Fix #5022, #5030, and #5031: Fix localization issues in Beta 0.11

[BenHenning]

Explanation

Fixes oppia/lessons-team#718
Fixes oppia/lessons-team#719
Fixes #5031

First, it was found during investigation that some images aren't showing up in Arabic or Brazilian Portuguese translations of lessons due to their image control character (that indicates to Android an image needs to be rendered) being replaced by empty reading text. This has been addressed with a more robust content description check and three new tests to detect this case and addresses oppia/lessons-team#718 and oppia/lessons-team#719. This was determined due to an excellent investigation & finding by @adhiamboperes.

Second, some specific images wouldn't render in Arabic specifically. This was due to a bug in AndroidSVG (our dependency library used for rendering SVG files). Per https://github.com/oppia/androidsvg/blob/4bc1d26412f0fb9fd4ef263fa93f6a64f4d4dbcf/androidsvg/src/main/java/com/caverock/androidsvg/utils/CSSFontVariationSettings.java#L73 AndroidSVG uses DecimalFormat to format certain CSS font settings (particularly in the case for dynamic fonts), and AndroidSVG isn't following the conventional advice of https://developer.android.com/reference/java/util/Locale (under "Be wary of the default locale") and https://docs.oracle.com/javase/8/docs/api/java/text/DecimalFormat.html (specific advice to create formatters for specific locales). In summary, AndroidSVG was formatting certain CSS parameters to floats without accounting for the fact that the default locale may change the numeral system when formatting those decimals. A patch to Oppia's fork of AndroidSVG fixes this problem (oppia/androidsvg@1265eb1). Note that this fix will also be included in the later AndroidSVG changes that the app will later be moved to. Additionally, these are unfortunately not changes that can be easily tested without finding specific SVGs to include in tests, and either scanning logcat or relying on screenshot testing. A more suitable testing strategy would be end-to-end tests vis-a-vis #3839.

Third, bullets weren't rendering correctly in Arabic views of lessons (#5031). This was due to two issues:

1. The previous logic was always assuming a rightward shift to center the bullet which doesn't work in RTL.
2. The previous logic was using the canvas width to invert X coordinates rather than the clipped space. This wasn't caught previously because the main tested content for bullets was a test revision card whose content spans the width of the screen (which corresponds to the drawing area of the canvas). However, lessons use a more constrained area due to content margins, so the bullets were being placed off screen in the margin whitespace that was being clipped.

See the following screenshots for a demonstration of the new content description tests failing without the fix, and passing with it:

Without fix	With fix

A few caveats with the bullet fixes:

• The FAQ bullets will still not render correctly in Arabic (technically in all non-English languages) due to [BUG]: Bullet lists aren't showing in FAQs for all non-English languages #5034 which requires fixes in the translation strings. They're actually missing the necessary markup elements needed to show their bullet lists correctly.
• The FAQ style fix can only be tested via basic measurement verification, but fortunately the existing test did require an update and changed in the expected way (the calculated x position of the bullets in the test have moved left 18 pixels). This isn't nearly as good as a visual test, but that will require Intoduce screenshot-testing framework #1815 to be finished, first.

Note that this PR will need to be cherry-picked into the 0.11 release branch.

Essential Checklist

• The PR title and explanation each start with "Fix #bugnum: " (If this PR fixes part of an issue, prefix the title with "Fix part of #bugnum: ...".)
• Any changes to scripts/assets files have their rationale included in the PR explanation.
• The PR follows the style guide.
• The PR does not contain any unnecessary code changes from Android Studio (reference).
• The PR is made from a branch that's not called "develop" and is up-to-date with "develop".
• The PR is assigned to the appropriate reviewers (reference).

For UI-specific PRs only

Demonstrations of fixes for oppia/lessons-team#718 (note that, for the first Arabic screenshots, the lesson image isn't actually localized in the content; also, the Naija image does load without the PR fixes but it can be a bit delayed):

Before (without fixes)	After (with the fixes of this PR)

Demonstrations of fixes for oppia/lessons-team#719 (only some images actually hit this bug):

Before (without fixes)	After (with the fixes of this PR)

Demonstrations of fixes for oppia/lessons-team#718, oppia/lessons-team#719, and #5031 (since Arabic lessons have conflated failures for images, and some of these images also demonstrate the bullet list issue):

Before (without fixes)	After (with the fixes of this PR)

[BenHenning]

PTAL @adhiamboperes. I've self-reviewed, and an earlier commit before the AndroidSVG version update passed all CI checks. I verified locally that Gradle is building (since I forgot to push the gradle change during the previous commit's almost-finished run), so I'm fairly confident CI will pass.

[adhiamboperes]

Everything looks good to me. I also tested all the changes locally and the linked issues are already fixed. Thanks @BenHenning!

[oppiabot]

Unassigning @adhiamboperes since they have already approved the PR.

[oppiabot]

Hi @BenHenning, this PR is ready to be merged. Please address any remaining comments prior to merging, and feel free to merge this PR once the CI checks pass and you're happy with it. Thanks!

[adhiamboperes]

Re: Caveat(1) on bullet fixes in the PR description: This has now been fixed and merged in.