Ejemplo n.º 1
0
def single_input(args, config=None):
    """
    Single Input Mode works to convert a single input XML file into EPUB.

    This is probably the most typical use case and is the most highly
    configurable, see the argument parser and oaepub --help
    """
    if config is None:
        config = get_config_module()
    #Determination of input type and processing
    #Fetch by URL
    if 'http:' in args.input:
        raw_name = u_input.url_input(args.input)
        abs_input_path = os.path.join(LOCAL_DIR, raw_name+'.xml')
        parsed_article = Article(abs_input_path, validation=args.no_dtd_validation)
    #Fetch by DOI
    elif args.input[:4] == 'doi:':
        raw_name = u_input.doi_input(args.input)
        abs_input_path = os.path.join(LOCAL_DIR, raw_name+'.xml')
        parsed_article = Article(abs_input_path, validation=args.no_dtd_validation)
    #Local XML input
    else:
        abs_input_path = utils.get_absolute_path(args.input)
        raw_name = u_input.local_input(abs_input_path)
        parsed_article = Article(abs_input_path, validation=args.no_dtd_validation)

    #Generate the output path name, this will be the directory name for the
    #output. This output directory will later be zipped into an EPUB
    output_name = os.path.join(utils.get_output_directory(args), raw_name)

    #Make the EPUB
    make_epub(parsed_article,
              outdirect=output_name,
              explicit_images=args.images,   # Explicit image path
              batch=False,
              config=config)

    #Cleanup removes the produced output directory, keeps the ePub file
    if args.clean:  # Defaults to False, --clean or -c to toggle on
        shutil.rmtree(output_name)

    #Running epubcheck on the output verifies the validity of the ePub,
    #requires a local installation of java and epubcheck.
    if args.no_epubcheck:
        epubcheck('{0}.epub'.format(output_name), config)
Ejemplo n.º 2
0
def main(argv=None):
    args = docopt(__doc__,
                  argv=argv,
                  version='OpenAccess_EPUB v.' + __version__,
                  options_first=True)

    c_file = args['COLLECTION_FILE']
    c_file_root = utils.file_root_name(c_file)
    abs_input_path = utils.get_absolute_path(c_file)

    if not args['--log-to']:
        log_to = os.path.join(os.path.dirname(abs_input_path),
                              c_file_root + '.log')
    else:
        log_to = args['--log-to']

    #Basic logging configuration
    oae_logging.config_logging(args['--no-log-file'], log_to,
                               args['--log-level'], args['--silent'],
                               args['--verbosity'])

    command_log = logging.getLogger('openaccess_epub.commands.collection')

    #Load the config module, we do this after logging configuration
    config = openaccess_epub.utils.load_config_module()

    #Quit if the collection file is not there
    if not os.path.isfile(c_file):
        command_log.critical('File does not exist {0}'.format(c_file))
        sys.exit('Unable to continue')

    command_log.info('Parsing collection file: {0}'.format(c_file))
    with open(c_file, 'r') as f:
        inputs = [line.strip() for line in f.readlines()]

    #Get the output directory
    if args['--output'] is not None:
        output_directory = utils.get_absolute_path(args['--output'])
    else:
        if os.path.isabs(config.default_output):  # Absolute remains so
            output_directory = config.default_output
        else:  # Else rendered relative to input
            abs_dirname = os.path.dirname(abs_input_path)
            output_directory = os.path.normpath(
                os.path.join(abs_dirname, config.default_output))

    output_directory = os.path.join(output_directory, c_file_root)
    command_log.info(
        'Processing collection output in {0}'.format(output_directory))

    if os.path.isdir(output_directory):
        utils.dir_exists(output_directory)
    try:
        os.makedirs(output_directory)
    except OSError as err:
        if err.errno != 17:
            command_log.exception(
                'Unable to recursively create output directories')

    #Instantiate collection NCX and OPF
    navigation = Navigation(collection=True)
    package = Package(collection=True, title=c_file_root)

    #Copy over the basic epub directory
    make_epub_base(output_directory)

    epub_version = None

    #Iterate over the inputs
    for xml_file in inputs:
        xml_path = utils.evaluate_relative_path(
            os.path.dirname(abs_input_path), xml_file)
        parsed_article = Article(xml_path,
                                 validation=not args['--no-validate'])
        if epub_version is None:  # Only set this once, no mixing!
            if args['--epub2']:
                epub_version = 2
            elif args['--epub3']:
                epub_version = 3
            else:
                epub_version = parsed_article.publisher.epub_default

        navigation.process(parsed_article)
        package.process(parsed_article)

        #Get the Digital Object Identifier
        doi = parsed_article.get_DOI()
        journal_doi, article_doi = doi.split('/')

        #Get the images
        openaccess_epub.utils.images.get_images(output_directory,
                                                args['--images'], xml_path,
                                                config, parsed_article)

        parsed_article.publisher.render_content(output_directory, epub_version)

    if epub_version == 2:
        navigation.render_EPUB2(output_directory)
        package.render_EPUB2(output_directory)
    elif epub_version == 3:
        navigation.render_EPUB3(output_directory)
        package.render_EPUB3(output_directory)
    epub_zip(output_directory)

    #Cleanup removes the produced output directory, keeps the EPUB
    if not args['--no-cleanup']:
        command_log.info('Removing {0}'.format(output_directory))
        shutil.rmtree(output_directory)

    #Running epubcheck on the output verifies the validity of the ePub,
    #requires a local installation of java and epubcheck.
    if not args['--no-epubcheck']:
        epub_name = '{0}.epub'.format(output_directory)
        openaccess_epub.utils.epubcheck(epub_name, config)
Ejemplo n.º 3
0
def main(argv=None):
    args = docopt(__doc__,
                  argv=argv,
                  version='OpenAccess_EPUB v.' + __version__,
                  options_first=True)

    c_file = args['COLLECTION_FILE']
    c_file_root = utils.file_root_name(c_file)
    abs_input_path = utils.get_absolute_path(c_file)

    if not args['--log-to']:
        log_to = os.path.join(os.path.dirname(abs_input_path),
                              c_file_root + '.log')
    else:
        log_to = args['--log-to']

    #Basic logging configuration
    oae_logging.config_logging(args['--no-log-file'],
                               log_to,
                               args['--log-level'],
                               args['--silent'],
                               args['--verbosity'])

    command_log = logging.getLogger('openaccess_epub.commands.collection')

    #Load the config module, we do this after logging configuration
    config = openaccess_epub.utils.load_config_module()

    #Quit if the collection file is not there
    if not os.path.isfile(c_file):
        command_log.critical('File does not exist {0}'.format(c_file))
        sys.exit('Unable to continue')

    command_log.info('Parsing collection file: {0}'.format(c_file))
    with open(c_file, 'r') as f:
        inputs = [line.strip() for line in f.readlines()]

    #Get the output directory
    if args['--output'] is not None:
        output_directory = utils.get_absolute_path(args['--output'])
    else:
        if os.path.isabs(config.default_output):  # Absolute remains so
            output_directory = config.default_output
        else:  # Else rendered relative to input
            abs_dirname = os.path.dirname(abs_input_path)
            output_directory = os.path.normpath(os.path.join(abs_dirname, config.default_output))

    output_directory = os.path.join(output_directory, c_file_root)
    command_log.info('Processing collection output in {0}'.format(output_directory))

    if os.path.isdir(output_directory):
        utils.dir_exists(output_directory)
    try:
        os.makedirs(output_directory)
    except OSError as err:
        if err.errno != 17:
            command_log.exception('Unable to recursively create output directories')

    #Instantiate collection NCX and OPF
    navigation = Navigation(collection=True)
    package = Package(collection=True, title=c_file_root)

    #Copy over the basic epub directory
    make_epub_base(output_directory)

    epub_version = None

    #Iterate over the inputs
    for xml_file in inputs:
        xml_path = utils.evaluate_relative_path(os.path.dirname(abs_input_path),
                                                xml_file)
        parsed_article = Article(xml_path, validation=not args['--no-validate'])
        if epub_version is None:  # Only set this once, no mixing!
            if args['--epub2']:
                epub_version = 2
            elif args['--epub3']:
                epub_version = 3
            else:
                epub_version = parsed_article.publisher.epub_default

        navigation.process(parsed_article)
        package.process(parsed_article)

        #Get the Digital Object Identifier
        doi = parsed_article.get_DOI()
        journal_doi, article_doi = doi.split('/')

        #Get the images
        openaccess_epub.utils.images.get_images(output_directory,
                                                args['--images'],
                                                xml_path,
                                                config,
                                                parsed_article)

        parsed_article.publisher.render_content(output_directory, epub_version)

    if epub_version == 2:
        navigation.render_EPUB2(output_directory)
        package.render_EPUB2(output_directory)
    elif epub_version == 3:
        navigation.render_EPUB3(output_directory)
        package.render_EPUB3(output_directory)
    epub_zip(output_directory)

    #Cleanup removes the produced output directory, keeps the EPUB
    if not args['--no-cleanup']:
        command_log.info('Removing {0}'.format(output_directory))
        shutil.rmtree(output_directory)

    #Running epubcheck on the output verifies the validity of the ePub,
    #requires a local installation of java and epubcheck.
    if not args['--no-epubcheck']:
        epub_name = '{0}.epub'.format(output_directory)
        openaccess_epub.utils.epubcheck(epub_name, config)