Fix dephell and CI hell #3

ddelange · 2024-07-19T20:07:29Z

fix some dephell and related CI hell by getting rid of conda and using install_requires and extras_require exclusively

consolidate CI requirements in top-level setup.py
fix build error due to naming confusion vaex vs vaex-meta
- top-level setup.py package name was/is vaex-meta
- packages/vaex-meta package name was/is vaex
- renamed packages/vaex-meta to packages/vaex
  - this setup.py is the one uploaded as https://pypi.org/project/vaex
  - top-level setup.py, vaex-meta, does not exist on pypi as only intended for local development)
- this avoids a build error when requiring a local uri in install_requires (vaex @ file://.../vaex-meta is not allowed)
unify requirements: move all non-CI requirements to the relevant setup.py, either to install_requires or to [all] extra
use setyp-python in CI
use uv pip install .[ci] in CI
- pip install . now also works
- pip install https://github.com/vaexio/vaex/archive/refs/heads/master.zip now also works
remove hacky pip-inside-pip code from top-level setup.py, replace with file:// specifiers
remove misc/conda folder containing outdated dependency derivatives (potential source of confusion)
remove conda* env files, as all requirements and their versiom ranges are now (mutually) exclusively found in setup.py files

ddelange · 2024-07-25T11:13:38Z

cc @maartenbreddels I think we're almost there, only a few things left:

new pandas brings some issues:

    def test_df_project():
        x = np.arange(10, dtype='i4')
        y = x**2
        df = vaex.from_arrays(x=x, y=y, z1=x+y, z2=x-y)
        # projecting 2 columns will drop 2 columns, which could be done in different order
        df_a = df[['x', 'y']]
        df_b = df[['x', 'y']]
        assert df_a.fingerprint() == df_b.fingerprint()
>       assert df_a.fingerprint() == 'dataframe-c13a4ab588272f03855ae5627731f7e5'
E       AssertionError: assert 'dataframe-d4...9ff888ba16e7c' == 'dataframe-c1...ae5627731f7e5'
E         
E         - dataframe-c13a4ab588272f03855ae5627731f7e5
E         + dataframe-d4565ca8187231a051a9ff888ba16e7c

and

    @pytest.mark.skipif(sys.version_info[:2] < (3, 7), reason="Python 36+Pandas has no Float64Dtype")
    def test_from_pandas():
        dd_dict = {
            'boolean': [True, True, False, None, True],
            'text': ['This', 'is', 'some', 'text', 'so...'],
            'text_missing': pd.Series(['Some', 'parts', None, 'missing', None], dtype='string'),
            'float': [1, 30, -2, 1.5, 0.000],
            'float_missing': [1, None, -2, 1.5, 0.000],
            'float_missing_masked': pd.Series([1, None, -2, 1.5, 0.000], dtype=pd.Float64Dtype()),
            'int_missing': pd.Series([1, None, 5, 1, 10], dtype='Int64'),
            'datetime_1': [pd.NaT, datetime.datetime(2019, 1, 1, 1, 1, 1), datetime.datetime(2019, 1, 1, 1, 1, 1), datetime.datetime(2019, 1, 1, 1, 1, 1), datetime.datetime(2019, 1, 1, 1, 1, 1)],
            'datetime_2': [pd.NaT, None, pd.NaT, pd.NaT, pd.NaT],
>           'datetime_3': [pd.Timedelta('1M'), pd.Timedelta('1D'), pd.Timedelta('100M'), pd.Timedelta('2D'), pd.Timedelta('1H')],
            'datetime_4': [pd.Timestamp('2001-1-1 2:2:11'), pd.Timestamp('2001-12'), pd.Timestamp('2001-10-1'), pd.Timestamp('2001-03-1 2:2:11'), pd.Timestamp('2001-1-1 2:2:11')],
            'datetime_5': [datetime.date(2010, 1, 1), datetime.date(2010, 1, 1), datetime.date(2010, 1, 1), datetime.date(2010, 1, 1), datetime.date(2010, 1, 1)],
            'datetime_6': [datetime.time(21, 1, 1), datetime.time(21, 1, 1), datetime.time(21, 1, 1), datetime.time(21, 1, 1), datetime.time(21, 1, 1)],
        }

tests/from_pandas_test.py:23: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
timedeltas.pyx:1844: in pandas._libs.tslibs.timedeltas.Timedelta.__new__
    ???
timedeltas.pyx:644: in pandas._libs.tslibs.timedeltas.parse_timedelta_string
    ???
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

>   ???
E   ValueError: Units 'M', 'Y' and 'y' do not represent unambiguous timedelta values and are not supported.

and

x = array(['2009-10-12T03:31:00.000000000'], dtype='datetime64[ns]')

    @register_function(scope='dt', as_property=True)
    def dt_weekofyear(x):
        """Returns the week ordinal of the year.
    
