Checks and fixups

Custom automatic fixups

You can also implement your own automatic fixup in addition to the standard ones and include them in AUTOFIX_LIST.

The automatic fixes are powerful, but can also cause damage; be careful when writing one.

For example, the following automatic fixup would replace every occurrence of string foo in translation with bar:

# -*- coding: utf-8 -*-
#
# Copyright © 2012 - 2019 Michal Čihař <michal@cihar.com>
#
# This file is part of Weblate <https://weblate.org/>
#
# This program is free software: you can redistribute it and/or modify
# it under the terms of the GNU General Public License as published by
# the Free Software Foundation, either version 3 of the License, or
# (at your option) any later version.
#
# This program is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
# GNU General Public License for more details.
#
# You should have received a copy of the GNU General Public License
# along with this program.  If not, see <https://www.gnu.org/licenses/>.
#

from django.utils.translation import ugettext_lazy as _

from weblate.trans.autofixes.base import AutoFix


class ReplaceFooWithBar(AutoFix):
    """Replace foo with bar."""

    name = _("Foobar")

    def fix_single_target(self, target, source, unit):
        if "foo" in target:
            return target.replace("foo", "bar"), True
        return target, False

To install custom checks, you need to provide a fully-qualified path to the Python class in the AUTOFIX_LIST, see Custom quality checks and auto fixes.

Customizing behavior

You can fine tune Weblate behavior (mostly checks) for each source string (in source strings review, see Additional info on source strings) or in the Component configuration (Translation flags). Some file formats also allow to specify flags directly in the format.

Here is a list of flags currently accepted:

rst-text
Treat text as RST document, effects Unchanged translation.
md-text
Treat text as Markdown document.
dos-eol
Use DOS end of line markers instead of Unix ones (\r\n instead of \n).
url
The string should consist of URL only.
max-length:N
Limit maximal length for string to N chars, see Maximum Length
xml-text
Treat text as XML document, affects Invalid XML markup and XML tags mismatch.
font-family
Define font family for rendering checks, see Managing fonts.
font-weight
Define font weight for rendering checks, see Managing fonts.
font-size
Define font size for rendering checks, see Managing fonts.
font-spacing
Define font spacing for rendering checks, see Managing fonts.
python-format, c-format, php-format, python-brace-format, javascript-format, c-sharp-format, java-format, java-messageformat, auto-java-messageformat
Treats all strings like format strings, affects Formatted strings, Formatted strings, Formatted strings, Formatted strings, Formatted strings, Formatted strings, Formatted strings, Formatted strings, Unchanged translation.
ignore-end-space
Skip the «Trailing space» quality check.
ignore-inconsistent
Skip the «Inconsistent» quality check.
ignore-translated
Skip the «Has been translated» quality check.
ignore-begin-newline
Skip the «Starting newline» quality check.
ignore-zero-width-space
Skip the «Zero-width space» quality check.
ignore-escaped-newline
Skip the «Mismatched n» quality check.
ignore-same
Skip the «Unchanged translation» quality check.
ignore-end-question
Skip the «Trailing question» quality check.
ignore-end-ellipsis
Skip the «Trailing ellipsis» quality check.
ignore-ellipsis
Skip the «Ellipsis» quality check.
ignore-python-brace-format
Skip the «Python brace format» quality check.
ignore-end-newline
Skip the «Trailing newline» quality check.
ignore-c-format
Skip the «C format» quality check.
ignore-javascript-format
Skip the «JavaScript format» quality check.
ignore-optional-plural
Skip the «Optional plural» quality check.
ignore-end-exclamation
Skip the «Trailing exclamation» quality check.
ignore-end-colon
Skip the «Trailing colon» quality check.
ignore-xml-invalid
Skip the «Invalid XML markup» quality check.
ignore-xml-tags
Skip the «XML tags mismatch» quality check.
ignore-python-format
Skip the «Python format» quality check.
ignore-plurals
Skip the «Missing plurals» quality check.
ignore-begin-space
Skip the «Starting spaces» quality check.
ignore-bbcode
Skip the «Mismatched BBcode» quality check.
ignore-multiple-failures
Skip the «Multiple failing checks» quality check.
ignore-php-format
Skip the «PHP format» quality check.
ignore-end-stop
Skip the «Trailing stop» quality check.
ignore-angularjs-format
Skip the «AngularJS interpolation string» quality check.
ignore-c-sharp-format
Skip the «C# format» quality check.
ignore-java-format
Skip the «Java format» quality check.

