research.py

# -*- coding: utf-8 -*-
"""Functions for the research space."""

__copyright__ = 'Copyright (c) 2019-2023, Utrecht University'
__license__   = 'GPLv3, see LICENSE'

import genquery
from pathvalidate import validate_filename, validate_filepath, ValidationError

import folder
import groups
import meta_form
from util import *

__all__ = ['api_research_folder_add',
           'api_research_folder_copy',
           'api_research_folder_move',
           'api_research_folder_delete',
           'api_research_folder_rename',
           'api_research_file_copy',
           'api_research_file_rename',
           'api_research_file_move',
           'api_research_file_delete',
           'api_research_system_metadata',
           'api_research_collection_details',
           'api_research_list_temporary_files',
           'api_research_manifest']


def folder_new_name_check(folder_name):
    if len(folder_name) == 0:
        return False, api.Error('missing_foldername', 'Missing folder name. Please add a folder name')

    # TODO remove when upgrade to GenQuery 2
    # This check should only be done on new folders, since may have old folders with apostrophes
    if '\'' in folder_name:
        return False, api.Error('invalid_foldername', 'It is not allowed to use apostrophes in a folder name')

    # Name should not contain '\\' or '/'
    if '/' in folder_name or '\\' in folder_name:
        return False, api.Error('invalid_foldername', 'It is not allowed to use slashes in the new folder name')

    # name should not be '.' or '..'
    if folder_name in ('.', '..'):
        return False, api.Error('invalid_foldername', 'it is not allowed to name the folder {}'.format(folder_name))

    return True, ""


@api.make()
def api_research_folder_add(ctx, coll, new_folder_name):
    """Add a new folder to a research folder.

    :param ctx:             Combined type of a callback and rei struct
    :param coll:            Collection to create new folder in
    :param new_folder_name: Name of the new folder

    :returns: Dict with API status result
    """
    coll_target = coll + '/' + new_folder_name

    valid_folder_name, error_response = folder_new_name_check(new_folder_name)
    if not valid_folder_name:
        return error_response

    try:
        validate_filepath(coll_target.decode('utf-8'))
    except ValidationError:
        return api.Error('invalid_foldername', 'This is not a valid folder name. Please choose another name for your folder')

    # not in home - a groupname must be present ie at least 2!?
    if not len(coll.split('/')) > 2:
        return api.Error('invalid_destination', 'It is not possible to add folder ' + new_folder_name + ' at this location')

    # in vault?
    target_group_name = coll_target.split('/')[3]
    if target_group_name.startswith('vault-'):
        return api.Error('not_allowed', 'It is not possible to add folders in the vault')

    # permissions ok for group?
    user_full_name = user.full_name(ctx)
    if groups.user_role(ctx, user_full_name, target_group_name) in ['none', 'reader']:
        return api.Error('not_allowed', 'You do not have sufficient permissions to add new folders')

    # Collection exists?
    if not collection.exists(ctx, coll):
        return api.Error('invalid_foldername', 'The selected folder to add a new folder to does not exist')

    # Folder not locked?
    if folder.is_locked(ctx, coll):
        return api.Error('not_allowed', 'The indicated folder is locked so no new folders can be added to it')

    # new collection exists?
    if collection.exists(ctx, coll_target):
        return api.Error('invalid_foldername', 'The folder already exists. Please choose another name')

    # All requirements OK
    try:
        collection.create(ctx, coll_target)
    except msi.Error:
        return api.Error('internal', 'Something went wrong. Please try again')

    return api.Result.ok()


def folder_copy_check(ctx, folder_path, new_folder_path, overwrite, copy=True):
    """Check whether can copy (or move) folder to new folder location.

