io.py

#!/usr/bin/env python3
"""
Methods related to IO.

File: pyneuroml/io.py

Copyright 2024 NeuroML contributors
"""

import inspect
import logging
import os
import sys
import textwrap
import typing
from typing import Optional

import neuroml.loaders as loaders
import neuroml.writers as writers
from neuroml import NeuroMLDocument

import lems.model.model as lems_model
from pyneuroml.errors import FILE_NOT_FOUND_ERR
from pyneuroml.validators import validate_neuroml2

logger = logging.getLogger(__name__)
logger.setLevel(logging.INFO)


def read_neuroml2_file(
    nml2_file_name: str,
    include_includes: bool = False,
    verbose: bool = False,
    already_included: Optional[list] = None,
    optimized: bool = False,
    check_validity_pre_include: bool = False,
) -> NeuroMLDocument:
    """Read a NeuroML2 file into a `nml.NeuroMLDocument`

    :param nml2_file_name: file of NeuroML 2 file to read
    :type nml2_file_name: str
    :param include_includes: toggle whether files included in NML file should also be included/read
    :type include_includes: bool
    :param verbose: toggle verbosity
    :type verbose: bool
    :param already_included: list of files already included
    :type already_included: list
    :param optimized: toggle whether the HDF5 loader should optimise the document
    :type optimized: bool
    :param check_validity_pre_include: check each file for validity before including
    :type check_validity_pre_include: bool
    :returns: nml.NeuroMLDocument object containing the read NeuroML file(s)
    """
    if already_included is None:
        already_included = []

    logger.info("Loading NeuroML2 file: %s" % nml2_file_name)

    if not os.path.isfile(nml2_file_name):
        logger.critical("Unable to find file: %s!" % nml2_file_name)
        sys.exit(FILE_NOT_FOUND_ERR)

    if nml2_file_name.endswith(".h5") or nml2_file_name.endswith(".hdf5"):
        nml2_doc = loaders.NeuroMLHdf5Loader.load(nml2_file_name, optimized=optimized)
    else:
        nml2_doc = loaders.NeuroMLLoader.load(nml2_file_name)

    base_path = os.path.dirname(os.path.realpath(nml2_file_name))

    if include_includes:
        if verbose:
            logger.info(
                "Including included files (included already: {})".format(
                    already_included
                )
            )

        incl_to_remove = []
        for include in nml2_doc.includes:
            incl_loc = os.path.abspath(os.path.join(base_path, include.href))
            if incl_loc not in already_included:
                inc = True  # type: typing.Union[bool, typing.Tuple[bool, str]]
                if check_validity_pre_include:
                    inc = validate_neuroml2(incl_loc, verbose_validate=False)

                if inc:
                    logger.debug(
                        "Loading included NeuroML2 file: {} (base: {}, resolved: {}, checking {})".format(
                            include.href,
                            base_path,
                            incl_loc,
                            check_validity_pre_include,
                        )
                    )
                    nml2_sub_doc = read_neuroml2_file(
                        incl_loc,
                        True,
                        verbose=verbose,
                        already_included=already_included,
                        check_validity_pre_include=check_validity_pre_include,
                    )
                    if incl_loc not in already_included:
                        already_included.append(incl_loc)

                    membs = inspect.getmembers(nml2_sub_doc)

                    for memb in membs:
                        if (
                            isinstance(memb[1], list)
                            and len(memb[1]) > 0
                            and not memb[0].endswith("_")
                        ):
                            for entry in memb[1]:
                                if memb[0] != "includes":
                                    logger.debug(
                                        "  Adding {!s} from: {!s} to list: {}".format(
                                            entry, incl_loc, memb[0]
                                        )
                                    )
                                    getattr(nml2_doc, memb[0]).append(entry)
                    incl_to_remove.append(include)
                else:
                    logger.warning("Not including file as it's not valid...")

        for include in incl_to_remove:
            nml2_doc.includes.remove(include)

    return nml2_doc


def write_neuroml2_file(
    nml2_doc: NeuroMLDocument,
    nml2_file_name: str,
    validate: bool = True,
    verbose_validate: bool = False,
    hdf5: bool = False,
) -> typing.Optional[typing.Union[bool, typing.Tuple[bool, str]]]:
    """Write a NeuroMLDocument object to a file using libNeuroML.

