fix: reset nodeIds during fullsnapshot (#412)

## Summary

Minor fixes to clear state during on fullsnapshot

<!-- CURSOR_SUMMARY -->
---

> [!NOTE]
> **Low Risk**
> Small, localized state-reset change in replay event generation; low
risk aside from potentially affecting downstream expectations about node
ID reuse.
> 
> **Overview**
> Resets session replay DOM-image mapping state when generating a
`FULL_SNAPSHOT` by clearing `nodeIds` at the start of
`generateCaptureFullEvents`.
> 
> This prevents stale image signature→node ID mappings from leaking
across multiple full captures in the same session, improving correctness
of subsequent mutation/removal events.
> 
> <sup>Written by [Cursor
Bugbot](https://cursor.com/dashboard?tab=bugbot) for commit
f7e7da8. This will update automatically
on new commits. Configure
[here](https://cursor.com/dashboard?tab=bugbot).</sup>
<!-- /CURSOR_SUMMARY -->

Author: abelonogov-ld

sdk/@launchdarkly/observability-android/lib/src/main/kotlin/com/launchdarkly/observability/replay/exporter/RRWebEventGenerator.kt

@@ -190,6 +190,7 @@ class RRWebEventGenerator(
      */
     fun generateCaptureFullEvents(exportFrame: ExportFrame): List<Event> {
         if (exportFrame.addImages.isEmpty()) return emptyList()
+        nodeIds.clear()
         val eventBatch = mutableListOf<Event>()
         knownKeyFrameId = exportFrame.keyFrameId