fix previousKey returning a predecessor for an absent key

[rootvector2]

AbstractSortedMapDecorator.previousKey and DualTreeBidiMap.previousKey compute the predecessor with headMap(key).lastKey() without first checking that key is present, so an absent key returns a predecessor instead of null, disagreeing with the sibling nextKey fixed in #728 and with every other OrderedMap.previousKey (AbstractLinkedMap, ListOrderedMap, TreeBidiMap, PatriciaTrie, SingletonMap all return null for an absent key); this guards with containsKey and uses NavigableMap.lowerKey when the map supports it, mirroring the merged nextKey. Found while addressing the nextKey review in #728, where a separate PR for previousKey was requested.

• Read the contribution guidelines for this project.
• Read the ASF Generative Tooling Guidance if you use Artificial Intelligence (AI).
• I used AI to create any part of, or all of, this pull request. Which AI tool was used to create this pull request, and to what extent did it contribute? Claude Code helped locate the inconsistency, draft the fix, and write the regression tests; I reviewed and verified the change and ran the full build.
• Run a successful build using the default Maven goal with mvn; that's mvn on the command line by itself.
• Write unit tests that match behavioral changes, where the tests fail if the changes to the runtime are not applied. This may not always be possible, but it is a best practice.
• Write a pull request description that is detailed enough to understand what the pull request does, how, and why.
• Each commit in the pull request should have a meaningful subject line and body. Note that a maintainer may squash commits during the merge process.