    :param ctx:             Combined type of a callback and rei struct
    :param folder_path:     Path to the folder to copy
    :param new_folder_path: Path to the new copy of the folder
    :param overwrite:       Overwrite folder if it already exists
    :param copy:            Whether a copy operation (True) or move (False) (just for logging purposes)

    :returns: 2-Tuple containing whether can copy/move, and the error if cannot
    """
    # Whether copy or move
    verb = 'copy' if copy else 'move'
    verb_past = 'copied' if copy else 'moved'
    if len(new_folder_path) == 0:
        return False, api.Error('missing_folder_path', 'Missing folder path. Please add a folder path')

    # TODO remove when upgrade to GenQuery 2
    if '\'' in new_folder_path:
        return False, api.Error('invalid_foldername', 'It is not allowed to use apostrophes in a folder name')

    try:
        validate_filepath(new_folder_path.decode('utf-8'))
    except ValidationError:
        return False, api.Error('invalid_foldername', 'This is not a valid folder name. Please choose another name for your folder')

    # Same folder path makes no sense.
    if folder_path == new_folder_path:
        return False, api.Error('invalid_folder_path', 'Origin and {} folder paths are equal. Please choose another destination'.format(verb))

    # Inside the same path makes no sense.
    if "{}/".format(folder_path) in new_folder_path:
        return False, api.Error('invalid_folder_path', 'Cannot {} folder inside itself. Please choose another destination'.format(verb))

    # not in home - a groupname must be present ie at least 2!?
    if not len(new_folder_path.split('/')) > 2:
        return False, api.Error('invalid_destination', 'It is not possible to {} folder at this location'.format(verb))

    # in vault?
    target_group_name = new_folder_path.split('/')[3]
    if target_group_name.startswith('vault-'):
        return False, api.Error('invalid_destination', 'It is not possible to {} folder to the vault'.format(verb))

    # permissions ok for group?
    user_full_name = user.full_name(ctx)
    if groups.user_role(ctx, user_full_name, target_group_name) in ['none', 'reader']:
        return False, api.Error('not_allowed', 'You do not have sufficient permissions to {} the selected folder'.format(verb))

    # Folder not locked?
    if folder.is_locked(ctx, new_folder_path):
        return False, api.Error('not_allowed', 'The indicated folder is locked and therefore the folder can not be {}'.format(verb_past))

    # Does original folder exist?
    if not collection.exists(ctx, folder_path):
        return False, api.Error('invalid_source', 'The original folder ' + folder_path + ' can not be found')

    # Collection exists in destination?
    if not overwrite and collection.exists(ctx, new_folder_path):
        return False, api.Error('invalid_destination', 'Folder with this name already exists in destination')

    return True, ""


@api.make()
def api_research_folder_copy(ctx, folder_path, new_folder_path, overwrite=False):
    """Copy a folder in a research folder.

    :param ctx:             Combined type of a callback and rei struct
    :param folder_path:     Path to the folder to copy
    :param new_folder_path: Path to the new copy of the folder
    :param overwrite:       Overwrite folder if it already exists

    :returns: Dict with API status result
    """
    valid, errorResponse = folder_copy_check(ctx, folder_path, new_folder_path, overwrite, True)
    if not valid:
        return errorResponse

    # All requirements OK
    try:
        collection.copy(ctx, folder_path, new_folder_path, force=overwrite)
    except msi.Error:
        return api.Error('internal', 'Something went wrong. Please try again')

    return api.Result.ok()


@api.make()
def api_research_folder_move(ctx, folder_path, new_folder_path, overwrite=False):
    """Move a folder in a research folder.

    :param ctx:             Combined type of a callback and rei struct
    :param folder_path:     Path to the folder to move
    :param new_folder_path: Path to the new folder
    :param overwrite:       Overwrite folder if it already exists

    :returns: Dict with API status result
    """
    valid, errorResponse = folder_copy_check(ctx, folder_path, new_folder_path, overwrite, False)
    if not valid:
        return errorResponse

    # All requirements OK
    try:
        collection.move(ctx, folder_path, new_folder_path, force=overwrite)
    except msi.Error:
        return api.Error('internal', 'Something went wrong. Please try again')

    return api.Result.ok()


@api.make()
def api_research_folder_rename(ctx, new_folder_name, coll, org_folder_name):
    """Rename an existing research folder.

