Repository navigation
Automated Sync from main to stable - #799
Merged
Merged
Conversation
Signed-off-by: Warren Snipes <contact@warrensnipes.dev>
… of duplicating them inline (mlflow#26461) Signed-off-by: Joshua Wong <joshua.wong@databricks.com> Co-authored-by: Isaac <no-reply@databricks.com>
Co-authored-by: copilot-swe-agent[bot] <198982749+Copilot@users.noreply.github.com> Co-authored-by: harupy <17039389+harupy@users.noreply.github.com>
Signed-off-by: Haoji Tang <haoji.tang@databricks.com> Co-authored-by: Codex <noreply@openai.com>
Co-authored-by: copilot-swe-agent[bot] <198982749+Copilot@users.noreply.github.com> Co-authored-by: harupy <17039389+harupy@users.noreply.github.com> Co-authored-by: Haoji Tang <haoji.tang@databricks.com>
Co-authored-by: copilot-swe-agent[bot] <198982749+Copilot@users.noreply.github.com> Co-authored-by: harupy <17039389+harupy@users.noreply.github.com> Co-authored-by: Haoji Tang <haoji.tang@databricks.com>
Signed-off-by: harupy <17039389+harupy@users.noreply.github.com> Co-authored-by: Claude Opus 5.5 (1M context) <noreply@anthropic.com> Co-authored-by: copilot-swe-agent[bot] <198982749+Copilot@users.noreply.github.com>
Signed-off-by: harupy <17039389+harupy@users.noreply.github.com> Co-authored-by: copilot-swe-agent[bot] <198982749+Copilot@users.noreply.github.com> Co-authored-by: harupy <17039389+harupy@users.noreply.github.com>
Co-authored-by: mlflow-app[bot] <mlflow-app[bot]@users.noreply.github.com>
Co-authored-by: copilot-swe-agent[bot] <198982749+Copilot@users.noreply.github.com> Co-authored-by: harupy <17039389+harupy@users.noreply.github.com>
Co-authored-by: copilot-swe-agent[bot] <198982749+Copilot@users.noreply.github.com> Co-authored-by: harupy <17039389+harupy@users.noreply.github.com>
…w#26439) Signed-off-by: Maha Yadavalli <271375718+maharanay22@users.noreply.github.com> Co-authored-by: Maha Yadavalli <271375718+maharanay22@users.noreply.github.com> Co-authored-by: Claude <noreply@anthropic.com> Co-authored-by: Aaron Teo <atwkdeveloper@gmail.com>
…26440) Signed-off-by: Maha Yadavalli <271375718+maharanay22@users.noreply.github.com> Signed-off-by: Aaron Teo <atwkdeveloper@gmail.com> Co-authored-by: Maha Yadavalli <271375718+maharanay22@users.noreply.github.com> Co-authored-by: Claude <noreply@anthropic.com> Co-authored-by: Aaron Teo <atwkdeveloper@gmail.com> Co-authored-by: Isaac <no-reply@databricks.com>
Co-authored-by: copilot-swe-agent[bot] <198982749+Copilot@users.noreply.github.com> Co-authored-by: harupy <17039389+harupy@users.noreply.github.com>
Co-authored-by: copilot-swe-agent[bot] <198982749+Copilot@users.noreply.github.com> Co-authored-by: harupy <17039389+harupy@users.noreply.github.com>
Co-authored-by: copilot-swe-agent[bot] <198982749+Copilot@users.noreply.github.com> Co-authored-by: harupy <17039389+harupy@users.noreply.github.com>
…6451) Signed-off-by: James Fletcher <james.fletcher@databricks.com>
Co-authored-by: copilot-swe-agent[bot] <198982749+Copilot@users.noreply.github.com> Co-authored-by: harupy <17039389+harupy@users.noreply.github.com>
Signed-off-by: harupy <17039389+harupy@users.noreply.github.com> Co-authored-by: copilot-swe-agent[bot] <198982749+Copilot@users.noreply.github.com> Co-authored-by: harupy <17039389+harupy@users.noreply.github.com>
Co-authored-by: mlflow-app[bot] <mlflow-app[bot]@users.noreply.github.com>
Co-authored-by: copilot-swe-agent[bot] <198982749+Copilot@users.noreply.github.com> Co-authored-by: harupy <17039389+harupy@users.noreply.github.com>
Signed-off-by: harupy <17039389+harupy@users.noreply.github.com> Co-authored-by: copilot-swe-agent[bot] <198982749+Copilot@users.noreply.github.com> Co-authored-by: harupy <17039389+harupy@users.noreply.github.com>
