Table of Contents
Translations missing key elements such as a variable or link could potentially break your website or app. To help prevent this, Transifex automatically checks every translation against a set of conditions.
Whenever a condition is not met, e.g. an HTML element is missing, the translator will be shown either a warning or error message telling them what needs to be fixed. Translators can still save a translation when there's a warning. But if there's an error, they'll need to address the issue before they'll be able to save the translation.
Configuring translation checks
Each file format has a set of translation checks by default. You can change whether a failed check results in an error or warning, or turn off a check altogether. You need to have an administrator role in order to perform the following changes.
To customize translation checks:
Go to your Organization Settings from the main navigation
Click on Translation Checks in the left menu
Click on a file format to see the checks available for that file format
Click on the toggles next to each rule to change whether a check should show a warning, error message, or be off.
Hit Save to confirm the changes
You can always restore the translation check settings by clicking the Set to default checks next to the Save button.
Note The file format for Transifex Live projects is "Transifex Format (.json) checks". |
Checking for custom variables
If you use custom variables in your code, you can set up custom translation checks and make sure they are left untouched in the translation. This is done by adding a custom check for a file format, and specifying either the characters your custom variables start and end with or the exact characters your variables consist of.
To do this:
Head to the Translation Checks settings page
Click on the file format you want to add custom checks for
Click Set up custom placeholders
In the popup:
if you provide a start and end value, then the system will try to detect text which starts and ends with the specified words. So, in order to protect variables like {address}, %email%, [username] etc. during the translation process, you will just need to complete the fields "Starts with" and "Ends with" as shown below:
if you want the system to look for exact matches instead (e.g. %ld, %temperature), then you can enter the variable in the "Enter a custom variable" field including all the characters the variable consists of
Click Add placeholder so as to finish adding this check. You can add more if you like by creating more placeholders.
Note When adding multiple placeholder delimiters, keep in mind that the delimiters are executed in the order they are added.
In the following example, both placeholders will be executed. However, if the order was reversed then only the first expression would be executed. |
Like the default checks, you can change whether a failed check results in an error or warning, or turn it off all together.
Translation checks based on regular expressions
With custom translation checks, you, as an organization administrator, can add your own regular expressions and ensure that nothing will break your website or application due to an invalid translation.
These checks can be created per file format as follows:
Click on the file format where you want to create your custom translation check
Click 'Add a translation check' at the bottom of the checks list
Provide a description for the check you want to create - This message will be displayed in the editor every time your rule is violated
Add your custom regular expression
Define where the custom translation check will be applied i.e. Translations only (the translation check will fail if there is a match) or Translations, and source strings (the translation check will fail if the matches found in the source string are different than the matches found in the translation). In case of the latter, you also have the option to enable the setting "Respect the matching order" which ensures that the order in source is preserved in the translation text
Test the regular expression you created before saving your configuration
Define the status of the check (Error, Warning, Off)
Note
|
Use Cases - Examples
Description: Match Quotes
When the source text contains double or single quotes and you want to check that there are quotes also in the translation text, you can use the following regular expression:
Regex: \"|\'
You can omit the double or the single quote from this regex, if you want to check only for one of them, eg. \"
This expression instructs the system to find if double or single quotes exist in the translation text. Depending on where this check will be applied, translations only or both source and translation string, the following can happen:
if the check is on translations only and the translation text does not contain double or single quotes, then no error will be returned.
if the check is on translations only and the translation text contains double or single quotes, an error is returned
if the check is on the source and translation strings and the source text contains double or single quotes, whereas the translation text does not, an error will be returned
if the check is on the source and translation strings and both the source and translation text contain double or single quotes, no error is returned
The setup will look like this:
Your regex expression can easily be tested using the test area.
Message returned when translation check fails:
Message returned when translation check is successful:
More examples can be found below:
Description: Special characters (excluding whitespaces)
You can use the following regular expression if you want to check whether there are characters apart from alphanumeric ones and whitespace in the translation text or in both the source and translation texts. Examples of special characters are: !@#$%&* and others.
Regex: [^a-zA-Z0-9\s]
-----------------------------------------------------------------------------------------------------------------------------
Description: Spaces, tabs, or newline at end of a string
Often translators add a space or a tab or a newline at the end of their translation text, which is not present in the source text. To check whether there are such cases, you can use the following regular expression.
Regex: [\s\t\n]+$
Using the + in the regular expression means that there will be a match even if spaces or tabs or newlines are more than one. The $ sign denotes the end of the string. If you apply it to both the source and the translation text, then the following can happen:
If both the source and the translation text contain a space at the end, no error will be returned
If either the source or the translation text contains a space at the end, an error will be returned
If the translation check is applied only to the translation text, an error will be raised every time there is a space, a tab, or a newline at the end of the string, regardless if the source text contains them or not.
-----------------------------------------------------------------------------------------------------------------------------
Description: Find whitespace at the end or beginning of a string
Similar to the example above, if you want to locate whitespace at the beginning or at the end of a string, you can use the following regular expression.
