DataPuller/DataPuller.py

#! /usr/bin/python3

import os
import argparse
import shutil
import filecmp

# Argument Parsing
parser = argparse.ArgumentParser(
    prog="DataPuller",
    description="Check differences between folders and update files if needed",
)

parser.add_argument("-i", "--input", help="Path to input folder")
parser.add_argument("-o", "--output", help="Path to output folder")
parser.add_argument(
    "-q",
    "--quiet",
    action="store_true",
    help="Suppress non-essential commandline output",
)
parser.add_argument(
    "-y", "--confirm", action="store_true", help="Auto-confirm user queries"
)
parser.add_argument(
    "-d", "--delete", action="store_true", help="Auto-confirm file deletion"
)
parser.add_argument(
    "--dry-run",
    action="store_true",
    help="List affected files without copying or deleting",
)
parser.add_argument(
    "-s",
    "--shallow-diff",
    action="store_true",
    help="Use filecmp.cmp() shallow mode for file diff (faster, but might miss changes or generate false positives)",
)

args = parser.parse_args()

# Paths
if not args.input:
    inpath = input("Source Path: ")
else:
    inpath = args.input
if not args.output:
    outpath = input("Target Path: ")
else:
    outpath = args.output


def get_files(path):
    file_info = []
    if not os.path.exists(path):
        return file_info

    for folder, _, files in os.walk(path):
        for file in files:
            full_path = os.path.join(folder, file)
            rel_path = os.path.relpath(full_path, path)
            file_info.append(rel_path)
    return file_info


def main():
    startstring = " FOLDER-SYNC 0.1.3 "
    if not args.quiet:
        print(
            len(startstring) * "=" + "\n" + startstring + "\n" + len(startstring) * "="
        )

    # Test Folder availability
    if not os.path.exists(inpath):
        print(f"ERROR: The input folder ({inpath}) is unavailable or non-existant.")
        if not args.quiet:
            input("\nPress ENTER to exit...")
        exit(1)

    if not os.path.exists(outpath):
        print(f"ERROR: The output folder ({outpath}] is unavailable or non-existant.")
        if not args.quiet:
            input("\nPress ENTER to exit...")
        exit(1)

    print("Analyzing files. Please wait...\n")

    source = get_files(inpath)
    target = get_files(outpath)

    new = []
    changed = []
    deleted = []

    # Check for file additions and changes
    for path in source:
        if path not in target:
            new.append(path)
        elif not filecmp.cmp(
            os.path.join(inpath, path),
            os.path.join(outpath, path),
            shallow=args.shallow_diff,
        ):
            changed.append(path)

    # Check for file deletions
    for path in target:
        if path not in source:
            deleted.append(path)

    if not new and not changed and not deleted:
        if not args.quiet:
            print("No changes found.")
            input("\nPress ENTER to exit...")
        exit(0)

    # Print overview
    if not args.quiet:
        print("Found the following changes:\n")
        if new:
            print("Added:")
            for i in new:
                print(f" + {i}")
            print()
        if changed:
            print("Changed:")
            for i in changed:
                print(f" ~ {i}")
        if deleted:
            print("Deleted:")
            for i in deleted:
                print(f" - {i}")

    # Copy and update files
    if not args.dry_run:
        if new or changed:
            if args.confirm:
                copy_confirm = "y"
            else:
                copy_confirm = (
                    input("Write new and changed files? (y/N): ").strip().lower()
                )
            if copy_confirm in ["ja", "j", "yes", "y"]:
                if not args.quiet:
                    print("\nCopying files...")
                for path in new + changed:
                    source_path = os.path.join(inpath, path)
                    target_path = os.path.join(outpath, path)

                    # Create new folders if needed
                    os.makedirs(os.path.dirname(target_path), exist_ok=True)

                    # Copy file
                    try:
                        shutil.copy(source_path, target_path)
                        if not args.quiet:
                            print(f"Successfully copied {path}")
                    except Exception as error:
                        print(f"Failed to copy {path} with error {error}")
                        if not args.confirm:
                            if not input("Continue? (y/N): ") in [
                                "ja",
                                "j",
                                "yes",
                                "y",
                            ]:
                                exit(1)
            elif not args.quiet:
                print("\nSkipping changes...")

    # Delete files
    if not args.dry_run:
        if deleted:
            alert = " ALERT: Some files only exist at the destination. "
            if not args.quiet:
                print(f"{len(alert) * '*'}\n{alert}\n{len(alert) * '*'}")
            if args.delete:
                delete_confirm = "y"
            else:
                delete_confirm = (
                    input("Delete files on the target? (y/N): ").strip().lower()
                )
            if delete_confirm in ["ja", "j", "yes", "y"]:
                if not args.quiet:
                    print("\nDeleting files...")
                for path in deleted:
                    target_path = os.path.join(outpath, path)
                    try:
                        os.remove(target_path)
                        if not args.quiet:
                            print(f"Successfully deleted {path}")
                    except Exception as error:
                        print(f"Failed to delete {path} with error {error}")
                        if not args.confirm:
                            if not input("Continue? (y/N): ") in [
                                "ja",
                                "j",
                                "yes",
                                "y",
                            ]:
                                exit(1)
            elif not args.quiet:
                print("\nSkipping deletion...")

    if not args.quiet:
        if not args.dry_run:
            print("\nFolder sync finished!")
        else:
            print("\nDry run complete!")
        input("\nPress ENTER to exit...")


if __name__ == "__main__":
    main()