[sadpandajoe]

SUMMARY

tests/unit_tests/mcp_service/test_native_tool_surface.py::test_native_inventory_report_matches_registered_tools has been failing on master since #44558 ("feat(mcp): add caller-scoped PNG chart previews", 1e18d01), which turns the required unit-tests-required check red.

Root cause: #44558 added a PNGPreview variant to ChartPreviewContent and a png format to GetChartPreviewRequest in superset/mcp_service/chart/schemas.py. The output schemas of generate_chart, update_chart and get_chart_preview each grew by 496 bytes, which exceeds the inventory's headroom of 2% or 200 bytes. The PR did not regenerate tests/unit_tests/mcp_service/native_tool_inventory.json. Its CI could not catch that: the last CI run on the PR (head 1478574) used base cb8ff28 from 2026-10-09, which came before the native tool inventory and its test were added (#45081 / #45084, 2026-10-10). So the PR was green on a stale base and merged without being checked against the inventory. The earlier master commit #44657 also touched chart/schemas.py, but regenerating the inventory at its head gives a file identical to the recorded one, so all of the growth comes from #44558.

The growth is intended: it is the new PNG preview variant. As the test's docstring describes for intentional changes, this PR regenerates the recorded inventory with SUPERSET_MCP_UPDATE_TOOL_INVENTORY=1. Only generate_chart, update_chart, get_chart_preview and the catalog totals change. No other tool's entry changes, and no code or test logic changes.

TESTING INSTRUCTIONS

1. Check out this branch and install the backend dev requirements.
2. Run pytest tests/unit_tests/mcp_service/test_native_tool_surface.py. Every test passes, including test_native_inventory_report_matches_registered_tools, which fails on current master.
3. Optionally, run SUPERSET_MCP_UPDATE_TOOL_INVENTORY=1 pytest tests/unit_tests/mcp_service/test_native_tool_surface.py::test_native_inventory_report_matches_registered_tools and confirm git diff is empty. The committed inventory is exactly what the generator produces.

ADDITIONAL INFORMATION

• Has associated issue: Fixes test_native_inventory_report_matches_registered_tools fails on master after PNG chart previews #45510
• Required feature flags:
• Changes UI
• Includes DB Migration (follow approval process in SIP-59)
  • Migration is atomic, supports rollback & is backwards-compatible
  • Confirm DB migration upgrade and downgrade tested
  • Runtime estimates and downtime expectations provided
• Introduces new feature or API
• Removes existing feature or API

🤖 Generated with Claude Code

[codeant-ai-for-open-source]

CodeAnt PR Risk: Low Risk

• The PR appears safe to merge; the visible changes update tool-inventory size metrics and show no functional code changes.

Assessed commit: b88192cf89b0

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 83.64%. Comparing base (1e18d01) to head (b88192c).

Additional details and impacted files

@@             Coverage Diff             @@
##           master   #45511       +/-   ##
===========================================
+ Coverage   65.15%   83.64%   +18.49%     
===========================================
  Files        3053     3053               
  Lines      200638   200638               
  Branches    46820    46820               
===========================================
+ Hits       130717   167827    +37110     
+ Misses      66981    29566    -37415     
- Partials     2940     3245      +305     

Flag	Coverage Δ
hive	34.91% <ø> (ø)
mysql	52.87% <ø> (ø)
postgres	52.88% <ø> (-0.01%)	⬇️
presto	36.66% <ø> (ø)
python	87.16% <ø> (+34.00%)	⬆️
sqlite	52.64% <ø> (ø)
unit	81.30% <ø> (?)

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.

[rusackas]

LGTM, exactly the mechanical regeneration described, only the three touched tools' entries and the catalog totals changed. Good catch on why #44558's own CI missed this (stale base predating the inventory test).