…s on Databricks (mlflow#26414) Signed-off-by: harupy <17039389+harupy@users.noreply.github.com> Signed-off-by: Harutaka Kawamura <hkawamura0130@gmail.com> Co-authored-by: Isaac <no-reply@databricks.com> Co-authored-by: Claude Opus 5.5 (1M context) <noreply@anthropic.com> Co-authored-by: Copilot Autofix powered by AI <175728472+Copilot@users.noreply.github.com>
…mlflow#26512) Co-authored-by: copilot-swe-agent[bot] <198982749+Copilot@users.noreply.github.com> Co-authored-by: harupy <17039389+harupy@users.noreply.github.com>
Signed-off-by: Haoji Tang <haoji.tang@databricks.com> Co-authored-by: Isaac <no-reply@databricks.com>
Signed-off-by: B-Step62 <yuki.watanabe@databricks.com> Co-authored-by: Codex <noreply@openai.com>
Signed-off-by: harupy <17039389+harupy@users.noreply.github.com> Co-authored-by: copilot-swe-agent[bot] <198982749+Copilot@users.noreply.github.com> Co-authored-by: harupy <17039389+harupy@users.noreply.github.com>
Co-authored-by: copilot-swe-agent[bot] <198982749+Copilot@users.noreply.github.com> Co-authored-by: harupy <17039389+harupy@users.noreply.github.com>
Signed-off-by: Haoji Tang <haoji.tang@databricks.com>
Co-authored-by: copilot-swe-agent[bot] <198982749+Copilot@users.noreply.github.com> Co-authored-by: harupy <17039389+harupy@users.noreply.github.com>
…low#26748) Signed-off-by: dependabot[bot] <support@github.com> Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>
Co-authored-by: copilot-swe-agent[bot] <198982749+Copilot@users.noreply.github.com> Co-authored-by: harupy <17039389+harupy@users.noreply.github.com>
Co-authored-by: copilot-swe-agent[bot] <198982749+Copilot@users.noreply.github.com> Co-authored-by: harupy <17039389+harupy@users.noreply.github.com>
Signed-off-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com> Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com> Co-authored-by: Harutaka Kawamura <hkawamura0130@gmail.com>
…s are saved (mlflow#26713) Signed-off-by: Yashwanth Kiran <yashwanthkiran2624@gmail.com>
Co-authored-by: copilot-swe-agent[bot] <198982749+Copilot@users.noreply.github.com> Co-authored-by: harupy <17039389+harupy@users.noreply.github.com>
Co-authored-by: mlflow-app[bot] <mlflow-app[bot]@users.noreply.github.com>
…ments (mlflow#26760) Signed-off-by: harupy <17039389+harupy@users.noreply.github.com> Co-authored-by: copilot-swe-agent[bot] <198982749+Copilot@users.noreply.github.com> Co-authored-by: harupy <17039389+harupy@users.noreply.github.com>
Co-authored-by: copilot-swe-agent[bot] <198982749+Copilot@users.noreply.github.com> Co-authored-by: harupy <17039389+harupy@users.noreply.github.com>
…upload` to a fixed length (mlflow#26756) Signed-off-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com> Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com> Co-authored-by: Harutaka Kawamura <hkawamura0130@gmail.com>
mlflow#26765) Signed-off-by: mlflow-app[bot] <191841109+mlflow-app[bot]@users.noreply.github.com> Co-authored-by: mlflow-app[bot] <191841109+mlflow-app[bot]@users.noreply.github.com>
…ry span twice with a shared tracer provider (mlflow#26734) Signed-off-by: Levi Long <long.levi823@gmail.com> Co-authored-by: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Signed-off-by: Khaled Sulayman <ksulayma@redhat.com> Signed-off-by: Khaled Sulayman <55605301+khaledsulayman@users.noreply.github.com> Signed-off-by: Nehanth Narendrula <nehanthnarendrula@gmail.com> Signed-off-by: Robin Narsingh Ranabhat <robinnarsingha123@gmail.com> Signed-off-by: Nehanth <nehanthnarendrula@gmail.com> Signed-off-by: Hamid Moghani <hmoghani@gmail.com> Signed-off-by: Mike Sager <53917070+msager27@users.noreply.github.com> Signed-off-by: Ryan Barden <79597212+bardenr@users.noreply.github.com> Signed-off-by: Juntao Wang <37624318+DaoDaoNoCode@users.noreply.github.com> Signed-off-by: Aakanksha Duggal <aduggal@redhat.com> Signed-off-by: