docs: Remove tempfiles from docs

[MarieS-WiMLDS]

fixes #1178

[github-actions]

Coverage Report for backend

File	Stmts	Miss	Cover	Missing
venv/lib/python3.12/site-packages/skore
__init__.py	14	0	100%
__main__.py	8	1	80%	19
exceptions.py	3	0	100%
venv/lib/python3.12/site-packages/skore/cli
__init__.py	5	0	100%
cli.py	33	3	85%	104, 111, 117
color_format.py	43	3	90%	35–>40, 41–43
launch_dashboard.py	26	15	39%	36–57
quickstart_command.py	14	7	50%	37–51
venv/lib/python3.12/site-packages/skore/persistence
__init__.py	0	0	100%
venv/lib/python3.12/site-packages/skore/persistence/item
__init__.py	59	4	91%	89, 100–103
altair_chart_item.py	24	1	93%	14
cross_validation_reporter_item.py	105	12	88%	28–41, 125–126, 254, 257
item.py	24	1	96%	86
matplotlib_figure_item.py	30	1	94%	18
media_item.py	24	0	100%
numpy_array_item.py	22	1	92%	14
pandas_dataframe_item.py	31	1	94%	14
pandas_series_item.py	31	1	94%	14
pickle_item.py	29	0	100%
pillow_image_item.py	29	1	94%	14
plotly_figure_item.py	25	1	93%	14
polars_dataframe_item.py	29	1	94%	14
polars_series_item.py	24	1	93%	14
primitive_item.py	25	2	91%	13–15
sklearn_base_estimator_item.py	28	1	94%	14
skrub_table_report_item.py	10	1	86%	11
venv/lib/python3.12/site-packages/skore/persistence/repository
__init__.py	3	0	100%
item_repository.py	59	5	91%	15–16, 202–203, 226
view_repository.py	19	2	86%	9–10
venv/lib/python3.12/site-packages/skore/persistence/storage
__init__.py	4	0	100%
abstract_storage.py	22	0	100%
disk_cache_storage.py	33	1	95%	44
in_memory_storage.py	20	0	100%
venv/lib/python3.12/site-packages/skore/persistence/view
__init__.py	0	0	100%
view.py	5	0	100%
venv/lib/python3.12/site-packages/skore/project
__init__.py	3	0	100%
create.py	52	8	88%	116–122, 132–133, 140–141
load.py	22	3	88%	45–47
open.py	14	0	100%
project.py	50	2	95%	13, 172
venv/lib/python3.12/site-packages/skore/sklearn
__init__.py	5	0	100%
_base.py	140	4	97%	91, 94, 168–>173, 183–184
find_ml_task.py	45	1	95%	71–>87, 80–>87, 86
types.py	2	0	100%
venv/lib/python3.12/site-packages/skore/sklearn/_cross_validation
__init__.py	6	0	100%
metrics_accessor.py	163	0	100%
report.py	87	0	100%
venv/lib/python3.12/site-packages/skore/sklearn/_estimator
__init__.py	6	0	100%
metrics_accessor.py	265	4	97%	149–158, 183–>236, 191, 434–>437, 783–>786
report.py	120	0	99%	215–>221, 223–>225
utils.py	11	11	0%	1–19
venv/lib/python3.12/site-packages/skore/sklearn/_plot
__init__.py	4	0	100%
precision_recall_curve.py	121	1	98%	229–>246, 317
prediction_error.py	97	0	100%
roc_curve.py	128	0	100%
utils.py	88	0	100%
venv/lib/python3.12/site-packages/skore/sklearn/cross_validation
__init__.py	2	0	100%
cross_validation_helpers.py	47	4	90%	104–>136, 123–126
cross_validation_reporter.py	35	1	95%	187
venv/lib/python3.12/site-packages/skore/sklearn/cross_validation/plots
__init__.py	0	0	100%
compare_scores_plot.py	29	24	16%	10, 28–116
timing_plot.py	29	24	17%	10, 26–123
venv/lib/python3.12/site-packages/skore/sklearn/train_test_split
__init__.py	0	0	100%
train_test_split.py	36	2	94%	16–17
venv/lib/python3.12/site-packages/skore/sklearn/train_test_split/warning
__init__.py	8	0	100%
high_class_imbalance_too_few_examples_warning.py	17	3	78%	16–18, 80
high_class_imbalance_warning.py	18	2	88%	16–18
random_state_unset_warning.py	11	1	87%	15
shuffle_true_warning.py	9	0	91%	44–>exit
stratify_is_set_warning.py	11	1	87%	15
time_based_column_warning.py	22	1	89%	17, 69–>exit
train_test_split_warning.py	5	1	80%	21
venv/lib/python3.12/site-packages/skore/ui
__init__.py	0	0	100%
app.py	32	8	68%	26, 55–60, 72–74
dependencies.py	7	1	86%	12
project_routes.py	61	0	100%
venv/lib/python3.12/site-packages/skore/utils
__init__.py	0	0	100%
_accessor.py	7	0	100%
_logger.py	21	4	84%	14–18
_patch.py	11	5	46%	19–35
_progress_bar.py	30	0	100%
_show_versions.py	31	0	100%
TOTAL	2768	183	92%

Tests	Skipped	Failures	Errors	Time
612	3 💤	0 ❌	0 🔥	2m 52s ⏱️

[auguste-probabl]

I don't love this solution because it makes all the examples dependent on one-another. If example 1 fails halfway through, then clear won't be run and example 2 will start with a non-empty project, which could make it fail, etc.

[auguste-probabl]

But the clear method is fine IMO

[MarieS-WiMLDS]

I don't love this solution because it makes all the examples dependent on one-another. If example 1 fails halfway through, then clear won't be run and example 2 will start with a non-empty project, which could make it fail, etc.

Is this likely to happen? Or it will trigger a red CI and therefore shouldn't land in main? (unless we force it)

[auguste-probabl]

I don't love this solution because it makes all the examples dependent on one-another. If example 1 fails halfway through, then clear won't be run and example 2 will start with a non-empty project, which could make it fail, etc.

Is this likely to happen? Or it will trigger a red CI and therefore shouldn't land in main? (unless we force it)

Yes, tests fail all the time, and we would prefer that 1 example failed rather than all of them; that way it's easier to understand what went wrong and fix things

Another thing I thought of later is that this problem prevents us from running the examples in parallel

[MarieS-WiMLDS]

I can't find any alternative solution for now to answer to all our goals. I created an issue to formalize this: #1200. In the mean time, if it's ok for you, I'd like to proceed with this solution to remove tempfile for this afternoon.

[github-actions]

Documentation preview @ dc28842

[auguste-probabl]

If example 1 fails halfway through, then clear won't be run and example 2 will start with a non-empty project, which could make it fail, etc.

To mitigate this could you follow https://sphinx-gallery.github.io/stable/configuration.html#abort-build-on-first-fail? This way if an example fails, we can see quicker where the problem comes from (rather than 10 other examples failing just because the first one failed)

[sylvaincom]

Could we also apply this to the quick start example? 🙏

[sylvaincom]

Many thanks!

[auguste-probabl]

Nice!

2 approvals already.