codeant-ai-for-open-source[bot] commented on code in PR #41606: URL: https://github.com/apache/superset/pull/41606#discussion_r3588437479
########## superset/mcp_service/dashboard/tool/manage_dashboard_roles.py: ########## @@ -0,0 +1,329 @@ +# Licensed to the Apache Software Foundation (ASF) under one +# or more contributor license agreements. See the NOTICE file +# distributed with this work for additional information +# regarding copyright ownership. The ASF licenses this file +# to you under the Apache License, Version 2.0 (the +# "License"); you may not use this file except in compliance +# with the License. You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, +# software distributed under the License is distributed on an +# "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY +# KIND, either express or implied. See the License for the +# specific language governing permissions and limitations +# under the License. + +""" +Manage dashboard access roles FastMCP tool + +Adds/removes role-based dashboard access via explicit operations. Companion +to ``manage_dashboard_owners`` — dashboard access roles are dropped from the +generic ``update_dashboard`` tool because a full-replacement access-control +list silently widens or narrows who can see a dashboard. + +"Roles" are modeled as ROLE-type entries in the dashboard's Subject-based +``viewers`` list (apache/superset#38831 replaced the legacy +``roles``/``DASHBOARD_RBAC`` relationship with a unified Subject model +covering User/Role/Group, gated by the ``ENABLE_VIEWERS`` feature flag +instead of ``DASHBOARD_RBAC``). Any USER- or GROUP-type viewers already on +the dashboard are preserved untouched by this tool. +""" + +import logging +from typing import Any + +from fastmcp import Context +from sqlalchemy.exc import SQLAlchemyError +from superset_core.mcp.decorators import tool, ToolAnnotations + +from superset.commands.dashboard.exceptions import ( + DashboardAccessDeniedError, + DashboardNotFoundError, +) +from superset.exceptions import SupersetSecurityException +from superset.extensions import db, event_logger +from superset.mcp_service.dashboard.schemas import ( + ManageDashboardRolesRequest, + ManageDashboardRolesResponse, +) +from superset.mcp_service.system.schemas import serialize_subject_object +from superset.mcp_service.utils.url_utils import get_superset_base_url +from superset.subjects.types import SubjectType + +logger: logging.Logger = logging.getLogger(__name__) + + +def _find_and_authorize_dashboard( + identifier: int | str, +) -> tuple[Any, ManageDashboardRolesResponse | None]: + """Return (dashboard, None) on success or (None, error_response) on failure. + + Mirrors the helper in ``update_dashboard``: avoids ImportError before + Flask app initialisation by co-locating the imports it needs with the + call site rather than importing them at module load time. + """ + from superset import security_manager + from superset.daos.dashboard import DashboardDAO + + try: + dashboard = DashboardDAO.get_by_id_or_slug(identifier) + except DashboardAccessDeniedError: + # get_by_id_or_slug re-checks view access and raises access-denied + # for dashboards the caller cannot see; surface it as the + # structured permission_denied response instead of an unhandled + # error. + return None, ManageDashboardRolesResponse( + permission_denied=True, + error=( + "You do not have permission to access this dashboard. " + "Ask the user to grant access; do not retry." + ), + ) + except DashboardNotFoundError: + return None, ManageDashboardRolesResponse( + error=f"Dashboard not found: {identifier!r}", + ) + except SQLAlchemyError: + logger.exception("Database error looking up dashboard %r", identifier) + return None, ManageDashboardRolesResponse( + error="Failed to look up dashboard due to a database error.", + ) + + if dashboard is None: + return None, ManageDashboardRolesResponse( + error=f"Dashboard not found: {identifier!r}", + ) + + try: + security_manager.raise_for_editorship(dashboard) + except SupersetSecurityException: + return None, ManageDashboardRolesResponse( + permission_denied=True, + error=( + f"You don't have permission to edit dashboard " + f"'{dashboard.dashboard_title}' (ID: {dashboard.id})." + ), + ) + + return dashboard, None + + +def _dashboard_url(dashboard: Any) -> str: + """Build the user-facing dashboard URL, preferring slug over id.""" + return