Fix LT-22121: Analysis guesser should not guess ras for rAs

[jtmaxwell3]

A06-Kurdish uses A to represent a separate character from a. But the icu locale thinks that A is the capital of a. This means that the analysis guesser guesses analyses based on ras for rAs. One easy way to fix this would be to change the A to U+0391 GREEK CAPITAL LETTER ALPHA, but the data is from outside the project. An alternative is to only try lowercase analyses if the first letter of the wordform is uppercase. This will still guess ras for RAs, but getting it to guess rAs instead would require a lot more work since there are several places in FieldWorks where it is assumed that you use ToLower to lowercase words.

---

This change is 

[github-actions]

LCM Tests

    16 files  ±0      16 suites  ±0   3m 5s ⏱️ -2s
 2 831 tests +1   2 811 ✅ +1   20 💤 ±0  0 ❌ ±0 
11 272 runs  +4  11 104 ✅ +4  168 💤 ±0  0 ❌ ±0 

Results for commit 6610e6c. ± Comparison against base commit 69b6c0a.

♻️ This comment has been updated with latest results.

[imnasnainaec]

Would it be worth adding a test to https://github.com/sillsdev/liblcm/blob/LT-22121/tests/SIL.LCModel.Tests/DomainServices/AnalysisGuessServicesTests.cs?

[jasonleenaylor]

Reviewed 1 of 1 files at r1, 1 of 1 files at r2, all commit messages.
Reviewable status: complete! all files reviewed, all discussions resolved (waiting on @jtmaxwell3)