Nana Nosirova <10577112+nananosirova@users.noreply.github.com> Signed-off-by: mprahl <mprahl@users.noreply.github.com> Co-authored-by: Khaled Sulayman <55605301+khaledsulayman@users.noreply.github.com> Co-authored-by: Nehanth Narendrula <nehanthnarendrula@gmail.com> Co-authored-by: Claude Fable 5.1 <noreply@anthropic.com> Co-authored-by: Robin Narsingh Ranabhat <robinnarsingha123@gmail.com> Co-authored-by: Copilot Autofix powered by AI <175728472+Copilot@users.noreply.github.com> Co-authored-by: Hamid Moghani <hmoghani@gmail.com> Co-authored-by: Mike Sager <53917070+msager27@users.noreply.github.com> Co-authored-by: Codex <noreply@openai.com> Co-authored-by: Ryan Barden <79597212+bardenr@users.noreply.github.com> Co-authored-by: Juntao Wang <37624318+DaoDaoNoCode@users.noreply.github.com> Co-authored-by: Aakanksha Duggal <aduggal@redhat.com> Co-authored-by: Codex <codex@openai.com> Co-authored-by: Nana Nosirova <10577112+nananosirova@users.noreply.github.com>
Signed-off-by: mprahl <mprahl@users.noreply.github.com>
Reconstructed net downstream changes from v3.17.0 to 29439eb. Omit covered backport a3dbabf (upstream original 9b574f2). Original history preserved in backup/odh-before-snapshot-f566a42d-20261016. Source commits touching this group (including historical ancestry): 1dfef78 keep: Fix E2E readiness, fixtures, and dataset refresh cff2624 keep: Fix dashboard E2E startup and health checks for Konflux images 3d4ecdb keep: Fix Konflux build detection and image tags for E2E tests aa18b3a keep: Rely on the upstream runtime AI Gateway toggle 14dd969 keep: Fix 3.17 traces UI regressions in standalone and federated mode 1cafe96 keep: Align migration snapshot string inference across pandas versions 5538a48 keep: Fix v3.17.0 rebase CI compatibility checks c47f44a keep: Record the MLflow v3.17.0 rebase and local validation b08c113 keep: UI changes (squashed) df936d7 keep: Fork scaffolding (squashed) Co-authored-by: Codex <noreply@openai.com> Signed-off-by: Humair Khan <HumairAK@users.noreply.github.com>
Reconstructed net downstream changes from v3.17.0 to 29439eb. Omit covered backport a3dbabf (upstream original 9b574f2). Original history preserved in backup/odh-before-snapshot-f566a42d-20261016. Source commits touching this group (including historical ancestry): 26b7b01 keep: Hide telemetry and AI gateway UI when disabled 1cafe96 keep: Align migration snapshot string inference across pandas versions 5538a48 keep: Fix v3.17.0 rebase CI compatibility checks 6d4568b keep: Include the Starlette TestClient dependency in test requirements b08c113 keep: UI changes (squashed) fa81b17 keep: Backend changes (squashed) df936d7 keep: Fork scaffolding (squashed) Co-authored-by: Codex <noreply@openai.com> Signed-off-by: Humair Khan <HumairAK@users.noreply.github.com>
Reconstructed net downstream changes from v3.17.0 to 29439eb. Omit covered backport a3dbabf (upstream original 9b574f2). Original history preserved in backup/odh-before-snapshot-f566a42d-20261016. Source commits touching this group (including historical ancestry): 8603f5b keep: Ignore zod SNYK-JS-ZOD-20510278 until @a2ui supports zod 4 c3f59e0 keep: Wait for model registration and isolate E2E retries 1dfef78 keep: Fix E2E readiness, fixtures, and dataset refresh 26b7b01 keep: Hide telemetry and AI gateway UI when disabled a3dbabf drop: Restore `View All` for time-filtered OSS traces (mlflow#26593) aa18b3a keep: Rely on the upstream runtime AI Gateway toggle 14dd969 keep: Fix 3.17 traces UI regressions in standalone and federated mode 4ffef06 keep: Preserve host routing for single chat sessions 2f67566 keep:override SimpleSelect dropdown style b08c113 keep: UI changes (squashed) df936d7 keep: Fork scaffolding (squashed) Co-authored-by: Codex <noreply@openai.com> Signed-off-by: Humair Khan <HumairAK@users.noreply.github.com>