    :param ctx:             Combined type of a callback and rei struct
    :param new_folder_name: New folder name
    :param coll:            Parent collection of folder
    :param org_folder_name: Current name of the folder

    :returns: Dict with API status result
    """
    coll_target = coll + '/' + new_folder_name

    valid_folder_name, error_response = folder_new_name_check(new_folder_name)
    if not valid_folder_name:
        return error_response

    try:
        validate_filepath(coll_target.decode('utf-8'))
    except ValidationError:
        return api.Error('invalid_foldername', 'This is not a valid folder name. Please choose another name for your folder')

    # Same name makes no sense
    if new_folder_name == org_folder_name:
        return api.Error('invalid_foldername', 'Origin and target folder names are equal. Please choose another name')

    # not in home - a groupname must be present ie at least 2!?
    if not len(coll.split('/')) > 2:
        return api.Error('invalid_destination', 'It is not possible to add folder ' + org_folder_name + ' at this location')

    # in vault?
    target_group_name = coll_target.split('/')[3]
    if target_group_name.startswith('vault-'):
        return api.Error('not_allowed', 'It is not possible to rename folders in the vault')

    # permissions ok for group?
    user_full_name = user.full_name(ctx)
    if groups.user_role(ctx, user_full_name, target_group_name) in ['none', 'reader']:
        return api.Error('not_allowed', 'You do not have sufficient permissions to rename the selected folder')

    # Collection exists?
    if not collection.exists(ctx, coll):
        return api.Error('invalid_foldername', 'The selected folder does not exist')

    # Folder not locked?
    if folder.is_locked(ctx, coll):
        return api.Error('not_allowed', 'The indicated folder is locked and therefore can not be renamed')

    # new collection exists?
    if collection.exists(ctx, coll_target):
        return api.Error('invalid_foldername', 'The folder already exists. Please choose another name')

    # All requirements OK
    try:
        collection.rename(ctx, coll + '/' + org_folder_name, coll_target)
    except msi.Error:
        return api.Error('internal', 'Something went wrong. Please try again')

    return api.Result.ok()


@api.make()
def api_research_folder_delete(ctx, coll, folder_name):
    """Delete a research folder.

    :param ctx:         Combined type of a callback and rei struct
    :param coll:        Parent collection of folder to delete
    :param folder_name: Name of folder to delete

    :returns: Dict with API status result
    """
    coll_target = coll + '/' + folder_name

    # Not in home - a groupname must be present ie at least 2!?
    if not len(coll.split('/')) > 2:
        return api.Error('invalid_target', 'It is not possible to delete folder ' + folder_name + ' at this location')

    # Name should not contain '\\' or '/'.
    if '/' in folder_name or '\\' in folder_name:
        return api.Error('invalid_foldername', 'It is not allowed to use slashes in folder name that will be deleted')

    # in vault?
    target_group_name = coll_target.split('/')[3]
    if target_group_name.startswith('vault-'):
        return api.Error('not_allowed', 'It is not possible to delete folders from the vault')

    # permissions ok for group?
    user_full_name = user.full_name(ctx)
    if groups.user_role(ctx, user_full_name, target_group_name) in ['none', 'reader']:
        return api.Error('not_allowed', 'You do not have sufficient permissions to delete the selected folder')

    # Folder not locked?
    if folder.is_locked(ctx, coll):
        return api.Error('not_allowed', 'The indicated folder is locked and therefore can not be deleted')

    # Collection exists?
    if not collection.exists(ctx, coll_target):
        return api.Error('invalid_target', 'The folder to delete does not exist')

    # All requirements OK
    try:
        collection.remove(ctx, coll_target)
    except msi.Error:
        return api.Error('internal', 'Something went wrong. Please try again')

    return api.Result.ok()


@api.make()
def api_research_list_temporary_files(ctx, coll):
    """Get list of temporary files to be cleaned up.

    :param ctx:  Combined type of a callback and rei struct
    :param coll: Parent collection of folder to delete

    :returns: List of files that possibly require cleaning up
    """
    list_cleanup_files = []

    for uw_file in config.temporary_files:
        if "?" in uw_file or "*" in uw_file:
            wildcard_file = uw_file.replace('%', '\\%').replace('_', '\\_').replace('?', '_').replace('*', '%')
            iter = genquery.row_iterator(
                "DATA_NAME, COLL_NAME",
                "COLL_NAME like '" + coll + "%' AND DATA_NAME LIKE '" + wildcard_file + "'",
                genquery.AS_LIST, ctx
            )
        else:
            iter = genquery.row_iterator(
                "DATA_NAME, COLL_NAME",
                "COLL_NAME like '" + coll + "%' AND DATA_NAME = '" + uw_file + "'",
                genquery.AS_LIST, ctx
            )

        for row in iter:
            list_cleanup_files.append([row[1], row[0]])

    return list_cleanup_files


@api.make()
def api_research_file_copy(ctx, filepath, new_filepath, overwrite=False):
    """Copy a file in a research folder.

