feat: add AIConfigTracker, metrics & resumption tokens (AIC-2664)

[ctawiah]

Requirements

• I have added test coverage for new or changed functionality
• I have followed the repository's pull request submission guidelines
• I have validated my changes against all supported platform versions

Related issues

AIC-2664 — Step 4: AITRACK. Stacked on #173 (Step 3).

Describe the solution you've provided

Implements the AITRACK surface to spec, thread-safe by construction:

• LDAIConfigTracker gains the full method set: trackDuration, trackTimeToFirstToken, trackSuccess/trackError, trackFeedback, trackTokens, trackToolCall(s), trackJudgeResult, plus the trackDurationOf and trackMetricsOf wrappers, getTrackData(), getResumptionToken(), and getSummary(). Event names match the spec and the JS/Python SDKs.
• LDAIConfigTrackerImpl (internal): per-run UUID runId; record-once metrics use atomic claim-before-emit (AtomicBoolean) so exactly one event is emitted under concurrency (trackSuccess/trackError share one guard). Tool-call and judge-result events are not once-only; tool calls accumulate in a CopyOnWriteArrayList and getSummary().getToolCalls() returns an immutable snapshot.
• Validation: negative durations and token counts are clamped to zero; individual token counts emit only when positive; null args are guarded; a null judge score is treated as "no score" and is distinct from a legitimate 0.0. trackMetricsOf records an error and rethrows on both operation and extractor failures.
• Resumption tokens (ResumptionTokens): URL-safe Base64 (no padding) of canonical JSON in fixed key order runId, configKey, variationKey, version, graphKey. variationKey is always emitted for cross-SDK parity; modelName/providerName are not carried (restored trackers report ""). Decoding strictly type-validates each field and rejects malformed or oversized (>4 KB) tokens.
• Wiring: LDAIClientImpl now produces real trackers from createTracker() (a fresh runId per call) and adds LDAIClient.createTracker(token, context) to reconstruct a run across process boundaries. The placeholder NoOpAIConfigTracker is removed.
• The new tracking value types are consolidated in LDAITrackingTypes (mirroring the LDAIConfigTypes pattern).

Tests

• Resumption token: byte-compatible with fixed fixtures, round-trips, strict decode (missing/mistyped/oversized rejected), escaping.
• Tracker: per-event semantics, at-most-once, clamping, judge-result zero-vs-null, trackMetricsOf rethrow + error, summary, and concurrency (N threads → exactly one once-only event, intact tool-call list).
• Client wiring: tracker carries variation/model metadata, token reconstruction shares runId, each createTracker() starts a new run.

Out of scope (per ticket)

No Judge/Evaluator (Step 5), no AIGRAPH createGraphTracker, no provider-specific trackOpenAIMetrics/trackBedrockMetrics (post-1.0).

Made with Cursor

---

Note

Medium Risk
New public API surface and telemetry behavior on every AI config evaluation; resumption token parsing handles untrusted input with size limits and strict validation.

Overview
Replaces the placeholder no-op LDAIConfigTracker with a full AITRACK implementation that emits AI run metrics through LDClient.trackMetric, keyed by a per-run runId.

LDAIConfigTracker is expanded from a stub to the full API: duration, time-to-first-token, success/error, feedback, tokens, tool calls, judge results, trackDurationOf / trackMetricsOf, getTrackData(), getResumptionToken(), and getSummary(). LDAITrackingTypes adds the public value types (TokenUsage, Metrics, JudgeResult, etc.).

LDAIConfigTrackerImpl sends spec-aligned $ld:ai:* events with shared correlation fields; record-once metrics use atomic guards for thread safety. ResumptionTokens encodes/decodes cross-process resumption (cross-SDK byte fixtures in tests). LDAIClientImpl wires real trackers (new UUID per createTracker() on configs) and adds LDAIClient.createTracker(token, context) for deferred events; NoOpAIConfigTracker is removed.

^{Reviewed by Cursor Bugbot for commit 19d0f4f. Bugbot is set up for automated code reviews on this repo. Configure here.}