir,participant) for participant in os.listdir(participants_dir)]

participants_train = participant_folders[:25]
participants_val  = participant_folders[25:]
# participants_val  = participant_folders[25:30]
# participants_test  = participant_folders[-5:]

def get_file_stem(path):
    base=os.path.basename(path)
    return os.path.splitext(base)[0]

def read_metadata(df_path):  
    #read df
    df = pd.read_csv(df_path,sep=" ",header= None)
    df.columns = ["video_path","frames","label"]
    return df

def df_to_txt(df,dir_path):
    df.to_csv(dir_path, header=None, index=None, sep=' ', mode='a')

participants_val

def get_data(data_folders,subset,dataset_dir):
    df_dict = {
            "video_path": [],
            "label": [],
            "frames": []
            }
    for folder in data_folders:
        participant = get_file_stem(folder)
        df_path = os.path.join(folder,"FullDataset",f"HARClips_dataset_{participant}.txt")
        with open(df_path,"rb") as f:
            df = json.load(f)
        for k,v in df.items():
            df_dict[k] = df_dict[k] + v
    df_path = os.path.join(dataset_dir,f"{subset}.txt")
#     with open(df_path,"w") as f:
#         json.dump(df_dict,f)
    df_to_txt(pd.DataFrame.from_dict(df_dict),df_path)
    print("Finished")
        
        
        
#         segments_file = os.path.join(folder,"FullDataset","segments_data.npy")
#         labels_file =  os.path.join(folder,"FullDataset","targets_data.npy")
#         segments.append(np.load(segments_file))
#         labels.append(np.load(labels_file))
#     segments = np.concatenate(segments)
#     labels = np.concatenate(labels)
#     if not(os.path.isdir(os.path.join(dataset_dir,subset))):
#         os.makedirs(os.path.join(dataset_dir,subset))
#     np.save(os.path.join(dataset_dir,subset,"data"),segments)
#     np.save(os.path.join(dataset_dir,subset,"targets"),labels)

def get_data(data_folders,subset,dataset_dir,dataset_root):
    segments = []
    labels = []
    for folder in data_folders:
        participant = get_file_stem(folder)
        segments_file = os.path.join(folder,dataset_root,"segments_data.npy")
        labels_file =  os.path.join(folder,dataset_root,"targets_data.npy")
        segments.append(np.load(segments_file))
        labels.append(np.load(labels_file))
        
    segments = np.concatenate(segments)
    labels = np.concatenate(labels)
    if not(os.path.isdir(os.path.join(dataset_dir,subset))):
        os.makedirs(os.path.join(dataset_dir,subset))
    np.save(os.path.join(dataset_dir,subset,"data"),segments)
    np.save(os.path.join(dataset_dir,subset,"targets"),labels)
    

folders  = [participants_train,participants_val]
subsets  =["Train","Val"]
dataset_root = "CompleteData"
dataset_dir  = r"C:\Users\jeuux\Desktop\Carrera\MoAI\TFM\AnnotatedData\Accelerometer_Data\Datasets\HAR_Dataset_participants"
for folders,subset in zip(folders,subsets):
    test = get_data(folders,subset,dataset_dir,dataset_root)

from sklearn.preprocessing import StandardScaler, LabelEncoder
import joblib

base_dir = r"C:\Users\jeuux\Desktop\Carrera\MoAI\TFM\AnnotatedData\Accelerometer_Data\Datasets\HAR_Dataset_participants\Train" 
targets = np.load(os.path.join(base_dir,"targets.npy"))
data = np.load(os.path.join(base_dir,"data.npy"))

encoder  = LabelEncoder().fit(targets)
scaler = StandardScaler().fit(data.reshape(-1,9))

df_path = r"C:\Users\jeuux\Desktop\Carrera\MoAI\TFM\AnnotatedData\FinalDatasets\Participants\0304d\FullDataset\HARClips_dataset_0304d.txt"

#read df
df = pd.read_csv(df_path,sep=" ",header= None)
# df.columns = ["video_path","frames","label"]

import json 
with open(df_path,"rb") as f:
    df= json.load(f)

df["video_path"]

dataset_path = r"C:\Users\jeuux\Desktop\Carrera\MoAI\TFM\AnnotatedData\FinalDatasets\Datasets\HAR_dataset_v1"

folders  = [participants_train,participants_test,participants_val]
subsets  =["Train","Test","Val"]
for folders,subset in zip(folders,subsets):
    test = get_data(folders,subset,dataset_path)

df = pd.DataFrame.from_dict(test)

df

base_dir = r"C:\Users\jeuux\Desktop\Carrera\MoAI\TFM\AnnotatedData\FinalDatasets\Datasets\HAR_AL"
for index,(id1,id2) in enumerate(zip(range(30,69,5),range(35,71,5))):
    
    if(id1==65):
        id2 = 68
        
    dataset_path = os.path.join(base_dir,f"HAR_dataset_AL_v{index}")
    if not(os.path.isdir(dataset_path)):
        os.makedirs(dataset_path)
        
    participant_list = participant_folders[id1:id2]
    get_data(participant_list,"Train",dataset_path)
    
             
    

#concat every subtable for each participant
participant_folders=[os.path.join(participant_folder,folder) for folder in os.listdir(participant_folder)]

for participant_folder in tqdm(participant_folders):  
    folder=os.path.join(participant_folder,data_type)
    table_arr=[]
    for table in os.listdir(folder):
        table_arr.append(pd.read_csv(os.path.join(folder,table)))
    df=pd.concat(table_arr)
    participant=os.path.basename(participant_folder)
    df.to_csv(os.path.join(folder,"{0}_raw_{1}.csv".format(data_type,participant)),index=False)
    print("Save table of participant : {}".format(participant))