LT-21899: Parse Prioritizer look at following word as well as preceding word

[jtmaxwell3]

I added code that gathers statistics about how often each approved analysis occurs in the context of a previous word and in the context of a following word. This is used by GetContextScore in ComparePriorityCounts to order the analyses for a particular occurrence.

---

This change is 

[github-actions]

LCM Tests

    16 files  ±0      16 suites  ±0   2m 53s ⏱️ -1s
 2 828 tests +1   2 808 ✅ +1   20 💤 ±0  0 ❌ ±0 
11 260 runs  +4  11 081 ✅ +4  179 💤 ±0  0 ❌ ±0 

Results for commit d7b1b32. ± Comparison against base commit 93d7e90.

♻️ This comment has been updated with latest results.

[jasonleenaylor]

src/SIL.LCModel/DomainServices/AnalysisGuessServices.cs line 342 at r1 (raw file):

		/// <param name="occurrence"></param>
		/// <returns></returns>
		private IAnalysis GetBestAnalysis(ContextCount counts, AnalysisOccurrence occurrence)

The code in this method looks suspicious. I'm not sure how only checking the previousWordForm could be correct. I still have more code to read, but this has me scratching my head at the moment.

[jtmaxwell3]

The code is actually correct, although it does look suspicious. I will fix it.

[jasonleenaylor]

Reviewed 1 of 1 files at r2, 1 of 1 files at r3, all commit messages.
Reviewable status: complete! all files reviewed, all discussions resolved (waiting on @jtmaxwell3)