Signed-off-by: Humair Khan <HumairAK@users.noreply.github.com>
Co-authored-by: Codex <noreply@openai.com> Signed-off-by: Humair Khan <HumairAK@users.noreply.github.com>
Regenerate container requirements and validate the released wheel against the installed snapshot on Python 3.12. The release test suite passes 452 tests with 1 skipped, and all 66 Skills Registry route/method pairs pass authorization coverage. Co-authored-by: Codex <noreply@openai.com> Signed-off-by: Humair Khan <HumairAK@users.noreply.github.com>
Use the installed psycopg2 driver for default PostgreSQL URLs and require psycopg only for explicit postgresql+psycopg URLs. This avoids an unnecessary pip installation failing with permission denied in the non-root ODH image. Validated the exact cf52c9c ODH image as UID 1001 in an isolated Compose stack: HTTP health returned OK and a PostgreSQL query succeeded through psycopg2. Required pre-commit checks passed. Co-authored-by: Codex <noreply@openai.com> Signed-off-by: Humair Khan <HumairAK@users.noreply.github.com>
Rebase ODH MLflow onto interim upstream snapshot f566a42
Sync odh -> RHDS
…-dashboard Carry the downstream keep commits from #408 onto the f566a42 snapshot, which already includes the upstream Skill Registry, so the two backports there are dropped. PatternFly styling: - Fill the PF override gaps on the Skill Registry pages: room for the copy button on code snippets, the version-list hover carve-out keyed on aria-selected, primary action before the overflow menu, native file pickers as PF secondary buttons, and sidebar font and label fixes. - Give filter rows, typeahead comboboxes and toolbar search fields the PF form-control frame. Module federation and embedding: - Expose ./MlflowSkillRegistryWrapper (default basename /ai-hub/skills/registry) with relative routes for the catalog and both detail URL forms and a breadcrumb reporter. Embedded, the catalog and detail pages leave the title and breadcrumb to the host. - On the host's Agent traces page (GenAI workflow locked), list only experiments whose mlflow.experimentKind starts with genai_development, tag experiments created there as genai_development, and name the root breadcrumb "Agent traces". - On the host's Experiments page, bring back the GenAI / Model training toggle. Picking GenAI shows a notice that links to the experiment on the Agent traces page; leaving the experiment resets the toggle. Backend: - When a trace is written (start_trace, log_spans including OTLP, v2 start_trace), set mlflow.experimentKind=genai_development_inferred on experiments that have no kind yet, cached per workspace and experiment for an hour. The check and insert run in one session on the primary database and only insert, so a kind that a read replica hasn't caught up on, or that is set concurrently, is never overwritten. A failure never fails the trace write. - In the skill registry migration (e7d1f4b2a9c6), give the same inferred kind to experiments in any workspace that already have traces but no kind, so experiments traced before the upgrade still show on the Agent traces page. It lives in that upstream migration, not a new revision, so the head revision stays aligned with upstream. Also reset the URL between ExperimentListView tests, since upstream's new search-filter test leaves a filter in it. Co-authored-by: Nana Nosirova <10577112+nananosirova@users.noreply.github.com> Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Signed-off-by: Juntao Wang <juntwang@redhat.com>
keep: Restyle and federate Skill Registry, embed Agent traces for odh…
| @@ -0,0 +1,51 @@ | |||
| function removeHtmlComments(text) { | |||
| return text.replace(/<!--[\s\S]*?-->/g, ""); | |||