Примечание

Generally the rule is named ignore-* for any check, using its identifier, so you can use this even for your custom checks.

These flags are understood both in Component configuration settings, per source string settings and in translation file itself (eg. in GNU Gettext).

Managing fonts

Добавлено в версии 3.7.

The Maximum size of translation check needs fonts to properly calculate dimensions of rendered text. The fonts can be managed in Weblate in the font management tool which you can find as Fonts under Tools menu of your translation project.

You can upload TrueType or OpenType fonts, configure font groups and use those in the the check.

The font groups allow you to define different fonts for different languages, what is typically needed for non latin languages:

../_images/font-group-edit.png

The font groups are identified by name, which can not contain whitespace or special chars to be easy to use in check definition:

../_images/font-group-list.png

After upload the font family and style is automatically recognized:

../_images/font-edit.png

You can have number of fonts loaded into Weblate:

../_images/font-list.png

Writing own checks

Weblate comes with wide range of quality checks (see Quality checks), though they might not 100% cover all you want to check. The list of performed checks can be adjusted using CHECK_LIST and you can also add custom checks. All you need to do is to subclass weblate.checks.Check, set few attributes and implement either check or check_single methods (first one if you want to deal with plurals in your code, the latter one does this for you). You will find below some examples.

To install custom checks, you need to provide a fully-qualified path to the Python class in the CHECK_LIST, see Custom quality checks and auto fixes.

Checking translation text does not contain «foo»

This is a pretty simple check which just checks whether translation does not contain string «foo».

# -*- coding: utf-8 -*-
#
# Copyright © 2012 - 2019 Michal Čihař <michal@cihar.com>
#
# This file is part of Weblate <https://weblate.org/>
#
# This program is free software: you can redistribute it and/or modify
# it under the terms of the GNU General Public License as published by
# the Free Software Foundation, either version 3 of the License, or
# (at your option) any later version.
#
# This program is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
# GNU General Public License for more details.
#
# You should have received a copy of the GNU General Public License
# along with this program.  If not, see <https://www.gnu.org/licenses/>.
#
"""Simple quality check example."""

from django.utils.translation import ugettext_lazy as _

from weblate.checks.base import TargetCheck


class FooCheck(TargetCheck):

    # Used as identifier for check, should be unique
    # Has to be shorter than 50 chars
    check_id = "foo"

    # Short name used to display failing check
    name = _("Foo check")

    # Description for failing check
    description = _("Your translation is foo")

    # Real check code
    def check_single(self, source, target, unit):
        return "foo" in target

Checking Czech translation text plurals differ

Check using language information to verify that two plural forms in Czech language are not same.

# -*- coding: utf-8 -*-
#
# Copyright © 2012 - 2019 Michal Čihař <michal@cihar.com>
#
# This file is part of Weblate <https://weblate.org/>
#
# This program is free software: you can redistribute it and/or modify
# it under the terms of the GNU General Public License as published by
# the Free Software Foundation, either version 3 of the License, or
# (at your option) any later version.
#
# This program is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
# GNU General Public License for more details.
#
# You should have received a copy of the GNU General Public License
# along with this program.  If not, see <https://www.gnu.org/licenses/>.
#
"""Quality check example for Czech plurals."""

from django.utils.translation import ugettext_lazy as _

from weblate.checks.base import TargetCheck


class PluralCzechCheck(TargetCheck):

    # Used as identifier for check, should be unique
    # Has to be shorter than 50 chars
    check_id = "foo"

    # Short name used to display failing check
    name = _("Foo check")

    # Description for failing check
    description = _("Your translation is foo")

    # Real check code
    def check_target_unit(self, sources, targets, unit):
        if self.is_language(unit, ("cs",)):
            return targets[1] == targets[2]
        return False

    def check_single(self, source, target, unit):
        """We don't check target strings here."""
        return False