v0.53.2

What's Changed

• perf(weave): per-source union-all for custom-attrs schema query by @gtarpenning in #7502
• perf(weave): compute numeric-histogram bounds once (3 map scans -> 2) by @gtarpenning in #7504
• perf(weave): page-prefetch two-pass for agents list-view attribution by @gtarpenning in #7449
• perf(weave): prune the agents cost list-read via the two-pass by @gtarpenning in #7453
• chore(inference): fix 500 in catalog API by @jamie-rasmussen in #7515
• fix(weave_ts): re-route agent spans to the active project on weave.init() by @fixit-agent[bot] in #7512
• chore(weave_ts): Bump version, update changelog. by @drtangible in #7520
• perf(weave): two-pass filter CTE for calls_complete heavy reads by @gtarpenning in #7486
• chore(weave): add column_visibility to ComparisonViewDefinition by @gtarpenning in #7519
• perf(weave): add tuned skip indexes on calls_complete (trace_id + attributes_dump) by @gtarpenning in #7510
• fix(weave): soft-delete calls_complete via lightweight UPDATE, reclaim async by @gtarpenning in #7509
• feat(weave): Base64 and Data URL content ref conversion by @zbirenbaum in #7489
• chore(weave_ts): Move type to keep typedoc on Turn by @drtangible in #7521
• feat(weave_ts): nest LLM/Tool/SubAgent under a SubAgent by @rgao-coreweave in #7077
• chore(weave): require the (weave_ts) PR-title scope for TS SDK changes by @fixit-agent[bot] in #7522
• perf(weave): cache published content refs to dedupe repeat base64 blobs on ingest by @gtarpenning in #7527
• feat(weave): first-class sdk feedback for agent spans + turns by @gtarpenning in #7524
• fix(weave): capture Bedrock cache tokens in streaming converse by @fixit-agent[bot] in #7525
• fix(weave): pass through external refs embedded in JSON strings on read by @gtarpenning in #7530
• chore(weave): rename Turn/SubAgent factory methods to start_*() by @rgao-coreweave in #7526
• chore(weave): update costs by @jwlee64 in #7533
• perf(weave): batch content object inserts in calls_complete by @gtarpenning in #7528
• perf(weave): negative-cache mimetype-rejected base64 blobs on ingest by @gtarpenning in #7529
• fix(weave): map request-input validation failures to 400 not 500 by @gtarpenning in #7409
• perf(weave): narrow eval cost pricing to predict calls by @ro31337 in #7536
• fix(weave): idempotent migrations, additive 029, and partial-migration auto-recovery by @gtarpenning in #7496
• fix(weave): use real Bedrock cache token field names by @ro31337 in #7540
• fix(weave): drop stalling calls_complete reclaim DELETE by @gtarpenning in #7542
• feat(weave): report deleted versions in ObjDeleteRes by @ruhiparvatam in #7544
• perf(weave): skip latest alias write for content objects in calls_complete by @gtarpenning in #7537
• fix(weave): stabilize client-side digests across non-string dict keys by @gtarpenning in #7539
• chore(weave): cap openai below 2.45.0 for openai_agents shard by @gtarpenning in #7546
• feat(weave): nest LLM/Tool/SubAgent under Turn & SubAgent, add SubAgent.start_subagent() by @rgao-coreweave in #7133
• chore(weave): add gpt-5.6 by @jwlee64 in #7550
• fix(weave): agents-view conversation label uses opening user turn by @gtarpenning in #7545
• feat(weave): add agent eval span fields by @bcsherma in #7469
• feat(weave): filter agent conversations by signals (tags + ratings) - backend by @jtschoonhoven in #7477
• feat(weave): server-side ingest sampling for the spans model by @ro31337 in #7531
• fix(weave): fix duplicate migration numbers by @jtschoonhoven in #7553
• fix(weave): price W&B Inference models under the id usage records by @gtarpenning in #7549
• feat(weave): stamp eval metadata on agent spans by @bcsherma in #7476
• fix(weave): support google-adk 2.3.0+ trace_inference_result signature by @fixit-agent[bot] in #7560
• chore(weave): Update Remote Scoring samples with info about MT-Saas by @mscavezze-cw in #7563
• chore(weave): Remote Scoring docs and sample use released Weave by @mscavezze-cw in #7565
• perf(weave): flush file chunks and content objects concurrently in calls_complete by @gtarpenning in #7538
• fix(weave): include Anthropic prompt cache tokens in summary input_tokens by @ro31337 in #7561
• fix(weave): relax agent-search matched-message role to str by @gtarpenning in #7568
• feat(weave): denormalize scorer_trace_id onto feedback by @gtarpenning in #7555
• fix(weave): apply signal filters to agent stats by @jtschoonhoven in #7566
• perf(weave): batch content-object writes on the v1 call/upsert_batch path by @gtarpenning in #7564
• fix(weave): guard ref converter recursion against deeply-nested payloads by @gtarpenning in #7578
• fix(weave): trace direct openai calls under otel_v2 by @gtarpenning in #7570
• fix(weave): allow span_name in agent-spans group_by/filter by @gtarpenning in #7581
• fix(weave): normalize Claude Code chat projection by @bcsherma in #7579
• fix(weave_ts): propagate conversation attributes across runIsolated frames by @rgao-coreweave in #7577
• chore(weave_ts): bump TS SDK version to 0.16.3 by @rgao-coreweave in #7582
• chore(weave): add docstring to StringPrompt Python Class by @ngrayluna in #7583
• feat(weave_ts): default JS client to calls_complete ingest path by @gtarpenning in #7386
• feat(weave): export api contract by @gtarpenning in #7571

New Contributors

• @fixit-agent[bot] made their first contribution in #7512

Full Changelog: v0.53.1...v0.53.2