| data = json.loads(session_file.read_text()) | ||
| except FileNotFoundError: | ||
| # Missing, or removed by a concurrent expiry. | ||
| return None |
| return Session.from_dict(data) | ||
|
|
||
| @staticmethod | ||
| def create(context: dict[str, Any] | None = None, working_dir: Path | None = None) -> Session: |
| _expire_session(entry.stem) | ||
| except ValueError: | ||
| # Not a valid session id, so nothing else is kept for it. | ||
| entry.unlink(missing_ok=True) |
| self.send_header("Content-Type", content_type) | ||
| self.send_header("Content-Length", str(len(body))) | ||
| for key, value in (headers or {}).items(): | ||
| self.send_header(key, value) |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Automated Sync from main to stable
This PR automatically syncs the
mainbranch to thestablebranch by opening a pull request frommainintostable.Sync Summary
Latest commit:
00cdc4b12Merge pull request keep: Restyle and federate Skill Registry, embed Agent traces for odh… #798 from opendatahub-io/masterTotal commits to sync: 172
Source:
https://github.com/red-hat-data-services/mlflow.git@mainTarget:
https://github.com/red-hat-data-services/mlflow.git@stablePR head:
mainCommits to be synced
00cdc4b12Merge pull request keep: Restyle and federate Skill Registry, embed Agent traces for odh… #798 from opendatahub-io/master06275d92fkeep: Restyle and federate Skill Registry, embed Agent traces for odh-dashboardfdc78f56fMerge pull request Sync odh -> RHDS #797 from opendatahub-io/master320eab5e5Merge pull request chore(deps): update registry.access.redhat.com/ubi9/python-311 docker digest to 0cac2a0 (rhoai-3.3) #409 from opendatahub-io/rebase-snapshot-f566a42d950bf506akeep: Match Compose PostgreSQL dependencies to the configured drivercf52c9cbckeep: Use released Kubernetes auth plugin 2.1.0 for Skills Registryc4c395aa9keep: Record interim upstream snapshot rebase and validation3f20770b5Merge preserved ODH history into upstream snapshot f566a42b92bd3f16keep: UI changes for upstream snapshot f566a42 (squashed)c483ba76akeep: Backend changes for upstream snapshot f566a42 (squashed)5d231b457keep: Fork scaffolding for upstream snapshot f566a42 (squashed)f566a42d6Show tracking connection in skill registry snippets (Show tracking connection in skill registry snippets mlflow/mlflow#26768)fb519dbbcAdd the skill registry (Add the skill registry mlflow/mlflow#26736)d5a65b716Fixstrands,semantic_kernelandhaystackautolog exporting every span twice with a shared tracer provider (Fixstrands,semantic_kernelandhaystackautolog exporting every span twice with a shared tracer provider mlflow/mlflow#26734)1c34998bcPreservetimestamp=0inmlflow.log_metricandmlflow.log_metrics(Preservetimestamp=0inmlflow.log_metricandmlflow.log_metricsmlflow/mlflow#26765)74da7ae63Pad Azure block IDs inAzureBlobArtifactRepository.create_multipart_uploadto a fixed length (Pad Azure block IDs inAzureBlobArtifactRepository.create_multipart_uploadto a fixed length mlflow/mlflow#26756)be8f73367RestoreGITHUB_TOKENfor triage issue comments (RestoreGITHUB_TOKENfor triage issue comments mlflow/mlflow#26763)fa64cc386Publish triage PRs withmlflow-appwhile retaining approval requirements (Publish triage PRs withmlflow-appwhile retaining approval requirements mlflow/mlflow#26760)73b1a7dd5Update model catalog from upstream sources (Update model catalog from upstream sources mlflow/mlflow#26757)e03aec38fConstrain Transformers for PyTorch 2.5/2.6 model tests (Constrain Transformers for PyTorch 2.5/2.6 model tests mlflow/mlflow#26752)b5d4d8e50FixMlflowStoragemarking Optuna trials finished before their values are saved (FixMlflowStoragemarking Optuna trials finished before their values are saved mlflow/mlflow#26713)60608bcccRefresh the datasets list after creating a dataset (Refresh