Commit e0649221 authored by Vít Starý Novotný's avatar Vít Starý Novotný
Browse files

Replace image scratch

parent f0424619

ahisto_ocr/cli.py

+1 −1
Original line number Diff line number Diff line
@@ -50,7 +50,7 @@ def main(input_dir: str, output_dir: str, pero_ocr: bool, super_resolution: bool 
            LOGGER.info(f'Pulling image {name}')

            client.images.pull(name)



    check_docker_image('scratch')

    check_docker_image('gitlab.fi.muni.cz:5050/nlp/ahisto-modules/empty')

    check_docker_image('gitlab.fi.muni.cz:5050/nlp/ahisto-modules/ocr-eval')

    check_docker_image('gitlab.fi.muni.cz:5050/nlp/ahisto-modules/ocr-fileformat')

    check_docker_image('gitlab.fi.muni.cz:5050/nlp/ahisto-modules/ocr-tesseract')

ahisto_ocr/ocr.py

+3 −3
Original line number Diff line number Diff line
@@ -30,7 +30,7 @@ def run_tesseract(client, preprocessing_volume, second_pass_volume) -> None: 
            'mode': 'rw',

        },

    }

    with create_temporary_docker_container(client, 'scratch', command='cmd',

    with create_temporary_docker_container(client, 'gitlab.fi.muni.cz:5050/nlp/ahisto-modules/empty', command='cmd',

                                           volumes=volumes) as container:

        text = extract_text_file_from_container(container, '/input/list.txt')

        add_text_file_to_container(container, '/output/list.txt', text)
@@ -135,7 +135,7 @@ def run_pero_ocr(client, preprocessing_volume, pero_ocr_volume, gpus: str) -> No 
    run_docker_container(client, 'gitlab.fi.muni.cz:5050/nlp/ahisto-modules/ocr-pero', command=command, volumes=volumes,

                         environment={'LC_ALL': 'C'})



    with create_temporary_docker_container(client, 'scratch', command='cmd',

    with create_temporary_docker_container(client, 'gitlab.fi.muni.cz:5050/nlp/ahisto-modules/empty', command='cmd',

                                           volumes=volumes) as container:

        text = extract_text_file_from_container(container, '/input/list.txt')

        add_text_file_to_container(container, '/output/list.txt', text)
@@ -165,7 +165,7 @@ def combine_results(client, tesseract_volume, pero_ocr_volume, output_volume): 
    ]

    run_docker_container(client, 'gitlab.fi.muni.cz:5050/nlp/ahisto-modules/ocr-eval', command=command, volumes=volumes)



    with create_temporary_docker_container(client, 'scratch', command='cmd',

    with create_temporary_docker_container(client, 'gitlab.fi.muni.cz:5050/nlp/ahisto-modules/empty', command='cmd',

                                           volumes=volumes) as container:

        text = extract_text_file_from_container(container, '/pero-ocr/list.txt')

        add_text_file_to_container(container, '/output/list.txt', text)

ahisto_ocr/preprocessing.py

+2 −2
Original line number Diff line number Diff line
@@ -34,7 +34,7 @@ def apply_super_resolution(client, input_volume, postprocessing_volume, gpus: st 
    ]

    run_docker_container(client, 'gitlab.fi.muni.cz:5050/nlp/ahisto-modules/ocr-eval', command=command, volumes=volumes)



    with create_temporary_docker_container(client, 'scratch', command='cmd',

    with create_temporary_docker_container(client, 'gitlab.fi.muni.cz:5050/nlp/ahisto-modules/empty', command='cmd',

                                           volumes=volumes) as container:

        text = extract_text_file_from_container(container, '/input/list.txt')

        add_text_file_to_container(container, '/output/list.txt', text)
@@ -55,7 +55,7 @@ def apply_super_resolution(client, input_volume, postprocessing_volume, gpus: st 
                         environment={'NVIDIA_VISIBLE_DEVICES': gpus},

                         command=command, volumes=volumes)



    with create_temporary_docker_container(client, 'scratch', command='cmd',

    with create_temporary_docker_container(client, 'gitlab.fi.muni.cz:5050/nlp/ahisto-modules/empty', command='cmd',

                                           volumes=volumes) as container:

        filenames = [

            '{}_noise_scale.png'.format(Path(filename).with_suffix(''))

ahisto_ocr/volume.py

+2 −2
Original line number Diff line number Diff line
@@ -14,7 +14,7 @@ LOGGER = getLogger(__name__) 
def copy_input_to(client, volume, input_dir: Path, input_images: List[Path]) -> None:

    LOGGER.info(f'Copying {len(input_images)} input images from {input_dir}/ to Docker volume {volume.short_id}')

    volumes = {volume.name: {'bind': '/input', 'mode': 'rw'}}

    with create_temporary_docker_container(client, 'scratch', command='cmd',

    with create_temporary_docker_container(client, 'gitlab.fi.muni.cz:5050/nlp/ahisto-modules/empty', command='cmd',

                                           volumes=volumes) as container:

        with TemporaryFile('w+b') as tf:

            with tarfile.open(fileobj=tf, mode='w') as tar:
@@ -32,7 +32,7 @@ def copy_input_to(client, volume, input_dir: Path, input_images: List[Path]) -> 
def copy_output_from(client, volume, input_images: List[Path], output_dir: Path) -> None:

    LOGGER.info(f'Copying {len(input_images)} OCR texts from Docker volume {volume.short_id} to {output_dir}/')

    volumes = {volume.name: {'bind': '/output', 'mode': 'ro'}}

    with create_temporary_docker_container(client, 'scratch', command='cmd',

    with create_temporary_docker_container(client, 'gitlab.fi.muni.cz:5050/nlp/ahisto-modules/empty', command='cmd',

                                           volumes=volumes) as container:

        with TemporaryFile('w+b') as tf:

            for chunk in container.get_archive('/output')[0]: