tei2eoatex.py

#!/usr/bin/env python3


"""eoatei -> eoatex"""

__version__= "1.0"
__author__ = "samuel"
__date__="20191114"


from utils.load_config import load_config, exec_command, check_executable, copy_dir_overwrite, ToLog, ToFile
from utils.run_xslt import run_xslt

import logging
import argparse
from pathlib import Path
import os
import shutil

BASE_DIR = Path( __file__ ).resolve().parent
SCRIPT_PATH = Path( __file__ )
SCRIPT_NAME = SCRIPT_PATH.stem

DEFAULT_INPUT_DIR = \
    Path(os.environ['INPUT_DIR'] if 'INPUT_DIR' in os.environ else './input')

DEFAULT_OUTPUT_DIR = \
    Path(os.environ['OUTPUT_DIR'] if 'OUTPUT_DIR' in os.environ else './output')

DEFAULT_DEPENDENCIES_DIR = \
    Path(os.environ['DEPENDENCIES_DIR'] if 'DEPENDENCIES_DIR' in os.environ else './dependencies')

EOA_SCRIPTS_DIR = \
    Path(os.environ['EOA_SCRIPTS_DIR'])


def copy_files(
        input_dir,
        output_dir,
):
    for f in input_dir.iterdir():
        logging.debug( f"copy dir: {f}" )
        if f.is_dir():
            shutil.copytree(
                f,
                output_dir / f.name
            )
        else:
            shutil.copy(
                    f,
                    output_dir / f.name
            )

if __name__ == '__main__':

    parser = argparse.ArgumentParser(
            formatter_class=argparse.ArgumentDefaultsHelpFormatter
    )
    parser.add_argument(
            "-c", "--config",
            default = BASE_DIR / "config" / "eoaconvert.cfg",
            help="Name of config file"
    )
    parser.add_argument(
            "--log-level",
            default = "INFO",
            help="log level: choose between DEBUG, INFO, WARNING, ERROR, CRITICAL"
    )
    parser.add_argument(
            "-o", "--output-dir",
            help = f"output directory. default: {DEFAULT_OUTPUT_DIR}/PUBLICATION_NAME/eoatex",
            type = Path,
    )
    parser.add_argument(
            "-f", "--filename",
            default = Path("*.xml"),
            type = Path,
            help = "xml file inside PUBLICATION_DIR, or absolute path. Patterns like '*.xml' are also acceptable"
    )
    parser.add_argument(
            "-p", "--param",
            action = 'append',
            default = [],
            help="xslt params"
    )
    parser.add_argument(
            "-!", "--overwrite",
            action = "store_true",
            default = False,
            help="overwrite files at OUTPUT_DIR"
    )
    parser.add_argument(
            "PUBLICATION_DIR",
            help = "directory containing the publication (including resources like pictures, etc.)",
            type = Path,
    )

    args = parser.parse_args()

    publ_dir = args.PUBLICATION_DIR
    input_file = \
        args.filename if args.filename . is_absolute() else list(publ_dir . glob( str(args.filename) ))[0]
    OUTPUT_DIR = \
            args.output_dir if args.output_dir is not None else (DEFAULT_OUTPUT_DIR / publ_dir.resolve().stem) / "eoatex"
    log_dir = OUTPUT_DIR / "log"

    if not input_file.is_file():
        import errno
        raise( Exception(
            f"not a valid input file: {input_file}"
        ) )
    if OUTPUT_DIR.exists():
        if args.overwrite:
            shutil.rmtree( OUTPUT_DIR )
        else:
            raise( Exception( f"output directory already existing: '{OUTPUT_DIR}'!" ) )
    if not OUTPUT_DIR.exists():
            OUTPUT_DIR.mkdir(
                    parents=True
            )

    CONFIG_FILE = args.config
    CONFIG = load_config(
            CONFIG_FILE,
            args.log_level,
            (log_dir / SCRIPT_NAME) . with_suffix( ".log" ),
    )

    check_executable( "saxon" )

    output_file = OUTPUT_DIR / "main.tex"

    copy_files(
            input_dir = input_file.parent,
            output_dir = output_file.parent,
    )

    run_xslt(
            input_file = input_file,
            xslt_file = EOA_SCRIPTS_DIR / "stylesheets/tei2eoatex.xsl",
            output_file = output_file,
            params = args.param,
            exec_command_args = {
                'output_to' : ToFile( log_dir / SCRIPT_NAME / "saxon.log" )
            },
    )