the datasets list after creating a dataset mlflow/mlflow#26682)1f790dcfbInclude Gemini thinking tokens in Gateway output usage (Include Gemini thinking tokens in Gateway output usage mlflow/mlflow#26746)2d09f29c8Respect issue authors’ offers to fix bugs (Respect issue authors’ offers to fix bugs mlflow/mlflow#26750)2a5a87a90Bump the actions-minor group across 2 directories with 4 updates (Bump the actions-minor group across 2 directories with 4 updates mlflow/mlflow#26748)e71963015Addsession_levelto session-level scorers for hierarchical sessions (Addsession_levelto session-level scorers for hierarchical sessions mlflow/mlflow#26443)a3caeb850Update model catalog from upstream sources (Update model catalog from upstream sources mlflow/mlflow#26749)f3bdbe2f3Updateuv.lock(Updateuv.lockmlflow/mlflow#26745)107b93ec3Count Geminithoughts_token_countas output tokens inmlflow.gemini.autolog(Count Geminithoughts_token_countas output tokens inmlflow.gemini.autologmlflow/mlflow#26744)c2372af7dAllowsession_idto be a list inupdate_current_traceandtracing.contextto identify hierarchical sessions (Allowsession_idto be a list inupdate_current_traceandtracing.contextto identify hierarchical sessions mlflow/mlflow#26413)c4736e82dRemove ignored workflow filters fromdev/README.mdbadge links (Remove ignored workflow filters fromdev/README.mdbadge links mlflow/mlflow#26737)a4236e272Pindatasetsfor sentence-transformers versions below 5.2 (Pindatasetsfor sentence-transformers versions below 5.2 mlflow/mlflow#26735)8c45a446eKeep separate bars for same-named runs in bar chart image export (Keep separate bars for same-named runs in bar chart image export mlflow/mlflow#26732)34119f827HandleNonedefault for datetimeParamSpecinto_dict(HandleNonedefault for datetimeParamSpecinto_dictmlflow/mlflow#26729)fde9a9498Don't modify the caller's records inEvaluationDataset.merge_records(Don't modify the caller's records inEvaluationDataset.merge_recordsmlflow/mlflow#26730)6485ea856Record dataset, scorer, and agent lineage onmlflow.genai.evaluateruns (Record dataset, scorer, and agent lineage onmlflow.genai.evaluateruns mlflow/mlflow#26543)f1d2fe097Remove date cutoff from PR auto-close policy (Remove date cutoff from PR auto-close policy mlflow/mlflow#26728)e784ca564Notify issue reporters when triage files a PR (Notify issue reporters when triage files a PR mlflow/mlflow#26727)3cd892586Remove the maintainer-only gate from triage PRs (Remove the maintainer-only gate from triage PRs mlflow/mlflow#26725)602fd20b0Update model catalog from upstream sources (Update model catalog from upstream sources mlflow/mlflow#26724)b033c7686Normalize dataset list timestamps inuseDatasetsPageQuery(Normalize dataset list timestamps inuseDatasetsPageQuerymlflow/mlflow#26721)71d9a4018Update version to 3.17.1.dev0 (Update version to 3.17.1.dev0 mlflow/mlflow#26718)28be22e98Update model catalog from upstream sources (Update model catalog from upstream sources mlflow/mlflow#26716)bfb44702dRespectMLFLOW_DISABLE_TELEMETRYin the UI banner and settings (RespectMLFLOW_DISABLE_TELEMETRYin the UI banner and settings mlflow/mlflow#26699)ae1df4963Preservemodel_apiwhen recreating judges in DSPy alignment optimizers (Preservemodel_apiwhen recreating judges in DSPy alignment optimizers mlflow/mlflow#26707)a9f5d6ad5Update CHANGELOG for 3.17.1 (Update CHANGELOG for 3.17.1 mlflow/mlflow#26708)94140d9e9Let the MLflow Assistant's gateway provider work on an authenticated server (Let the MLflow Assistant's gateway provider work on an authenticated server mlflow/mlflow#25851)be57ca891Support Decision API in make_judge() (Support Decision API in make_judge() mlflow/mlflow#26578)2f83d1e1dLet authenticated remote clients edit their own MLflow Assistant provider settings (Let authenticated remote