Merge branch 'kiene/md5-script-for-characterization-2' into 'main'

#! /usr/bin/env python3

#   (C) 2022-2024 IVAS codec Public Collaboration with portions copyright Dolby International AB, Ericsson AB,

#   Fraunhofer-Gesellschaft zur Foerderung der angewandten Forschung e.V., Huawei Technologies Co. LTD.,

#   Koninklijke Philips N.V., Nippon Telegraph and Telephone Corporation, Nokia Technologies Oy, Orange,

#   Panasonic Holdings Corporation, Qualcomm Technologies, Inc., VoiceAge Corporation, and other

#   contributors to this repository. All Rights Reserved.

#   This software is protected by copyright law and by international treaties.

#   The IVAS codec Public Collaboration consisting of Dolby International AB, Ericsson AB,

#   Fraunhofer-Gesellschaft zur Foerderung der angewandten Forschung e.V., Huawei Technologies Co. LTD.,

#   Koninklijke Philips N.V., Nippon Telegraph and Telephone Corporation, Nokia Technologies Oy, Orange,

#   Panasonic Holdings Corporation, Qualcomm Technologies, Inc., VoiceAge Corporation, and other

#   contributors to this repository retain full ownership rights in their respective contributions in

#   the software. This notice grants no license of any kind, including but not limited to patent

#   license, nor is any license granted by implication, estoppel or otherwise.

#   Contributors are required to enter into the IVAS codec Public Collaboration agreement before making

#   contributions.

#   This software is provided "AS IS", without any express or implied warranties. The software is in the

#   development stage. It is intended exclusively for experts who have experience with such software and

#   solely for the purpose of inspection. All implied warranties of non-infringement, merchantability

#   and fitness for a particular purpose are hereby disclaimed and excluded.

#   Any dispute, controversy or claim arising under or in relation to providing this software shall be

#   submitted to and settled by the final, binding jurisdiction of the courts of Munich, Germany in

#   accordance with the laws of the Federal Republic of Germany excluding its conflict of law rules and

#   the United Nations Convention on Contracts on the International Sales of Goods.

import argparse

from pathlib import Path

from hashlib import md5

from collections import Counter

def get_hash_line_for_file(file: Path, experiment_dir: Path):

    with open(file, "rb") as f:

        hash = md5(f.read()).hexdigest()

    filepath = file.relative_to(experiment_dir)

    # always print forward slashes even on windows to be able to diff the result files

    hashline = f"{str(filepath.as_posix())} {hash}\n"

    return hashline

def main(experiment_dir, out_file):

    wav_files = sorted(experiment_dir.glob("proc_output*/**/*c[0-9][0-9].wav"))

    hashlines = [get_hash_line_for_file(f, experiment_dir) for f in wav_files]

    count = Counter([line.split()[-1] for line in hashlines])

    duplicates = [line for line in hashlines if count[line.split()[-1]] != 1]

    if len(duplicates) != 0:

        print("Found duplicate hashes in these lines:")

        for dup in duplicates:

            print(dup)

    with open(out_file, "w") as f:

        f.writelines(hashlines)

if __name__ == "__main__":

    parser = argparse.ArgumentParser(

        description="Cross-platform script for generating MD5 hashes of output files for the characterization testing experiments."

    )

    parser.add_argument(

        "experiment_dir",

        type=Path,

        help="Directory of the respective experiment (e.g. experiments/characterization/P800-12",

    )

    parser.add_argument(

        "out_file", type=Path, help="Output text file with filenames and Hashes"

    )

    args = parser.parse_args()

    main(args.experiment_dir, args.out_file)