f"{get_superset_base_url()}/dashboard/{dashboard.slug or dashboard.id}/" + + +def _viewer_role_ids(dashboard: Any) -> list[int]: + """Role IDs behind the dashboard's ROLE-type viewer subjects.""" + return [ + subject.role_id + for subject in dashboard.viewers + if subject.type == SubjectType.ROLE + ] + + +def _other_viewers(dashboard: Any) -> list[Any]: + """Non-ROLE-type viewers (USER/GROUP subjects), preserved untouched.""" + return [ + subject for subject in dashboard.viewers if subject.type != SubjectType.ROLE + ] + + +def _compute_new_role_ids( + dashboard: Any, request: ManageDashboardRolesRequest, viewers_enabled: bool +) -> tuple[list[int] | None, list[int] | None, ManageDashboardRolesResponse | None]: + """Load current viewer roles and apply add/remove operations. + + Returns ``(current_role_ids, new_role_ids, None)`` on success or + ``(None, None, error_response)`` when the initial lazy-load fails or a + removal targets an unassigned role. The pre-call ``current_role_ids`` + is returned so the caller can report true added/removed deltas. + """ + try: + current_role_ids = _viewer_role_ids(dashboard) + except SQLAlchemyError as db_err: + logger.error( + "Failed to load roles for dashboard %s: %s", + request.identifier, + db_err, + exc_info=True, + ) + return ( + None, + None, + ManageDashboardRolesResponse( + viewers_enabled=viewers_enabled, + error="Failed to load dashboard roles due to a database error.", + ), + ) + + unknown_removals = sorted(set(request.remove_role_ids) - set(current_role_ids)) + if unknown_removals: + return ( + None, + None, + ManageDashboardRolesResponse( + viewers_enabled=viewers_enabled, + error=( + f"Cannot remove role IDs that are not currently assigned: " + f"{unknown_removals}. Current role IDs: " + f"{sorted(current_role_ids)}." + ), + ), + ) + + new_role_ids = [ + role_id + for role_id in current_role_ids + if role_id not in request.remove_role_ids + ] + for role_id in request.add_role_ids: + if role_id not in new_role_ids: + new_role_ids.append(role_id) + + return current_role_ids, new_role_ids, None + + +@tool( + tags=["mutate"], + class_permission_name="Dashboard", + method_permission_name="write", + annotations=ToolAnnotations( + title="Manage dashboard access roles", + readOnlyHint=False, + destructiveHint=True, + ), +) +def manage_dashboard_roles( + request: ManageDashboardRolesRequest, ctx: Context +) -> ManageDashboardRolesResponse: + """ + Add or remove dashboard access roles with explicit operations. + + Dashboard access roles restrict who can view a dashboard to members of + the listed roles, on top of normal Superset permissions. An empty roles + list means "no role restriction" — the dashboard is visible per standard + permissions instead. This only takes effect when the ``ENABLE_VIEWERS`` + feature flag is enabled; the response's ``viewers_enabled`` field + reports whether it is, and ``warnings`` notes when a change was applied + but has no live effect. + + Roles are the ROLE-type entries in the dashboard's Subject-based + ``viewers`` list. Any USER- or GROUP-type viewers already on the + dashboard are left untouched. + + Unlike ``update_dashboard``'s dropped ``roles`` field, this tool never + accepts a full-replacement list — only + ``add_role_ids``/``remove_role_ids``. + + Privacy: the returned ``roles`` list is sanctioned only as confirmation + of the add/remove operation the caller explicitly requested on this + dashboard. Do not use it to answer "who can access X" for a dashboard + the caller did not ask to modify, and do not call this tool merely to + look up current roles — those remain off-limits per the server + instructions. + + Example:: + + manage_dashboard_roles(request={ + "identifier": 42, + "add_role_ids": [5], + }) + """ + from superset import is_feature_enabled + + ctx.info( + f"Managing dashboard roles: identifier={request.identifier} " + f"add={request.add_role_ids} remove={request.remove_role_ids}" + ) + + dashboard, auth_error = _find_and_authorize_dashboard(request.identifier) + if auth_error is not None: + return auth_error + + viewers_enabled = is_feature_enabled("ENABLE_VIEWERS") + warnings: list[str] = [] + if not viewers_enabled: + warnings.append( + "The ENABLE_VIEWERS feature flag is disabled on this instance; " + "dashboard viewers will be stored but have no effect on access " + "control until it is enabled." + ) + + current_role_ids, new_role_ids, compute_error = _compute_new_role_ids( + dashboard, request, viewers_enabled + ) + if compute_error is not None: + return compute_error + assert current_role_ids is not None # narrows for mypy + assert new_role_ids is not None # narrows for mypy + + try: + with event_logger.log_context(action="mcp.manage_dashboard_roles.apply"): + from superset.subjects.utils import subjects_from_roles + + other_viewers = _other_viewers(dashboard) + resolved_role_subjects = subjects_from_roles(new_role_ids) + resolved_role_ids = {subject.role_id for subject in resolved_role_subjects} + missing_role_ids = sorted(set(new_role_ids) - resolved_role_ids) + if missing_role_ids: + return ManageDashboardRolesResponse( + viewers_enabled=viewers_enabled, + error=( + f"One or more role IDs do not exist: " + f"{missing_role_ids}. Use list_roles to resolve " + "valid role IDs." + ), + ) + + dashboard.viewers = other_viewers + resolved_role_subjects + db.session.commit() # pylint: disable=consider-using-transaction + try: + db.session.refresh(dashboard) + except SQLAlchemyError: + logger.warning( + "Dashboard %s roles updated but refresh failed; " + "continuing with current values", + dashboard.id, + exc_info=True, + ) + + except SQLAlchemyError as db_err: + try: + db.session.rollback() # pylint: disable=consider-using-transaction + except SQLAlchemyError: + logger.warning( + "Database rollback failed during error handling", + exc_info=True, + ) + logger.error("Dashboard roles update failed: %s", db_err, exc_info=True) + return ManageDashboardRolesResponse( + viewers_enabled=viewers_enabled, + error="Failed to update dashboard roles due to a database error.", + ) + + final_role_ids = set(_viewer_role_ids(dashboard)) + ctx.info(f"Dashboard {dashboard.id} roles updated: {sorted(final_role_ids)}") + + return ManageDashboardRolesResponse( + roles=[ + info + for subject in dashboard.viewers + if subject.type == SubjectType.ROLE + and (info := serialize_subject_object(subject)) is not None + ], Review Comment: **Suggestion:** The tool computes `final_role_ids` and builds the `roles` response from `dashboard.viewers` after commit, but these reads are outside the SQLAlchemy error handling block. If relationship loading fails (for example after a refresh failure or transient DB issue), the mutation has already been persisted yet the response path raises an unhandled exception. Wrap post-commit viewer reads in DB-error handling and return a structured response with warnings. [logic error] <details> <summary><b>Severity Level:</b> Major ⚠️</summary> ```mdx - ❌ MCP dashboard roles tool can crash post-commit. - ⚠️ Roles updated but response fails, confusing callers. ``` </details> <details> <summary><b>Steps of Reproduction ✅ </b></summary> ```mdx 1. Start the FastMCP server via `init_fastmcp_server()` in `superset/mcp_service/app.py:950`, which imports `superset/mcp_service/dashboard/tool/__init__.py` and registers the `manage_dashboard_roles()` tool implemented at `superset/mcp_service/dashboard/tool/manage_dashboard_roles.py:199-234`. 2. From an MCP client, invoke `manage_dashboard_roles()` with a valid dashboard identifier and non-empty `add_role_ids`/`remove_role_ids`; the request schema `ManageDashboardRolesRequest` at `superset/mcp_service/dashboard/schemas.py:2-57` validates the operations, and the handler computes `new_role_ids` via `_compute_new_role_ids()` at `manage_dashboard_roles.py:134-186`. 3. Inside the `event_logger.log_context` block at `manage_dashboard_roles.py:263-281`, the tool resolves subjects from role IDs, updates `dashboard.viewers`, and calls `db.session.commit()` (line 282) followed by `db.session.refresh(dashboard)` at lines 283-285; a transient database error during or after refresh leaves the commit successful but the ORM session or `viewers` relationship in an error/expired state, with the refresh failure only logged at `manage_dashboard_roles.py:286-291`. 