HC Bubble Charts just give a blurb of plotly configuration info rather than the plots themselves

[operdeck]

pdstools version checks

• I have checked that this issue has not already been reported.

• I have confirmed this bug exists on the latest version of pdstools.

Issue description

For some customers, presumably when there are a lot of configurations, the Model Performance vs Action Success Rates (the Bubble Charts) chart does not give plots but gives a list of configurations.

Reproducible example

See HC checks

Expected behavior

This output is not useful. Expected an error message when we really can't handle it but rather just the actual plots. Perhaps truncated (but then with a warning/error and a hint to subset the data) to a few that probably make most sense.

Installed versions

Credits
Document created at: 29 Apr 2025 11:16:36

This notebook: pega-datascientist-tools/python/pdstools/reports/HealthCheck.qmd

Quarto runtime: 1.5.45
Pandoc: 3.6.4

Additional details from 'pdstools.show_versions()':
--- Version info ---
pdstools: 4.1.9
Platform: macOS-15.4-arm64-arm-64bit
Python: 3.12.3 (v3.12.3:f6650f9ad7, Apr 9 2024, 08:18:47) [Clang 13.0.0 (clang-1300.0.29.30)]

--- Dependencies ---
polars>=0.20: 1.27.1
typing_extensions: 4.13.2
onnxruntime==1.18.1; python_full_version < "3.10" and extra == "onnx": 1.20.1
onnxruntime==1.20.1; python_full_version >= "3.10" and extra == "onnx": 1.20.1

--- Dependency group: adm ---
plotly[express]>=6.0:
requests: 2.32.3

--- Dependency group: pega-io ---
polars_hash: 0.5.3
aioboto3: 14.1.0

--- Dependency group: api ---
httpx: 0.28.1
pydantic: 2.11.3
anyio: 4.9.0

--- Dependency group: healthcheck ---
requests: 2.32.3
great_tables>=0.13: 0.17.0
plotly[express]>=6.0:
papermill: 2.6.0
xlsxwriter>=3.0: 3.2.3
quarto: 0.1.0
pydot: 3.0.4

--- Dependency group: app ---
great_tables>=0.13: 0.17.0
papermill: 2.6.0
xlsxwriter>=3.0: 3.2.3
st-pages:
streamlit>=1.23: 1.44.1
plotly[express]>=6.0:
requests: 2.32.3
quarto: 0.1.0
pydot: 3.0.4

--- Dependency group: onnx ---
onnx==1.16.1: 1.16.1
scikit-learn==1.5.1:
skl2onnx==1.17.0: 1.17.0
httpx: 0.28.1
anyio: 4.9.0
pydantic: 2.11.3

--- Dependency group: all ---
great_tables>=0.13: 0.17.0
scikit-learn==1.5.1:
papermill: 2.6.0
httpx: 0.28.1
xlsxwriter>=3.0: 3.2.3
st-pages:
pydantic: 2.11.3
streamlit>=1.23: 1.44.1
plotly[express]>=6.0:
onnx==1.16.1: 1.16.1
skl2onnx==1.17.0: 1.17.0
requests: 2.32.3
anyio: 4.9.0
quarto: 0.1.0
pydot: 3.0.4

--- Dependency group: dev ---
pre-commit:

--- Dependency group: docs ---
myst-parser:
nbsphinx:
sphinx-autoapi:
furo:
sphinx:
sphinx-copybutton:
sphinx-argparse:

--- Dependency group: tests ---
great_tables>=0.13: 0.17.0
xlsxwriter>=3.0: 3.2.3
st-pages:
pytest-mock:
pydantic: 2.11.3
skl2onnx==1.17.0: 1.17.0
pytest:
requests: 2.32.3
moto:
scikit-learn==1.5.1:
papermill: 2.6.0
coverage:
httpx: 0.28.1
openpyxl: 3.1.5
streamlit>=1.23: 1.44.1
plotly[express]>=6.0:
testbook:
onnx==1.16.1: 1.16.1
pytest-httpx:
anyio: 4.9.0
pytest-cov:
quarto: 0.1.0
pydot: 3.0.4

[StijnKas]

Still a thing. We should consider explicitly failing here, and instructing the user to re-run the HC with filters applied.