    :param ctx:          Combined type of a callback and rei struct
    :param filepath:     Path to the file to copy
    :param new_filepath: Path to the new copy of the file
    :param overwrite:    Overwrite file if it already exists

    :returns: Dict with API status result
    """
    if len(new_filepath) == 0:
        return api.Error('missing_filepath', 'Missing file path. Please add a file path')

    # Same filepath makes no sense.
    if filepath == new_filepath:
        return api.Error('invalid_filepath', 'Origin and copy file paths are equal. Please choose another destination')

    _, org_data_name = pathutil.chop(filepath)
    # These are of the NEW filepath
    coll, data_name = pathutil.chop(new_filepath)
    try:
        validate_filename(data_name.decode('utf-8'))
    except Exception:
        return api.Error('invalid_filename', 'This is not a valid file name. Please choose another name')

    # TODO remove when upgrade to GenQuery 2
    # This check should only be done on new folders, since may have old folders with apostrophes
    if '\'' in coll:
        return api.Error('invalid_filepath', 'It is not allowed to copy a file to a folder with an apostrophe in the name')

    # not in home - a groupname must be present ie at least 2!?
    if not len(coll.split('/')) > 2:
        return api.Error('invalid_destination', 'It is not possible to copy files at this location')

    # Name should not contain '\\' or '/'
    if '/' in data_name or '\\' in data_name:
        return api.Error('invalid_filename', 'It is not allowed to use slashes in the new name of a file')

    # in vault?
    target_group_name = new_filepath.split('/')[3]
    if target_group_name.startswith('vault-'):
        return api.Error('invalid_destination', 'It is not possible to copy files in the vault')

    # permissions ok for group?
    user_full_name = user.full_name(ctx)
    if groups.user_role(ctx, user_full_name, target_group_name) in ['none', 'reader']:
        return api.Error('not_allowed', 'You do not have sufficient permissions to copy the selected file')

    # Folder not locked?
    if folder.is_locked(ctx, coll):
        return api.Error('not_allowed', 'The indicated folder is locked and therefore the indicated file can not be copied')

    # Does org file exist?
    if not data_object.exists(ctx, filepath):
        return api.Error('invalid_source', 'The original file ' + org_data_name + ' can not be found')

    # new filename already exists?
    if not overwrite and data_object.exists(ctx, new_filepath):
        return api.Error('invalid_destination', 'The file ' + data_name + ' already exists')

    # All requirements OK
    try:
        data_object.copy(ctx, filepath, new_filepath, force=overwrite)
    except msi.Error:
        return api.Error('internal', 'Something went wrong. Please try again')

    return api.Result.ok()


@api.make()
def api_research_file_rename(ctx, new_file_name, coll, org_file_name):
    """Rename a file in a research folder.

    :param ctx:           Combined type of a callback and rei struct
    :param new_file_name: New file name
    :param coll:          Parent collection of file
    :param org_file_name: Current name of the file

    :returns: Dict with API status result
    """
    if len(new_file_name) == 0:
        return api.Error('missing_filename', 'Missing filename. Please add a file name')

    try:
        validate_filename(new_file_name.decode('utf-8'))
    except Exception:
        return api.Error('invalid_filename', 'This is not a valid file name. Please choose another name')

    # Same name makes no sense
    if new_file_name == org_file_name:
        return api.Error('invalid_filename', 'Origin and target file names are equal. Please choose another name')

    path_target = coll + '/' + new_file_name

    # not in home - a groupname must be present ie at least 2!?
    if not len(coll.split('/')) > 2:
        return api.Error('invalid_destination', 'It is not possible to rename files at this location')

    # Name should not contain '\\' or '/'
    if '/' in new_file_name or '\\' in new_file_name:
        return api.Error('invalid_filename', 'It is not allowed to use slashes in the new name of a file')

    # in vault?
    target_group_name = path_target.split('/')[3]
    if target_group_name.startswith('vault-'):
        return api.Error('invalid_destination', 'It is not possible to rename files in the vault')

