fix(report): write HTML export as utf-8

[immu4989]

Description

export_and_save_report() writes the rendered HTML report via
open(output_file, "w") without an explicit encoding=, so the
file is written with locale.getpreferredencoding(False). On Windows
that's cp1252, which raises
UnicodeEncodeError: 'charmap' codec can't encode character ...
whenever the nbconvert HTML body contains non-Latin1 characters
(feature names with accents, user-supplied text in additional data,
dataset string values, etc.).

The reporter (@dalestee) diagnosed the issue and prescribed the fix in
the issue body. This PR applies that fix exactly: add
encoding="utf-8" to the open() call at
shapash/report/generation.py:93.

Grepped the rest of shapash/ for sibling sites — every other
write-side open() call is in binary mode and every read-side one
touches ASCII-only bundled JSON, so this is the only line that needs
to change.

Fixes #699

Type of change

• Bug fix (non-breaking change which fixes an issue)
• New feature (non-breaking change which adds functionality or feature that would cause existing functionality to not work as expected)
• This change requires a documentation update

How Has This Been Tested?

Added test_export_and_save_report_writes_utf8 in
tests/integration_tests/test_report_generation.py. It mocks
HTMLExporter and the local open binding in
shapash.report.generation, calls export_and_save_report, and
asserts that every write-mode open() call carries
encoding="utf-8".

The test uses mock.patch rather than monkeypatching locale.getpreferredencoding,
because Python 3.12 resolves open()'s default encoding at a lower
level than the locale module exposes — monkeypatching locale no
longer changes the actual behavior. A contract test on the call
signature is the most reliable way to catch a future regression on
CI runners that default to UTF-8.

• pytest tests/integration_tests/test_report_generation.py -q → 8 passed (7 pre-existing + 1 new)
• ruff format --check shapash/report/generation.py tests/integration_tests/test_report_generation.py → clean

Test Configuration:

• OS: macOS
• Python version: 3.12.13
• Shapash version: 2.8.1 (editable install of this branch)

Checklist:

• My code follows the style guidelines of this project
• I have performed a self-review of my own code
• I have commented my code, particularly in hard to understand areas — only a docstring on the new test linking the bug; the fix itself is one line
• I have made corresponding changes to the documentation — N/A, no user-facing API change
• My changes generate no new warnings
• I have added tests that prove my fix is effective or that my feature works
• New and existing unit tests pass locally with my changes
• Any dependent changes have been merged and published in downstream modules — N/A, no dependent changes

[guerinclement]

Hi @immu4989,
Thanks again for your contribution and your interest in shapash.

We have a wider ongoing work on the report generation to get rid of papermill/jupyter dependencies.
The fix for issue #699 will be shipped along the release of the feature.

I let your PR opened in the meantime but it will likely not be merged.

[immu4989]

Thanks @guerinclement, that makes total sense , and good to hear #699 will be addressed as part of the wider rework.

Happy to leave the PR open as a one line stopgap in case a Windows user reports the bug again before the new report module lands. Feel free to cherry pick the encoding="utf-8" change (or close this out) at any point , whichever is cleaner for you.

Looking forward to seeing the refactor land.