4. After exiting the `try`/`except` block, `manage_dashboard_roles()` computes `final_role_ids = set(_viewer_role_ids(dashboard))` at `manage_dashboard_roles.py:307`, where `_viewer_role_ids()` reads `dashboard.viewers` at `manage_dashboard_roles.py:118-124`, and then builds the `roles` list comprehension from `dashboard.viewers` at `manage_dashboard_roles.py:311-316`; these post-commit relationship loads may trigger lazy database access on the broken session, raising `SQLAlchemyError` that is not caught, so the MCP tool call fails even though the dashboard's role assignments were already updated in the database. ``` </details> [](https://app.codeant.ai/fix-in-ide?tool=cursor&prompt_id=f6a33f3d853d4597bad637220a7e8912&service=github&base_url=https%3A%2F%2Fgithub.com&org=apache&repo=apache%2Fsuperset) [](https://app.codeant.ai/fix-in-ide?tool=vscode-claude&prompt_id=f6a33f3d853d4597bad637220a7e8912&service=github&base_url=https%3A%2F%2Fgithub.com&org=apache&repo=apache%2Fsuperset) *(Use Cmd/Ctrl + Click for best experience)* <details> <summary><b>Prompt for AI Agent 🤖 </b></summary> ```mdx This is a comment left during a code review. **Path:** superset/mcp_service/dashboard/tool/manage_dashboard_roles.py **Line:** 307:316 **Comment:** *Logic Error: The tool computes `final_role_ids` and builds the `roles` response from `dashboard.viewers` after commit, but these reads are outside the SQLAlchemy error handling block. If relationship loading fails (for example after a refresh failure or transient DB issue), the mutation has already been persisted yet the response path raises an unhandled exception. Wrap post-commit viewer reads in DB-error handling and return a structured response with warnings. Validate the correctness of the flagged issue. If correct, How can I resolve this? If you propose a fix, implement it and please make it concise. Once fix is implemented, also check other comments on the same PR, and ask user if the user wants to fix the rest of the comments as well. if said yes, then fetch all the comments validate the correctness and implement a minimal fix ``` </details> <a href='https://app.codeant.ai/feedback?pr_url=https%3A%2F%2Fgithub.com%2Fapache%2Fsuperset%2Fpull%2F41606&comment_hash=ac704291187dc853343719f89198e0d14e452161a3b277131c46cb0c0b942697&reaction=like'>👍</a> | <a href='https://app.codeant.ai/feedback?pr_url=https%3A%2F%2Fgithub.com%2Fapache%2Fsuperset%2Fpull%2F41606&comment_hash=ac704291187dc853343719f89198e0d14e452161a3b277131c46cb0c0b942697&reaction=dislike'>👎</a> ########## superset/mcp_service/dashboard/tool/manage_dashboard_certification.py: ########## @@ -0,0 +1,206 @@ +# Licensed to the Apache Software Foundation (ASF) under one +# or more contributor license agreements. See the NOTICE file +# distributed with this work for additional information +# regarding copyright ownership. The ASF licenses this file +# to you under the Apache License, Version 2.0 (the +# "License"); you may not use this file except in compliance +# with the License. You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, +# software distributed under the License is distributed on an +# "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY +# KIND, either express or implied. See the License for the +# specific language governing permissions and limitations +# under the License. + +""" +Manage dashboard certification FastMCP tool + +Sets or clears the ``certified_by`` / ``certification_details`` badge +fields. Split out from the generic ``update_dashboard`` tool because +certification is a distinct governance concern from layout/theme/metadata +edits. +""" + +import logging +from typing import Any + +from fastmcp import Context +from sqlalchemy.exc import SQLAlchemyError +from superset_core.mcp.decorators import tool, ToolAnnotations + +from superset.commands.dashboard.exceptions import ( + DashboardAccessDeniedError, + DashboardNotFoundError, +) +from superset.exceptions import SupersetSecurityException +from superset.extensions import db, event_logger +from superset.mcp_service.dashboard.schemas import ( + ManageDashboardCertificationRequest, + ManageDashboardCertificationResponse, +) +from superset.mcp_service.utils.url_utils import get_superset_base_url + +logger: logging.Logger = logging.getLogger(__name__) + + +def _find_and_authorize_dashboard( + identifier: int | str, +) -> tuple[Any, ManageDashboardCertificationResponse | None]: + """Return (dashboard, None) on success or (None, error_response) on failure. + + Mirrors the helper in ``update_dashboard``: avoids ImportError before + Flask app initialisation by co-locating the imports it needs with the + call site rather than importing them at module load time. + """ + from superset import security_manager + from superset.daos.dashboard import DashboardDAO + + try: + dashboard = DashboardDAO.get_by_id_or_slug(identifier) + except DashboardAccessDeniedError: + # get_by_id_or_slug re-checks view access