        :returns: an expression containing the week ordinal of the year, extracted from a datetime column.
    
        Example:
    
        >>> import vaex
        >>> import numpy as np
        >>> date = np.array(['2009-10-12T03:31:00', '2016-02-11T10:17:34', '2015-11-12T11:34:22'], dtype=np.datetime64)
        >>> df = vaex.from_arrays(date=date)
        >>> df
          #  date
          0  2009-10-12 03:31:00
          1  2016-02-11 10:17:34
          2  2015-11-12 11:34:22
    
        >>> df.date.dt.weekofyear
        Expression = dt_weekofyear(date)
        Length: 3 dtype: int64 (expression)
        -----------------------------------
        0  42
        1   6
        2  46
        """
>       return _to_pandas_series(x).dt.weekofyear.values
E       AttributeError: 'DatetimeProperties' object has no attribute 'weekofyear'

and it looks like new graphql version (needed for more recent python versions) has an issue with the query syntax:

    def test_aggregates(df, schema):
        result = schema.execute("""
        {
            df {
                count
                min {
                    x
                    y
                }
                mean {
                    x
                    y
                }
                max {
                    x
                    y
                }
            }
        }
        """)
>       assert not result.errors
E       assert not [GraphQLError("Aggregation.__init__() got an unexpected keyword argument 'df'", locations=[SourceLocation(line=3, column=9)], path=['df'])]
E        +  where [GraphQLError("Aggregation.__init__() got an unexpected keyword argument 'df'", locations=[SourceLocation(line=3, column=9)], path=['df'])] = ExecutionResult(data={'df': None}, errors=[GraphQLError("Aggregation.__init__() got an unexpected keyword argument 'df'", locations=[SourceLocation(line=3, column=9)], path=['df'])]).errors

also something wrong with csr:

    def test_sparse_basics():
        df = vaex.from_arrays(x=x)
        df.add_columns(['s1', 's2'], s)
>       assert df.s1.tolist() == [0, 1, 2]

tests/sparse_test.py:12: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
/opt/hostedtoolcache/Python/3.12.4/x64/lib/python3.12/site-packages/vaex/expression.py:808: in tolist
    values = self.evaluate(i1=i1, i2=i2)
/opt/hostedtoolcache/Python/3.12.4/x64/lib/python3.12/site-packages/vaex/expression.py:1086: in evaluate
    return self.ds.evaluate(self, i1, i2, out=out, selection=selection, array_type=array_type, parallel=parallel)
/opt/hostedtoolcache/Python/3.12.4/x64/lib/python3.12/site-packages/vaex/dataframe.py:3095: in evaluate
    return self._evaluate_implementation(expression, i1=i1, i2=i2, out=out, selection=selection, filtered=filtered, array_type=array_type, parallel=parallel, chunk_size=chunk_size, progress=progress)
/opt/hostedtoolcache/Python/3.12.4/x64/lib/python3.12/site-packages/vaex/dataframe.py:6481: in _evaluate_implementation
    arrays[expression] = arrays[expression][0:end-start]  # materialize fancy columns (lazy, indexed)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

self = <vaex.column.ColumnSparse object at 0x7efd744a73b0>
slice = slice(0, 3, None)

    def __getitem__(self, slice):
        # not sure if this is the fastest
>       return self.matrix[slice, self.column_index].A[:,0]
E       AttributeError: 'csr_matrix' object has no attribute 'A'

also tensorflow wants an extension now:

______________ test_keras_model_classification[df_factory_numpy] _______________

tmpdir = local('/tmp/pytest-of-runner/pytest-0/test_keras_model_classificatio0')
df_iris = #    sepal_length    sepal_width    petal_length    petal_width    class_
0    5.9             3.0            4.2     ...   3.8            1.7             0.3            0
149  6.2             2.9            4.3             1.3            1

    def test_keras_model_classification(tmpdir, df_iris):
        df = df_iris
        copy = df.copy()
        features = ['sepal_width', 'petal_length', 'sepal_length', 'petal_width']
        target = 'class_'
    
        # Preprocessing
        df = df.ml.minmax_scaler(features)
        df = df.ml.one_hot_encoder(features=[target])
        features = df.get_column_names(regex='^minmax')
        targets = df.get_column_names(regex='^class__')
    
        # Create the neural netwrok model
        nn_model = K.Sequential()
        nn_model.add(K.layers.Dense(4, activation='tanh'))
        nn_model.add(K.layers.Dense(3, activation='softmax'))
        nn_model.compile(optimizer=K.optimizers.RMSprop(learning_rate=0.01),
                        loss=K.losses.categorical_crossentropy,
                        metrics=['accuracy'])
    
        X = df[features].values
        y = df[targets].values
        nn_model.fit(x=X, y=y, validation_split=0.05, epochs=11, verbose=0)
    
        keras_model = vaex.ml.tensorflow.KerasModel(features=features, prediction_name='pred', model=nn_model)
        df_trans = keras_model.transform(df)
        assert df_trans.pred.shape == (150, 3)
    
        state_path = str(tmpdir.join('state.json'))
>       df_trans.state_write(state_path)

tests/ml/tensorflow_test.py:47: 

...