    # permissions ok for group?
    user_full_name = user.full_name(ctx)
    if groups.user_role(ctx, user_full_name, target_group_name) in ['none', 'reader']:
        return api.Error('not_allowed', 'You do not have sufficient permissions to rename the selected file')

    # Folder not locked?
    if folder.is_locked(ctx, coll):
        return api.Error('not_allowed', 'The indicated folder is locked and therefore the indicated file can not be renamed')

    # Does org file exist?
    if not data_object.exists(ctx, coll + '/' + org_file_name):
        return api.Error('invalid_source', 'The original file ' + org_file_name + ' can not be found')

    # new filename already exists?
    if data_object.exists(ctx, path_target):
        return api.Error('invalid_destination', 'The selected filename ' + new_file_name + ' already exists')

    # Does data object have replica in intermediate or locked state? (replica is not actively being written to or opened for read)
    if data_object.has_replica_with_status(ctx, coll + '/' + org_file_name, [constants.replica_status.INTERMEDIATE_REPLICA,
                                                                             constants.replica_status.READ_LOCKED,
                                                                             constants.replica_status.WRITE_LOCKED]):
        return api.Error('not_allowed', 'Not allowed, this file is actively being written to')

    # All requirements OK
    try:
        data_object.rename(ctx, coll + '/' + org_file_name, path_target)
    except msi.Error:
        return api.Error('internal', 'Something went wrong. Please try again')

    return api.Result.ok()


@api.make()
def api_research_file_move(ctx, filepath, new_filepath, overwrite=False):
    """Move a file in a research folder.

    :param ctx:          Combined type of a callback and rei struct
    :param filepath:     Path to the file to move
    :param new_filepath: Path to the new location of the file
    :param overwrite:    Overwrite file if it already exists

    :returns: Dict with API status result
    """
    if len(new_filepath) == 0:
        return api.Error('missing_filepath', 'Missing file path. Please add a file path')

    # Same filepath makes no sense.
    if filepath == new_filepath:
        return api.Error('invalid_filepath', 'Origin and move file paths are equal. Please choose another destination')

    # These are of the NEW filepath
    coll, data_name = pathutil.chop(new_filepath)
    try:
        validate_filename(data_name.decode('utf-8'))
    except Exception:
        return api.Error('invalid_filename', 'This is not a valid file name. Please choose another name')

    # TODO remove when upgrade to GenQuery 2
    # This check should only be done on new folders, since may have old folders with apostrophes
    if '\'' in coll:
        return api.Error('invalid_filepath', 'It is not allowed to move a file to a folder with an apostrophe in the name')

    # not in home - a groupname must be present ie at least 2!?
    if not len(coll.split('/')) > 2:
        return api.Error('invalid_destination', 'It is not possible to move files to this location')

    # Name should not contain '\\' or '/'
    if '/' in data_name or '\\' in data_name:
        return api.Error('invalid_filename', 'It is not allowed to use slashes in the new name of a file')

    # in vault?
    target_group_name = new_filepath.split('/')[3]
    if target_group_name.startswith('vault-'):
        return api.Error('invalid_destination', 'It is not possible to move files in the vault')

    # permissions ok for group?
    user_full_name = user.full_name(ctx)
    if groups.user_role(ctx, user_full_name, target_group_name) in ['none', 'reader']:
        return api.Error('not_allowed', 'You do not have sufficient permissions to move the selected file')

    # Folder not locked?
    if folder.is_locked(ctx, coll):
        return api.Error('not_allowed', 'The indicated folder is locked and therefore the indicated file can not be moved')

    # Does org file exist?
    if not data_object.exists(ctx, filepath):
        return api.Error('invalid_source', 'The original file ' + data_name + ' can not be found')

    # new filename already exists?
    if not overwrite and data_object.exists(ctx, new_filepath):
        return api.Error('invalid_destination', 'The file ' + data_name + ' already exists')

    # All requirements OK
    try:
        if overwrite:
            data_object.copy(ctx, filepath, new_filepath, force=overwrite)
            data_object.remove(ctx, filepath)
        else:
            data_object.rename(ctx, filepath, new_filepath)
    except msi.Error:
        return api.Error('internal', 'Something went wrong. Please try again')

    return api.Result.ok()


@api.make()
def api_research_file_delete(ctx, coll, file_name):
    """Delete a file in a research folder.