and raises access-denied + # for dashboards the caller cannot see; surface it as the + # structured permission_denied response instead of an unhandled + # error. + return None, ManageDashboardCertificationResponse( + permission_denied=True, + error=( + "You do not have permission to access this dashboard. " + "Ask the user to grant access; do not retry." + ), + ) + except DashboardNotFoundError: + return None, ManageDashboardCertificationResponse( + error=f"Dashboard not found: {identifier!r}", + ) + except SQLAlchemyError: + logger.exception("Database error looking up dashboard %r", identifier) + return None, ManageDashboardCertificationResponse( + error="Failed to look up dashboard due to a database error.", + ) + + if dashboard is None: + return None, ManageDashboardCertificationResponse( + error=f"Dashboard not found: {identifier!r}", + ) + + try: + security_manager.raise_for_editorship(dashboard) + except SupersetSecurityException: + return None, ManageDashboardCertificationResponse( + permission_denied=True, + error=( + f"You don't have permission to edit dashboard " + f"'{dashboard.dashboard_title}' (ID: {dashboard.id})." + ), + ) + + return dashboard, None + + +def _dashboard_url(dashboard: Any) -> str: + """Build the user-facing dashboard URL, preferring slug over id.""" + return f"{get_superset_base_url()}/dashboard/{dashboard.slug or dashboard.id}/" + + +@tool( + tags=["mutate"], + class_permission_name="Dashboard", + method_permission_name="write", + annotations=ToolAnnotations( + title="Manage dashboard certification", + readOnlyHint=False, + destructiveHint=False, + ), +) +def manage_dashboard_certification( + request: ManageDashboardCertificationRequest, ctx: Context +) -> ManageDashboardCertificationResponse: + """ + Set or clear a dashboard's certification badge. + + ``certified_by`` and ``certification_details`` are independent optional + fields: omit (None) to leave a field unchanged, pass an empty string to + clear it, or pass a value to set it. Certification surfaces as a badge + next to the dashboard title in the UI. + + Example:: + + manage_dashboard_certification(request={ + "identifier": 42, + "certified_by": "Data Platform Team", + "certification_details": "Verified against source-of-truth metrics.", + }) + """ + ctx.info(f"Managing dashboard certification: identifier={request.identifier}") + + dashboard, auth_error = _find_and_authorize_dashboard(request.identifier) + if auth_error is not None: + return auth_error + + if request.certified_by is None and request.certification_details is None: + return ManageDashboardCertificationResponse( + certified_by=dashboard.certified_by, + certification_details=dashboard.certification_details, + dashboard_url=_dashboard_url(dashboard), + changed_fields=[], + warnings=["No fields provided; dashboard unchanged."], + ) + + changed_fields: list[str] = [] + warnings: list[str] = [] + + try: + with event_logger.log_context( + action="mcp.manage_dashboard_certification.apply" + ): + if request.certified_by is not None: + dashboard.certified_by = request.certified_by or None + changed_fields.append("certified_by") + + if request.certification_details is not None: + dashboard.certification_details = request.certification_details or None + changed_fields.append("certification_details") + + db.session.commit() # pylint: disable=consider-using-transaction + try: + db.session.refresh(dashboard) + except SQLAlchemyError: + logger.warning( + "Dashboard %s certification updated but refresh failed; " + "continuing with current values", + dashboard.id, + exc_info=True, + ) + warnings.append( + "Dashboard updated but post-update refresh failed; " + "returned values may not reflect database state." + ) + + except SQLAlchemyError as db_err: + try: + db.session.rollback() # pylint: disable=consider-using-transaction + except SQLAlchemyError: + logger.warning( + "Database rollback failed during error handling", + exc_info=True, + ) + logger.error("Dashboard certification update failed: %s", db_err, exc_info=True) + return ManageDashboardCertificationResponse( + error="Failed to update dashboard certification due to a database error.", + ) + + ctx.info( + f"Dashboard {dashboard.id} certification updated: changed={changed_fields}" + ) + + return ManageDashboardCertificationResponse( + certified_by=dashboard.certified_by, + certification_details=dashboard.certification_details, + dashboard_url=_dashboard_url(dashboard), + changed_fields=changed_fields, + warnings=warnings, Review