clients edit their own MLflow Assistant provider settings mlflow/mlflow#25847)beb1f3eb2Pass trace inputs and outputs to custom scorers that the server runs (Pass trace inputs and outputs to custom scorers that the server runs mlflow/mlflow#26690)5bb8e4be4Fix the docker job executor tests on Windows (Fix the docker job executor tests on Windows mlflow/mlflow#26701)364951ef7Installpsycopgalongsidepsycopg2-binaryin Docker Compose (Installpsycopgalongsidepsycopg2-binaryin Docker Compose mlflow/mlflow#26694)c50c1d6eeUpdate model catalog from upstream sources (Update model catalog from upstream sources mlflow/mlflow#26700)9d651c2f1Fixmlflow.openai.autolog()tracing for streams closed early (Fixmlflow.openai.autolog()tracing for streams closed early mlflow/mlflow#26667)68072cd61Build the sandbox fallback image from the server's own source (Build the sandbox fallback image from the server's own source mlflow/mlflow#26536)9b574f263RestoreView Allfor time-filtered OSS traces (RestoreView Allfor time-filtered OSS traces mlflow/mlflow#26593)cd70aa615Readgen_ai.usage.cache_write.input_tokensinGenAiTranslator(Readgen_ai.usage.cache_write.input_tokensinGenAiTranslatormlflow/mlflow#26680)d448f59e4Preventrun_dev_server.pyfrom orphaning starting children (Preventrun_dev_server.pyfrom orphaning starting children mlflow/mlflow#26666)8dfd1989bAllow distinct captions for triage after videos (Allow distinct captions for triage after videos mlflow/mlflow#26688)e1f00c2f4AddTraceAudioto custom trace views (AddTraceAudioto custom trace views mlflow/mlflow#26452)653102901Report the triage browser as online (Report the triage browser as online mlflow/mlflow#26686)96699be9fKeep triage UI servers alive, start them on demand, and stop them before exiting (Keep triage UI servers alive, start them on demand, and stop them before exiting mlflow/mlflow#26683)36a1db061Add an experimentaldockerjob executor for custom scorer runs (Add an experimentaldockerjob executor for custom scorer runs mlflow/mlflow#26430)97d106178Use live dev servers for UI bug triage (Use live dev servers for UI bug triage mlflow/mlflow#26679)4ec1dc3b0Usere.fullmatchin name/ID validators to reject trailing newlines (Usere.fullmatchin name/ID validators to reject trailing newlines mlflow/mlflow#26485)f2a1e4a64Disable experiments Delete button when no checked experiment is in the current list (Disable experiments Delete button when no checked experiment is in the current list mlflow/mlflow#26677)376f47a60Clarify video capture guidance for bug triage (Clarify video capture guidance for bug triage mlflow/mlflow#26678)2d181b1c9Load latest Databricks scorers from the scheduled-scorers REST config so feedback records the scorer version (Load latest Databricks scorers from the scheduled-scorers REST config so feedback records the scorer version mlflow/mlflow#26542)2b5eac9adPreserve datetime values inArray(datetime)prediction inputs (Preserve datetime values inArray(datetime)prediction inputs mlflow/mlflow#26671)efa4c49bfRetry transient Docker builds ingenerate-protos.sh(Retry transient Docker builds ingenerate-protos.shmlflow/mlflow#26673)2e195da10Update model catalog from upstream sources (Update model catalog from upstream sources mlflow/mlflow#26672)074c0fcf4Updateuv.lock(Updateuv.lockmlflow/mlflow#26670)6081fa6a2Fix nested objectpropertiesbeing stripped from function tool schemas (Fix nested objectpropertiesbeing stripped from function tool schemas mlflow/mlflow#26646)f9d3759e1Reset prompts list pagination when the search filter changes (Reset prompts list pagination when the search filter changes mlflow/mlflow#26665)274688c73Reset sidebar experiment menu when returning to the experiments list (Reset sidebar experiment menu when returning to the experiments list mlflow/mlflow#26661)e4e2440a3Add