Regex: ^\s+|\s$
The | sign is used as the logical OR and the ^ sign marks the beginning of the string. If you apply this check to the translation text, for example, the following can happen:
If the translation text contains one or more whitespaces at the beginning of a string, or whitespace at the end, an error will be returned
If the translation text doesn't start or end with whitespace, then no error will be returned
-----------------------------------------------------------------------------------------------------------------------------
Description: Bold text wrapped with **double asterisks**
Often source strings contain inline formatting. This formatting usually needs to be applied to the translation text too. If you want to ensure that your translators preserve the formatting of words emphasized in the source language when translating content, then you can use the following regular expression.
Regex: \*\*
When this check is applied to the translation and source strings, the following can happen:
If part of your source string is wrapped with **double asterisks** and this formatting is not detected in the translation text applied, then an error message will be returned upon saving the translation.
If part of your translation string is wrapped with **double asterisks** and such formatting is not found in the source string, then an error message will be returned upon saving the translation.
-----------------------------------------------------------------------------------------------------------------------------
Description: Find multiple whitespaces
Sometimes the translators add more than one whitespaces in their translations by mistake. To catch such cases you can use the following regular expression.
Regex: \s{2,}
The curly brackets signify how many times the previous token, in our case the whitespace, is repeated. The lower limit is 2 times; there is no upper limit. When this check is applied to the translation text, the following can happen:
If the translation text contains two or more consecutive whitespaces somewhere in the string (beginning, middle, end) an error will be returned
If the translation text doesn't contain two or more consecutive whitespaces, then no error will be returned
-----------------------------------------------------------------------------------------------------------------------------
Description: Placeholders that start and end with % (no whitespace in between)
Many times the source text contains variables, which the translator should leave intact while he translates the source string. If whitespaces are added by mistake, then the developers' code will break and the app or page etc. will not load correctly.
Let's take, for example, the following source string: "There are %no_of_countries% in %continent%". The translator adds the following as a greek translation: "Υπάρχουν %no_of countries% χώρες στην %continent%". There is whitespace added by accident in the first variable. When the following regular expression is used, these mistakes can be caught and fixed, before the translations reach production.
Regex: %(?!%)(?!%)\S+?%
The negative lookaheads (?!%) denote that the first character, the %, should not be followed directly by one or two same characters. It should be followed by one or more non-whitespace characters, and end with the % character.
Given the example above, and considering that you have a translation check with the above regular expression configured for both the source and the translation text, an error (or a warning) will be raised for the translation entered.
-----------------------------------------------------------------------------------------------------------------------------
Description: Match everything except an unwanted pattern
When you want to identify instances where any character except a list of excluded characters is enclosed in brackets, then you can use the following regular expression.
Regex: \[([^w|W])\]
To match any character except a list of excluded characters put the excluded characters between [^ and ]. The caret ^ must immediately follow the [ .
When this check is applied to the translation text only, the following can happen:
If the translation text contains a character enclosed in brackets and this character is either
worW, then no errors will be returnedIf the translation text contains a character enclosed in brackets and this character is neither
wnorW, then an error will be returned
-----------------------------------------------------------------------------------------------------------------------------
Description: Consecutive duplicate words
You can use the following regular expression to identify instances where translation text contains consecutive duplicate words.
Regex: \b(\w+)\s+\1\b
The \b character is a word boundary that allows you to perform a “whole words only” search. For instance, regexp \bJava\b will be found in Hello, Java!, where Java is a standalone word, but not in Hello, JavaScript!.
The \w+ character matches word characters. A word character is a character a-z, A-Z, 0-9, including _ (underscore). The parentheses around this group of characters create a capturing group, which allows you to refer to the matched characters later in the expression.
The \s+ character matches one or more whitespace characters (such as a space, tab, or newline).
Finally, the \1 at the end of the expression matches the exact same text matched by the first capturing group.
When this check is applied to the translation text only, the following can happen:
If the translation text contains consecutive duplicate words, then an error will be returned.
If the translation text contains the same word twice, but one does not follow after the other in order, then no errors will be returned.
-----------------------------------------------------------------------------------------------------------------------------
Description: Mismatched colon
When you want to check if both source and translation text end with a colon(:), then you can use the following regular expression.
Regex: ^(?!.*:$).*$
When this check is applied to translations and source strings, the following can happen:
If the source string contains a colon (:) character at the end of the text but the translation string doesn't, then an error will be returned.
If the source AND translation strings contain a colon (:) character at the end of the text, then no errors will be returned.
-----------------------------------------------------------------------------------------------------------------------------
Description: Preserve style and tone in translations
To check if the translation accurately reflects the style and tone of the original text, you can use a regular expression to identify words or phrases that shouldn't be there in the first place. For example:
Regex: \b(yo|dude|bro)\b
When this check is applied to translations only, the following can happen:
If the translation text contains at least one of the words included in the expression above (i.e., yo, dude, bro), then an error will be returned.
If the translation text doesn't contain any of the words included in the expression above (i.e., yo, dude, bro), then no errors will be returned.
On-demand translation checks
On-demand translation checks can be found under the settings page of your Transifex project. This feature allows you to run translation checks on translation strings that have already been submitted.
In order to run translation checks on demand, you need to:
Visit the Settings page of your project in Transifex
Click on the Workflow tab
Scroll down to the "Quality Assurance" section
Click on the highlighted text "Check now" as shown below:
After that, the system will start running translation checks on the existing translations.
When the task is completed, an email notification will be sent to the user who performed this action (admins, maintainers) where a detailed report will be included:
