concatenate_files.py

import fnmatch
import logging
import os
import sys

import yaml

CLIPBOARD_SIZE = 14500
RESERVED_MESSAGE = "\n --- This is not the complete content. To be continued. ---"
RESERVED_SIZE = len(RESERVED_MESSAGE.encode('utf-8'))
TEMP_FILE_NAME = 'temp_state.txt'

# Configure the logging settings
logging.basicConfig(filename='concatenate_files.log', level=logging.DEBUG)


def gather_files(fileset_name, filesets):
    # Load filesets if it's a path to a YAML file
    if isinstance(filesets, str):
        with open(filesets, 'r') as f:
            filesets = yaml.safe_load(f)

    # Get the file patterns for the specified fileset_name
    patterns = filesets.get(fileset_name, [])

    included_patterns = [p for p in patterns if not p.startswith('!')]
    excluded_patterns = [p[1:] for p in patterns if p.startswith('!')]

    matching_files = []

    # Walk through all files recursively starting from the current directory
    for dirpath, dirnames, filenames in os.walk('.'):
        for filename in filenames:
            full_path = os.path.join(dirpath, filename)
            relative_path = os.path.relpath(full_path)

            # Check if file matches any of the inclusion patterns
            include_file = any(fnmatch.fnmatch(relative_path, pattern) for pattern in included_patterns)

            # Check if file matches any of the exclusion patterns
            exclude_file = any(fnmatch.fnmatch(relative_path, pattern) for pattern in excluded_patterns)

            if include_file and not exclude_file:
                matching_files.append(relative_path)
                logging.debug(f"Included file: {relative_path}")

    return matching_files


def concatenate_files_for_set(fileset_name, filesets):
    temp_file_path = f"temp_{fileset_name}.txt"
    state_file_path = "state.yml"

    # Reserve bytes for the message at the end
    actual_buffer_size = CLIPBOARD_SIZE - len(RESERVED_MESSAGE.encode('utf-8'))

    # Load previous state
    current_part = 1
    if os.path.exists(state_file_path):
        with open(state_file_path, 'r') as state_file:
            state = yaml.safe_load(state_file)
            if state['fileset_name'] == fileset_name:
                current_part = state['current_part']
            else:
                # Reset the state for a different fileset
                if os.path.exists(temp_file_path):
                    os.remove(temp_file_path)
                current_part = 1

    # If the temporary file doesn't exist, it means this is the first call for this fileset.
    # Create the concatenated file from scratch.
    if not os.path.exists(temp_file_path):
        with open(temp_file_path, 'w', encoding='utf-8') as f:
            for filepath in gather_files(fileset_name, filesets):
                logging.info(f"Concatenating file: {filepath}")
                f.write(f"--- {filepath}\n")
                with open(filepath, 'r', encoding='utf-8') as file_content:
                    f.write(file_content.read())
                    f.write("\n")

    # Determine total parts
    total_size = os.path.getsize(temp_file_path)
    total_parts = max((total_size + CLIPBOARD_SIZE - 1) // CLIPBOARD_SIZE, 1)

    # Log relevant information
    logging.info(f"Fileset: {fileset_name}, Current Part: {current_part}, Total Parts: {total_parts}")

    # Read the appropriate part
    with open(temp_file_path, 'rb') as f:
        f.seek((current_part - 1) * CLIPBOARD_SIZE)
        output = f.read(actual_buffer_size).decode('utf-8')
        if current_part < total_parts:
            output += RESERVED_MESSAGE

    # Update state
    with open(state_file_path, 'w') as state_file:
        state = {
            'fileset_name': fileset_name,
            'current_part': current_part + 1
        }
        yaml.safe_dump(state, state_file)

    # If it's the last part, cleanup
    if current_part == total_parts:
        if os.path.exists(temp_file_path):
            os.remove(temp_file_path)
        if os.path.exists(state_file_path):
            os.remove(state_file_path)

    # Print the content to be copied to the clipboard
    print(output)

    if total_parts > 1:
        print(f"Part {current_part} of {total_parts}. Please run this command again to continue.", file=sys.stderr)


if __name__ == '__main__':
    if len(sys.argv) > 1 and sys.argv[1] == 'abort-cp':
        if os.path.exists(TEMP_FILE_NAME):
            os.remove(TEMP_FILE_NAME)
        print("Copy process aborted!")
    else:
        with open('fileset-definitions.yml', 'r') as f:
            filesets = yaml.safe_load(f)

        fileset_name = sys.argv[1]
        concatenate_files_for_set(fileset_name, filesets)