[Deprecation] Move to new Crowdln API

[TarekkMA]

We should move to API 2.0 before the end of the year

[welcome[bot]]

Hello! 👋 Thanks for logging this issue. Please remember we are all volunteers here, so some patience may be required before we can get to the issue. Also remember that the fastest way to get resolution on an issue is to propose a change directly, https://github.com/ankidroid/Anki-Android/wiki/Contributing

[mikehardy]

That's unfortunate, already plenty to do but it is always something in a large project :-) Thanks for logging this.

[tapish2000]

Can I take this issue?

[david-allison]

Can I take this issue?

This may be complex - I feel it might be better if you take a different issue while you're getting used to the project

[github-actions[bot]]

Hello 👋, this issue has been opened for more than 2 months with no activity on it. If the issue is still here, please keep in mind that we need community support and help to fix it! Just comment something like still searching for solutions and if you found one, please open a pull request! You have 7 days until this gets closed automatically

[mikehardy]

They have a javascript client which has examples on the README detailing pretty much exactly what we need to do

https://support.crowdin.com/api/v2/#section/Introduction/Crowdin-API-Clients

https://github.com/crowdin/crowdin-api-client-js

They also have other languages available, but as a scripting language I'd prefer to use typescript these days. It has the widest developer pool so most people know it, it comes in handy when cross-training for our github actions scripts, and we use it as a front-end web technology as well in the Anki ecosystem.

I suggest making a sub-directory in whatever the location is of the current script, and that subdirectory is a typescript module with package.json and everything. Execute ts-node, alter the github action to do a yarn in that sub-directory to install packages then the same ts-node execute on the script

The hard part here will be testing. I suggest making a new temporary crowdin project on a temporary fork, then porting the current Python "file push to crowdin" to the new typescript equivalent and making sure that crowdin sees the files in the right spot so we have continuity with the real project. Then port the "build project and download project" python to the typescript equivalent, then run our post-processing scripts as they are and push to the i18n_sync branch like in current github action

Not sure I'll have time to do this, but I did the research at least and that's the sketch I'd follow.

Incremental development and testing testing testing each feature. using a second temp project from a temp fork means you don't have to worry about official access tokens or anything, you can create them yourself as temp project owner so there's no blocker and no chance of messing up the main crowdin translations

[david-allison]

Would probably need to check with CrowdIn support about getting the test project verified as open source, we're very close to the 60k strings limit which pushes you to paid territory if you're not an open source app

[mikehardy]

As long as it's done quickly, no worries ;-). test could also just skip-on-purpose a couple files as long as they were in the same path location (like, to 01 through 05, skip 06-10 or whatever). The test needs to be representative, not complete IMHO

[krmanik]

I will give it a try.

[mikehardy]

Champion :trophy: ! let me know if you need any advice or anything, this is so important to project continuity that I really want to see this get done and I will help any way I can

[krmanik]

Okay, Thanks

[krmanik]

@mikehardy @david-allison

I have updated this shell script to find broken strings. The typescript approach is slower and may not catch the expected errors.

This is added in below script.

'<!--'
'%d%%'
'%s%%'
'%1$d%%'
'%1$.1f'
 '%1$.2f'
 '%2$.2f'
 '%3$.1f'

Will this script handle the expected errors?

#!/bin/bash
# Spot malformed string replacement patterns in Android localization files.
# Hopefully it will prevent this kind of bugs: https://code.google.com/p/ankidroid/issues/detail?id=359

# shellcheck disable=SC2016
EXIT_STATUS=0

VALUES=res/values/10-preferences.xml

pushd AnkiDroid/src/main > /dev/null || exit 1
if grep -RHn "%1$ s" $VALUES; then
  echo "Found '%1$ s'-related error"
  EXIT_STATUS=$((EXIT_STATUS + 1));
fi
if grep -RHn "%1$ d" $VALUES; then
  echo "Found '%1$ s'-related error"
  EXIT_STATUS=$((EXIT_STATUS + 1));
fi
if grep -RHn "%1" $VALUES | grep -v "%1\\$"; then
  echo "Found '%1'-related error"
  EXIT_STATUS=$((EXIT_STATUS + 1));
fi

# This is currently not working and I'm not sure why? It worked a couple days ago as of 20200516
echo "The next test will likely only run correctly on macOS. On Ubuntu it does not work well."
if grep -RHn '%' $VALUES |
 grep -v '<!--'     | # Filter comments
 grep -v '%d%%'     |
 grep -v '%s%%'     |
 grep -v '%1$d%%'   |
 grep -v 'pref_summary_percentage' |
 sed -e 's/%/\n%/g' | # Split lines that contain several expressions
 grep '%'           | # Filter out lines that do not contain expressions
 grep -v ' n% '     | # Lone % character, not a variable
 grep -v '(n%)'     | # Lone % character, not a variable
 grep -v 'n%<'      | # Same, at the end of the string
 grep -v '>n% '     | # Same, at the beginning of the string
 grep -v '%で'      | # Same, no spaces in Japanese
 grep -v '%s'       | # Single string variable
 grep -v '%d'       | # Single decimal variable
 grep -v '%[0-9][0-9]\?$s' | # Multiple string variable
 grep -v '%[0-9][0-9]\?$d' |  # Multiple decimal variable
 grep -v '%1$.1f'   | # ?
 grep -v '%1$.2f'   |
 grep -v '%2$.2f'   |
 grep -v '%3$.1f'   |
 grep -v '%.0f'     |
 grep -v '%.1f'     |
 grep -v '%\\n'     |
 grep -v 'stats_overview_card_types_'
then
 echo "Found grep errors in $VALUES"
 EXIT_STATUS=$((EXIT_STATUS + 1))
fi

if grep -RHn '％' $VALUES; then
  echo "Found errors in simple '%' grep"
  EXIT_STATUS=$((EXIT_STATUS + 1));
  fi

if grep -RHn "CDATA " $VALUES; then
  echo "Found CDATA-related errors"
  EXIT_STATUS=$((EXIT_STATUS + 1));
fi

${ANDROID_HOME}/cmdline-tools/latest/bin/lint --check StringFormatInvalid ./res
popd > /dev/null || exit 1
echo "Exiting with status $EXIT_STATUS"
exit $EXIT_STATUS

[krmanik]

The lint catch the errors that I am trying to implement in check.ts file.

The following lint check errors in localized strings also and it is faster.

Check string format errors like %1$ s (space after $) and other errors in <string>

lint --check StringFormatInvalid ./res

Check error for %1 $s (space after %1)

lint --check StringFormatCount ./res

Check if localized string contains one, few, many, other for quantity strings

lint --check MissingQuantity ./res

In these locales, it is usually an error to have a message which does not include a formatting argument (such as '%d'), since it will not be clear from the grammar what quantity the quantity string is describing.

lint --check ImpliedQuantity ./res

I think the lint should be used. http://tools.android.com/tips/lint-checks