Comment: **Suggestion:** After a successful commit, the response still reads `dashboard` attributes outside the database error handler. If `refresh` fails (or the session expires attributes on commit), accessing these fields can raise a new `SQLAlchemyError`, turning a successful write into an unhandled tool failure. Guard post-commit field reads with DB-error handling and return a structured response/warning path instead of directly dereferencing ORM attributes. [logic error] <details> <summary><b>Severity Level:</b> Major ⚠️</summary> ```mdx - ❌ MCP dashboard certification tool may error after commit. - ⚠️ Clients see tool failure, certification state still mutated. ``` </details> <details> <summary><b>Steps of Reproduction ✅ </b></summary> ```mdx 1. Start the FastMCP server via `init_fastmcp_server()` in `superset/mcp_service/app.py:950`, which registers dashboard tools from `superset/mcp_service/dashboard/tool/__init__.py`, including `manage_dashboard_certification()` defined at `superset/mcp_service/dashboard/tool/manage_dashboard_certification.py:119-137`. 2. From an MCP client, call `manage_dashboard_certification()` with a valid dashboard identifier and non-None certification fields; the request schema `ManageDashboardCertificationRequest` lives at `superset/mcp_service/dashboard/schemas.py:117-148`, and the handler applies changes and calls `db.session.commit()` inside the `event_logger.log_context` block at `manage_dashboard_certification.py:156-168`. 3. During or immediately after the post-commit `db.session.refresh(dashboard)` at `manage_dashboard_certification.py:169-171`, a transient database/connection issue occurs, leaving the commit successful but the ORM session in an error or expired state; the refresh failure is logged and a warning appended at `manage_dashboard_certification.py:172-181` without rolling back or wrapping subsequent attribute reads. 4. Execution continues past the `try`/`except` block to the final response construction at `manage_dashboard_certification.py:200-205`, which dereferences `dashboard.certified_by` and `dashboard.certification_details`; on an expired or erroring session these attribute accesses may trigger a lazy load that raises `SQLAlchemyError` outside any handler, causing the MCP tool invocation to fail even though the dashboard certification changes were successfully persisted. ``` </details> [](https://app.codeant.ai/fix-in-ide?tool=cursor&prompt_id=4a81760e48334106ae3beaf90208741b&service=github&base_url=https%3A%2F%2Fgithub.com&org=apache&repo=apache%2Fsuperset) [](https://app.codeant.ai/fix-in-ide?tool=vscode-claude&prompt_id=4a81760e48334106ae3beaf90208741b&service=github&base_url=https%3A%2F%2Fgithub.com&org=apache&repo=apache%2Fsuperset) *(Use Cmd/Ctrl + Click for best experience)* <details> <summary><b>Prompt for AI Agent 🤖 </b></summary> ```mdx This is a comment left during a code review. **Path:** superset/mcp_service/dashboard/tool/manage_dashboard_certification.py **Line:** 200:205 **Comment:** *Logic Error: After a successful commit, the response still reads `dashboard` attributes outside the database error handler. If `refresh` fails (or the session expires attributes on commit), accessing these fields can raise a new `SQLAlchemyError`, turning a successful write into an unhandled tool failure. Guard post-commit field reads with DB-error handling and return a structured response/warning path instead of directly dereferencing ORM attributes. Validate the correctness of the flagged issue. If correct, How can I resolve this? If you propose a fix, implement it and please make it concise. Once fix is implemented, also check other comments on the same PR, and ask user if the user wants to fix the rest of the comments as well. if said yes, then fetch all the comments validate the correctness and implement a minimal fix ``` </details> <a href='https://app.codeant.ai/feedback?pr_url=https%3A%2F%2Fgithub.com%2Fapache%2Fsuperset%2Fpull%2F41606&comment_hash=2099124990b6dc7d994b783822ec57814234080dd8e999c0182a6d20ade9728e&reaction=like'>👍</a> | <a href='https://app.codeant.ai/feedback?pr_url=https%3A%2F%2Fgithub.com%2Fapache%2Fsuperset%2Fpull%2F41606&comment_hash=2099124990b6dc7d994b783822ec57814234080dd8e999c0182a6d20ade9728e&reaction=dislike'>👎</a> -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