captions to triage videos (Add captions to triage videos mlflow/mlflow#26662)22838e03dDocument video recording for multi-action triage bugs (Document video recording for multi-action triage bugs mlflow/mlflow#26658)67bcc563fInstallffmpegfor triage video recordings (Installffmpegfor triage video recordings mlflow/mlflow#26653)7bcc4b5caFix unreadable explanatory text inRegisterModelFormin dark mode (Fix unreadable explanatory text inRegisterModelFormin dark mode mlflow/mlflow#26652)0bc19586bMake triage artifacts survive failed-job re-runs (Make triage artifacts survive failed-job re-runs mlflow/mlflow#26650)f568b5aafAllow proportionate verification for bug triage fixes (Allow proportionate verification for bug triage fixes mlflow/mlflow#26649)f39e81f97UsetextPrimaryfor evaluation score-change indicators (UsetextPrimaryfor evaluation score-change indicators mlflow/mlflow#26645)b8292f6d2Accept all NumPy integer dtypes for document IDs in retriever metrics (Accept all NumPy integer dtypes for document IDs in retriever metrics mlflow/mlflow#26643)1a72e08e0Show zero input/output token counts with their labels inTokensCellhover card (Show zero input/output token counts with their labels inTokensCellhover card mlflow/mlflow#26642)1dff5398fShow OpenAI Decisions in the v2 trace viewer (Show OpenAI Decisions in the v2 trace viewer mlflow/mlflow#26577)c18cca059Use theme border color for chat session metrics divider in dark mode (Use theme border color for chat session metrics divider in dark mode mlflow/mlflow#26640)b7a5c98c1Export list-form OpenAI system message content togen_ai.system_instructions(Export list-form OpenAI system message content togen_ai.system_instructionsmlflow/mlflow#26470)d2e9e06b7Join all text parts inGeminiAdapterresponses (Join all text parts inGeminiAdapterresponses mlflow/mlflow#26446)0e844f2f2Add hot-reload guidance for frontend fix verification (Add hot-reload guidance for frontend fix verification mlflow/mlflow#26633)3fd52d475Log traceback when GenAI semconv message conversion fails intranslate_span_to_genai(Log traceback when GenAI semconv message conversion fails intranslate_span_to_genaimlflow/mlflow#26632)37dc089f0Clear stale error inShowArtifactTextViewwhen loading a new artifact (Clear stale error inShowArtifactTextViewwhen loading a new artifact mlflow/mlflow#26631)225681ab7Update bug triage screenshot guidance (Update bug triage screenshot guidance mlflow/mlflow#26629)3881be79fAllow ready triage bot PRs incopilot-review.yml(Allow ready triage bot PRs incopilot-review.ymlmlflow/mlflow#26628)0ab57f5e4Require the sandbox for remote MLflow Assistant tool execution and cap remote permissions (Require the sandbox for remote MLflow Assistant tool execution and cap remote permissions mlflow/mlflow#25844)6a643eab8Fix audio artifact preview for logged models inShowArtifactAudioView(Fix audio artifact preview for logged models inShowArtifactAudioViewmlflow/mlflow#26626)a4d5f3e7bAccept lowercase comparators insearch_experimentsfor SQL stores (Accept lowercase comparators insearch_experimentsfor SQL stores mlflow/mlflow#26619)ccad0a99aFix boolean values rendering as empty cells in CSV artifact preview (Fix boolean values rendering as empty cells in CSV artifact preview mlflow/mlflow#26622)3190ac489Trigger labeling when a draft PR is ready for review (Trigger labeling when a draft PR is ready for review mlflow/mlflow#26620)12281e9d4Encourage useful visual evidence in triage PRs (Encourage useful visual evidence in triage PRs mlflow/mlflow#26623)a866c902aFix@mlflow.tracerecordingselfin span inputs for falsy instances (Fix@mlflow.tracerecordingselfin span inputs for falsy instances mlflow/mlflow#26617)Merging
GitHub automerge is enabled for this pull request once required checks pass.