    :param nml2_doc: NeuroMLDocument object to write to file
    :type nml2_doc: NeuroMLDocument
    :param nml2_file_name: name of file to write to
    :type nml2_file_name: str
    :param validate: toggle whether the written file should be validated
    :type validate: bool
    :param verbose_validate: toggle whether the validation should be verbose
    :type verbose_validate: bool
    :param hdf5: write to HDF5 file
    :type hdf5: bool
    """
    if hdf5 is True:
        writers.NeuroMLHdf5Writer.write(nml2_doc, nml2_file_name)
    else:
        writers.NeuroMLWriter.write(nml2_doc, nml2_file_name)

    if validate:
        return validate_neuroml2(nml2_file_name, verbose_validate)
    return None


def read_lems_file(
    lems_file_name: str,
    include_includes: bool = False,
    fail_on_missing_includes: bool = False,
    debug: bool = False,
) -> lems_model.Model:
    """Read LEMS file using PyLEMS. See WARNING below.

    WARNING: this is a general function that uses PyLEMS to read any files that
    are valid LEMS *even if they are not valid NeuroML*. Therefore, this
    function is not aware of the standard NeuroML LEMS definitions.

    To validate NeuroML LEMS files which need to be aware of the NeuroML
    standard LEMS definitions, please use the `validate_neuroml2_lems_file`
    function instead.
    """
    if not os.path.isfile(lems_file_name):
        logger.critical("Unable to find file: %s!" % lems_file_name)
        sys.exit(FILE_NOT_FOUND_ERR)

    model = lems_model.Model(
        include_includes=include_includes,
        fail_on_missing_includes=fail_on_missing_includes,
    )
    model.debug = debug

    model.import_from_file(lems_file_name)

    return model


def write_lems_file(
    lems_model: lems_model.Model, lems_file_name: str, validate: bool = False
) -> None:
    """Write a lems_model.Model to file using pyLEMS.

    :param lems_model: LEMS model to write to file
    :type lems_model: lems_model.Model
    :param lems_file_name: name of file to write to
    :type lems_file_name: str
    :param validate: toggle whether written file should be validated
    :type validate: bool
    """
    lems_model.export_to_file(lems_file_name)

    if validate:
        from lems.base.util import validate_lems

        validate_lems(lems_file_name)


def confirm_file_exists(filename: str) -> None:
    """Check if a file exists, exit if it does not.

    :param filename: the filename to check
    :type filename: str
    """
    if not os.path.isfile(filename):
        logger.critical("Unable to find file: %s!" % filename)
        sys.exit(FILE_NOT_FOUND_ERR)


def confirm_neuroml_file(filename: str) -> None:
    """Confirm that file exists and is a NeuroML file before proceeding with
    processing.

    :param filename: Names of files to check
    :type filename: str
    """
    # print('Checking file: %s'%filename)
    # Some conditions to check if a LEMS file was entered
    # TODO: Ideally we'd like to check the root node: checking file extensions is brittle
    confirm_file_exists(filename)
    if filename.startswith("LEMS_"):
        logger.warning(
            textwrap.dedent(
                """
            *************************************************************************************
            **  Warning, you may be trying to use a LEMS XML file (containing <Simulation> etc.)
            **  for a pyNeuroML option when a NeuroML2 file is required...
            *************************************************************************************
            """
            )
        )


def confirm_lems_file(filename: str) -> None:
    """Confirm that file exists and is a LEMS file before proceeding with
    processing.

    :param filename: Names of files to check
    :type filename: list of strings
    """
    # print('Checking file: %s'%filename)
    # Some conditions to check if a LEMS file was entered
    # TODO: Ideally we'd like to check the root node: checking file extensions is brittle
    confirm_file_exists(filename)
    if filename.endswith("nml"):
        logger.warning(
            textwrap.dedent(
                """
            *************************************************************************************
            **  Warning, you may be trying to use a NeuroML2 file for a pyNeuroML option
            **  when a LEMS XML file (containing <Simulation> etc.) is required...
            *************************************************************************************
            """
            )
        )