    :param ctx:       Combined type of a callback and rei struct
    :param coll:      Parent collection of file to delete
    :param file_name: Name of file to delete

    :returns: Dict with API status result
    """
    path_target = coll + '/' + file_name

    # not in home - a groupname must be present ie at least 2!?
    if not len(coll.split('/')) > 2:
        return api.Error('invalid_target', 'It is not possible to delete files at this location')

    # in vault?
    target_group_name = path_target.split('/')[3]
    if target_group_name.startswith('vault-'):
        return api.Error('not_allowed', 'It is not possible to delete files from the vault')

    # permissions ok for group?
    user_full_name = user.full_name(ctx)
    if groups.user_role(ctx, user_full_name, target_group_name) in ['none', 'reader']:
        return api.Error('not_allowed', 'You do not have sufficient permissions to delete the selected file')

    # Folder not locked?
    if folder.is_locked(ctx, coll):
        return api.Error('not_allowed', 'The indicated folder is locked and therefore the indicated file can not be deleted')

    # Data object exists?
    if not data_object.exists(ctx, path_target):
        return api.Error('invalid_target', 'The data object to delete does not exist')

    # All requirements OK
    try:
        data_object.remove(ctx, path_target)
    except msi.Error:
        return api.Error('internal', 'Something went wrong. Please try again')

    return api.Result.ok()


@api.make()
def api_research_system_metadata(ctx, coll):
    """Return collection statistics as JSON.

    :param ctx:  Combined type of a callback and rei struct
    :param coll: Research collection

    :returns: Dict with research system metadata
    """
    data_count = collection.data_count(ctx, coll)
    collection_count = collection.collection_count(ctx, coll)
    size = collection.size(ctx, coll)
    size_readable = misc.human_readable_size(size)

    result = "{} files, {} folders, total of {}".format(data_count, collection_count, size_readable)

    return {"Package size": result}


@api.make()
def api_research_collection_details(ctx, path):
    """Return details of a research collection."""
    if not collection.exists(ctx, path):
        return api.Error('nonexistent', 'The given path does not exist')

    # Check if collection is a research group.
    space, _, group, _ = pathutil.info(path)
    if space != pathutil.Space.RESEARCH:
        return {}

    basename = pathutil.chop(path)[1]

    # Retrieve user type.
    member_type = groups.user_role(ctx, user.full_name(ctx), group)

    # Retrieve research folder status.
    status = folder.get_status(ctx, path)

    # Check if user is datamanager.
    category = groups.group_category(ctx, group)
    is_datamanager = groups.user_is_datamanager(ctx, category, user.full_name(ctx))

    # Retrieve lock count.
    lock_count = meta_form.get_coll_lock_count(ctx, path)
    is_locked = folder.is_locked(ctx, path)

    return {"basename": basename,
            "status": status.value,
            "member_type": member_type,
            "is_datamanager": is_datamanager,
            "lock_count": lock_count,
            "is_locked": is_locked}


@api.make()
def api_research_manifest(ctx, coll):
    """Produce a manifest of data objects in a collection

    :param ctx:  Combined type of a callback and rei struct
    :param coll: Parent collection of data objects to include

    :returns: List of json objects with name and checksum
    """
    iter = genquery.row_iterator(
        "ORDER(DATA_NAME), DATA_SIZE, DATA_CHECKSUM",
        "COLL_NAME = '{}'".format(coll),
        genquery.AS_LIST, ctx
    )
    checksums = [{"name": row[0], "size": misc.human_readable_size(int(row[1])), "checksum": data_object.decode_checksum(row[2])} for row in iter]

    iter_sub = genquery.row_iterator(
        "ORDER(COLL_NAME), ORDER(DATA_NAME), DATA_SIZE, DATA_CHECKSUM",
        "COLL_PARENT_NAME like '{}%'".format(coll),
        genquery.AS_LIST, ctx
    )
    length = len(coll) + 1
    checksums_sub = [{"name": (row[0] + "/")[length:] + row[1], "size": misc.human_readable_size(int(row[2])), "checksum": data_object.decode_checksum(row[3])} for row in iter_sub]

    return checksums + checksums_sub