"""
Module: tokenization.refine_results.tokens_to_txt.py
====================================================
This module, a part of the `tokenization.refine_results` package, handles the conversion of tokenized data from a CSV file
to individual text files.
Functions
---------
- `tokenization_export_csv_columns_to_txt_file`: Exports a CSV file's contents to individual text files, with directories
for each column.
- `save_txt_files_to_directory`: Saves data from a dictionary into text files, organizing the files into directories based on keys.
Notes
-----
The module expects CSV files to have a specific structure, including a 'filename' column.
Please refer to the individual function docstrings for more detailed descriptions and examples of usage.
"""
import os
from datetime import datetime
import pandas as pd
from tqdm import tqdm
from iMaT.src.tokenization.utils import select_csv_file_2d_token_representation
from iMaT.src.utils.error_handling import handle_error
tokenizers_available_for_refining = ['CPWord', 'Octuple', 'OctupleMono', 'MuMIDI']
[docs]def tokenization_export_csv_columns_to_txt_file():
"""
Exports the columns of a CSV file to individual text files.
This function asks the user to select a CSV file, and then groups the DataFrame by filename. For each group,
it concatenates the values of each column into a string. Finally, it saves each column's concatenated string
into individual text files in directories named after each column. These directories are then bundled into
a single directory named 'extracted_data_[current date and time]'.
Parameters: None
Returns: None
See Also
--------
select_csv_file_2d_token_representation : Opens a file dialog allowing the user to select a CSV file.
save_txt_files_to_directory : Saves the refined data into text files in directories named after each column.
"""
try:
file_path = select_csv_file_2d_token_representation()
if file_path is None:
return
df = pd.read_csv(file_path)
# Initialize a dictionary to hold grouped column data
grouped_data = {}
# Check if the DataFrame contains the 'filename' column
if 'filename' in df.columns:
# Group the data by filename
grouped = df.groupby('filename')
# Loop through each group
for name, group in tqdm(grouped, desc='Grouping data', unit='group'):
column_data = {}
# Loop over the remaining columns in the group and concatenate the values into a string
for column in group.columns:
if column != 'filename':
column_data[column] = ' '.join(group[column].astype(str).values)
# Save the column data for the current group in the grouped_data dictionary
grouped_data[name] = column_data
# Save the data to the new text files
save_txt_files_to_directory(grouped_data, file_path)
else:
print("'filename' column does not exist in the DataFrame.")
except Exception as e:
handle_error(e)
[docs]def save_txt_files_to_directory(data, file_path):
"""
Saves a dictionary of data into text files in directories named after each key.
Each key-value pair in the data dictionary represents a filename and its associated data respectively.
For each filename, the function creates a directory and within that directory, it creates a text file for
each data column and writes the corresponding data into it. These directories are then bundled into a
single directory named 'extracted_data_[current date and time]'.
Parameters
----------
data : dict
The dictionary of data to be saved. Each key-value pair represents a filename and its associated data.
file_path : str
The original file path used to generate the new directory's name.
Returns
-------
str
The path to the newly created directory.
See Also
--------
os.path.dirname : Returns the directory component of a pathname.
os.makedirs : Recursively creates directories.
"""
try:
folder_path = os.path.dirname(file_path) # Get the directory path of the file
cleaned_csv_dir = os.path.join(folder_path, "extracted_data_" + datetime.now().strftime("%Y%m%d_%H%M%S"))
# Create new directory if it does not exist
os.makedirs(cleaned_csv_dir, exist_ok=True)
# Loop over the data
for filename, column_data in tqdm(data.items(), desc='Writing to file', unit='file'):
# Loop over each column data
for column_name, data in column_data.items():
# Create a new directory for the column
new_dir_path = os.path.join(cleaned_csv_dir, column_name)
os.makedirs(new_dir_path, exist_ok=True)
# Define the output file path
output_file_path = os.path.join(new_dir_path, f"{filename}_{column_name}.txt")
# Write the data to the output file
with open(output_file_path, 'w') as output_file:
output_file.write(data)
return cleaned_csv_dir
except Exception as e:
handle_error(e)