dataprocessingcopy.py

# -*- coding: utf-8 -*-
"""dataprocessing.ipynb

Automatically generated by Colaboratory.

Original file is located at
    https://colab.research.google.com/drive/10At7vh21OGTlE-Myv1NhAHi7l7NwBocQ
"""

import pandas as pd
import numpy as np
import os
from zipfile import ZipFile
import re
import json
import base64

from google.colab import drive
drive.mount('/content/drive')

path = "/content/drive/MyDrive/Duke/huggingface_project/data"

df = pd.read_excel(path+"/questionnaire-data.xlsx", header=2)

df["vviq_score"] = np.sum(df.filter(like = "vviq"), axis = 1)
df["osiq_score"] = np.sum(df.filter(like = "osiq"), axis = 1)
df["treatment"] = np.where(df.vviq_score > 40, "control", "aphantasia")

df = df.rename(columns={
    "Sub ID": "sub_id",
    df.columns[5]: "art_ability",
    df.columns[6]: "art_experience",
    df.columns[9]: "difficult",
    df.columns[10]: "diff_explanation"
})

df.columns = df.columns.str.lower()

df = df.drop(df.filter(like="unnamed").columns, axis = 1)

df = df.drop(df.filter(regex="(vviq|osiq)\d+").columns, axis = 1)

df[df.columns[df.dtypes == "object"]] = df[df.columns[df.dtypes == "object"]].astype("string")

df[df.columns] = df[df.columns].replace([np.nan,pd.NA, "nan","na","NA","n/a","N/A","N/a"], None)

data = {}
for ind, row in df.iterrows():
  data[row["sub_id"]] = {
      "subject_id": int(row["sub_id"]),
      "treatment": row["treatment"],
      "demographics": dict(df.iloc[ind][1:-1])
  }
  data[row["sub_id"]]["demographics"]["art_ability"] = int(data[row["sub_id"]]["demographics"]["art_ability"])
  data[row["sub_id"]]["demographics"]["vviq_score"] = int(data[row["sub_id"]]["demographics"]["vviq_score"])
  data[row["sub_id"]]["demographics"]["osiq_score"] = int(data[row["sub_id"]]["demographics"]["osiq_score"])

stored_images = {}
with ZipFile(path + "/Images.zip", "r") as zip:
  for image_file in zip.namelist():
    with zip.open(image_file, 'r') as fil:
      im = fil.read()
      im_encoded = base64.b64encode(im).decode("utf-8")
      stored_images[image_file.removesuffix(".jpg")] = im_encoded

def get_sub_files(subject, file_list):
  pattern = re.compile("^.*" + subject + "-[a-z]{3}\d-(kitchen|livingroom|bedroom).*")
  sub_files = [f for f in file_list if pattern.match(f)]
  sub = {
    "kitchen": {
        "perception": "",
        "memory": ""
    },
    "livingroom": {
        "perception": "",
        "memory": ""
    },
    "bedroom": {
        "perception": "",
        "memory": ""
    },
  }

  for fil in sub_files:
    if "kitchen" in fil:
      if "pic" in fil:
        sub["kitchen"]["perception"] = fil
      else:
        sub["kitchen"]["memory"] = fil
    elif "livingroom" in fil:
      if "pic" in fil:
        sub["livingroom"]["perception"] = fil
      else:
        sub["livingroom"]["memory"] = fil
    else:
      if "pic" in fil:
        sub["bedroom"]["perception"] = fil
      else:
        sub["bedroom"]["memory"] = fil
  return sub

with ZipFile(path + "/Aphantasia-Drawings.zip", "r") as zip:
  files = zip.namelist()
  aphan_subs = list({f.split("/")[0] for f in files})
  aphantasia_drawing_dataset = {}
  for s in aphan_subs:
    if int(s[3:]) in data.keys():
      data[int(s[3:])]["drawings"] = get_sub_files(s, files)
    else:
      data[int(s[3:])] = {"drawings": get_sub_files(s,files)}

with ZipFile(path + "/Control-Drawings.zip", "r") as zip:
  files = zip.namelist()
  cntrl_subs = list({f.split("/")[0] for f in files})
  full_control = {}
  for s in cntrl_subs:
    if int(s[3:]) in data.keys():
      data[int(s[3:])]["drawings"] = get_sub_files(s, files)
    else:
      data[int(s[3:])] = {"drawings": get_sub_files(s,files)}

stored_images["kitchen"] = stored_images.pop('high_sun_ajwbpqrwvknlvpeh')
stored_images["bedroom"] = stored_images.pop('low_sun_acqsqjhtcbxeomux')
stored_images["livingroom"] = stored_images.pop('low_sun_byqgoskwpvsbllvy')

def extract_images(subject, treatment):
  images_bytes = {
      "kitchen": {
          "perception": "",
          "memory": ""
      },
      "livingroom": {
          "perception": "",
          "memory": ""
      },
      "bedroom": {
          "perception": "",
          "memory": ""
      }
  }
  for room in ["kitchen", "livingroom", "bedroom"]:
    paths = data[subject]["drawings"].get(room).values()
    paths = [p for p in paths if p != ""]
    if treatment == "aphantasia":
      with ZipFile(path + "/Aphantasia-Drawings.zip", "r") as zip:
        for filename in paths:
          with zip.open(filename, 'r') as fil:
            im = fil.read()
            im_encoded = base64.b64encode(im).decode("utf-8")
            if "mem" in filename:
              images_bytes[room]["memory"] = im_encoded
            else:
              images_bytes[room]["perception"] = im_encoded
    else:
      with ZipFile(path + "/Control-Drawings.zip", "r") as zip:
        for filename in paths:
          with zip.open(filename, 'r') as fil:
            im = fil.read()
            im_encoded = base64.b64encode(im).decode("utf-8")
            if "mem" in filename:
              images_bytes[room]["memory"] = im_encoded
            else:
              images_bytes[room]["perception"] = im_encoded

  return images_bytes

missing = []
for i in data.keys():
  if "drawings" in data[i] and "treatment" in data[i]:
    data[i]["drawings"] = extract_images(i,data[i]["treatment"])
  else:
    missing.append(i)

for num in missing:
  data.pop(num, None)

for sub in data.keys():
  data[sub]["image"] = stored_images

subject_data_path = path + "/clean_data.json"

#with open(subject_data_path, "w", encoding="utf-8") as sub_data:
  #json.dump(data, sub_data, indent=2)

type(data)

da = pd.DataFrame(data)





flattened_data = []

for key, value in data.items():
    flattened_subject = pd.json_normalize(value, sep='_')
    flattened_data.append(flattened_subject)


da = pd.concat(flattened_data, ignore_index=True)

# Save the DataFrame to a Parquet file
da.to_parquet(path + 'data.parquet')