James Gardner: Home > Work > Code > ConversionKit > 0.3.2 > Chapter 6: Message Customization...

ConversionKit v0.3.2 documentation

Chapter 6: Message Customization and Internationalization

WARNING: None of this has been tested and most likely is incorrect

Customising Converter Error Messages

The ConversionKit converters all come with default messages to set on conversion objects if error conditions occur but very frequently you will want to customise these messages so that they make sense in the context in which they are being used.

Specifying Alternative Arguments

The error messages used by a converter are defined by the function arguments to the converter factory which produces the converter. All error message arguments start with the characters msg_.

As an example, the error message which is set if an email converter recieves a conversion whose value is empty is called msg_empty and by default has the value u'Please enter an email address'. If you want to customise this message you can do so by specifying an alternative when you call the stringToEmail() converter factory.

Here’s an example. First the imports:

>>> from conversionkit import Conversion
>>> from formconvert.email import stringToEmail

Then the standard version:

>>> email = stringToEmail()
>>> Conversion('').perform(email).error
u'Please enter an email address'

Here is a customised version:

>>> email = stringToEmail(
...     msg_empty = u'No value was entered'
... )
>>> Conversion('').perform(email).error
u'No value was entered'

Notice that the erorr now uses the customised message.

You can find out the error messages which are available to be customised by looking at the arguments to the converters in the API Documentation. Any arguments starting msg_ are customizable error messages.

Planning For Error Variables

Some error messages have certain values associated with them which you can use in the message using Python string interpolation of dictionary values.

For example the msg_socket error message for the stringToEmail() converter factory looks like this: u'An error occured when trying to connect to the server: %(error)s'. Here, the %(error)s text will be replaced with the actual error. Generally speaking, if the original error message contains variables it is wise to ensure your customised versions do too.

The state.conversionkit.message() Function

If you look at the source code for any of the ConversionKit converters you’ll see that the messages are all declared by being wrapped in a _() function. Tools such as Babel and GNU gettext can recognise this function and extract the message from the Python source code into a .po file for internationalization. A .pot file is created from the .po file and a translater translates the strings. The .po file is then compiled in a .mo file, known as a message catalog, and an internationalisation function can then pick the right translation to use for the particular user. The ConversionKit _() function itself doesn’t do anything except return the value passed to it.

Whenever ConversionKit generates an error message for a conversion it calls the conversionkit.message() function with three arguments:


The state argument passed to the conversion’s .perform() method when the converter was applied.


The error message as a string, whether it was the default or a customised message passed as a converter factory argument
A dictionary of any arguments which can be substituted into the message string.

If the code using ConversionKit supports internationalisation you can provide ConversionKit with a custom function which also gets passed these arguments and is responsible for returning a translated version of the error message and for substituting in any args.


At the moment ConversionKit only has messages translated into British English but if you would like to provide a variant in an alternative language the authors would be happy to accept a .pot file.

Here’s an example function which doesn’t return the messages in another language but does make all messages uppercase. The principle is the same:

>>> def message(state, message, args):
...     return message.upper() % args

To get ConversionKit to use your custom error message handler you need to set the function as part of the state object. Here’s an example which sets up the state with the message() function above and then tests it:

>>> from conversionkit import Conversion
>>> from bn import AttributeDict
>>> from formconvert.email import stringToEmail
>>> email = stringToEmail()
>>> state = AttributeDict()
>>> state['conversionkit'] = {'message': message}
>>> Conversion('').perform(email, state).error

You can imagine that a different function would be capable of translating the error message instead.

James Gardner: Home > Work > Code > ConversionKit > 0.3.2 > Chapter 6: Message Customization...