merge_imudata.py

import pandas as pd
import os.path
#from config import *

GESTURES = ["walk","still","flap"] #different names
MAIN_GESTURE= ["walk","still","flap"] # e.g walk
clean_dir = 'data_pipeline/clean_test_data/'
merge_dir = 'data_pipeline/merge_data/'
gesture = "r"
file_holder = []

# Merge the separate recordings into a complete file
def merge_data(gesture_name):
  dfs = pd.DataFrame()
  for path, currentDirectory, files in os.walk(clean_dir):
  # Find all the files same with the gesture name in the directory
    for file in files:
      if file.startswith(gesture_name):
        file_holder.append(file)
    # Concat the separate files (recordings) into a merged one
    for i in range(len(file_holder)):
      dir = os.path.join(path, str(file_holder[i]))
      df = pd.read_csv(dir, index_col=[0])
      df["activity"] = gesture_name
      dfs = dfs.append(df).reset_index(drop=True)
    # Save the separate files from one recording into one file
    mergeName = os.path.join(merge_dir, gesture_name + ".csv")
    dfs.to_csv(mergeName)

    # return the path of the merged data file
    return mergeName

def complete_file(gesture_lst, name):
  dfs = pd.DataFrame()
  for i in gesture_lst:
    dir = os.path.join(merge_dir, i + ".csv")
    df = pd.read_csv(dir, index_col=[0])
    dfs =dfs.append(df).reset_index(drop=True)
  new_name = os.path.join(merge_dir, name + ".csv")
  dfs.to_csv(new_name)

  return new_name



if __name__ == "__main__":
# lst = merge_data("walk")
# lst2 = merge_data("still")
# df = pd.read_csv(lst)
 dfs = complete_file(["walk","still","flap"],"walk_still_flap")