crowdin-edited context for python-based strings will be lost between releases

[indirectlylit]

Observed behavior

In https://github.com/learningequality/kolibri/pull/6867 @jredrejo added the ability to include context for Python-based strings. This means that context can be added by developers and uploaded to Crowdin, which is great.

However, often translators and non-developer LE staff (like @radinamatic) will edit and create context in Crowdin. For front-end strings, we have a mechanism (created by @nucleogenesis) of transferring crowdin-edited context back to the front-end $trs objects in source.

For backend strings, we'll need to check for and manually copy over any crowdin-edited context, or this information will be lost the next time we upload strings for a new release.

Expected behavior

context for backend strings should be automatically transferred from crowdin to source like the front-end strings are.

As a concrete example, context for this string was added via crowdin during the 0.14 translation process:

https://crowdin.com/translate/kolibri/3806/en-fa#279222

When running i18n-update, this context should be incorporated into the Python source files

User-facing consequences

high potential for lost context between releases

Steps to reproduce

• upload strings to crowdin in one branch
• edit context for a backend string
• download strings to Kolibri
• re-upload strings to crowdin in a new branch

Context

0.13.2