>       raise ValueError(
            "Invalid filepath extension for saving. "
            "Please add either a `.keras` extension for the native Keras "
            f"format (recommended) or a `.h5` extension. "
            "Use `model.export(filepath)` if you want to export a SavedModel "
            "for use with TFLite/TFServing/etc. "
            f"Received: filepath={filepath}."
        )
E       ValueError: Invalid filepath extension for saving. Please add either a `.keras` extension for the native Keras format (recommended) or a `.h5` extension. Use `model.export(filepath)` if you want to export a SavedModel for use with TFLite/TFServing/etc. Received: filepath=/tmp/tmpcu2kp0mw.

maartenbreddels

Awesome cleanup, great that you used uv!
Q: Why did we have to rename vaex-meta to meta?

maartenbreddels · 2024-08-01T10:45:33Z

packages/vaex-ml/setup.py

+      'annoy',
+      'scikit-learn',
+      'xgboost',
+      'lightgbm~=4.0',
+      'catboost',


these should be optional right? put them under all?

which ones exactly should go to all?

all these, since they were not a dependency before (more like a peer dependency - in npm speak)

maartenbreddels · 2024-08-01T10:46:59Z

setup.py

    ],
    packages=[],
    install_requires=[
-        'pip',
+        f"{package}[all] @ {(Path(__file__).parent / 'packages' / package).as_uri()}"


are you saying this works for an editable install? (awesome if yes!)

nope, this doesn't work for -e I'm afraid.

however, I think you could do pip install -e packages/* if you wanted.

In solara we do:
https://github.com/widgetti/solara/blob/master/requirements-dev.txt

But if we do this, we have to maybe update https://vaex.io/docs/installing.html

maartenbreddels · 2024-08-01T10:49:58Z

new pandas brings some issues:

We can simply update the keys, that is expected.

also tensorflow wants an extension now:

happy to disable that test and leave it up to a community member to fix it.

The rest of the errors would could solve by pinning in the ci optional requirements, what do you think?

ddelange · 2024-08-01T12:36:43Z

Q: Why did we have to rename vaex-meta to meta?

@maartenbreddels see PR description, second bullet punt last sub-bullet point

The rest of the errors would could solve by pinning

that is my hope too, although python 3.11 / 3.12 will need certain fresher libraries than were used before: e.g. graphql v3.

ddelange · 2024-08-01T12:57:38Z

packages/vaex-core/setup.py

-                         "nest-asyncio>=1.3.3", "pyarrow>=5.0.0", "frozendict!=2.2.0",
+setup_requires = ["numpy~=1.17"]
+install_requires_core = ["numpy~=1.17", "aplus", "tabulate>=0.8.3", "dask!=2022.4.0",
+                         "future>=0.15.2", "pyyaml", "six", "cloudpickle", "pandas",


pandas python 3.12 wheels were introduced in pandas v2.1.1, so that's the implicit minimum version CI needs to support. e.g. the error with 'M' as period was already changed in pandas v1.0. So I think vaex CI was construed with pandas v0.

ddelange · 2024-08-01T14:47:34Z

hmm ^ new commit unrelated to build, now there are build errors that weren't there last week 🤯

maartenbreddels · 2024-08-01T16:58:30Z

@maartenbreddels see PR description, second bullet punt last sub-bullet point

How did I miss that? 🤯

ddelange · 2024-08-02T08:09:22Z

vendor/boost/boost/mpl/aux_/integral_wrapper.hpp:73:31: error: integer value -1 is outside the valid range of values [0, 3] for this enumeration type [-Wenum-constexpr-conversion]
    typedef AUX_WRAPPER_INST( BOOST_MPL_AUX_STATIC_CAST(AUX_WRAPPER_VALUE_TYPE, (value - 1)) ) prior;

this clang boost compile error fixed by downgrading to macos-13. maybe github added a new version of clang on macos-14 last week...

ddelange · 2024-08-02T08:27:07Z

also tensorflow wants an extension now:

happy to disable that test and leave it up to a community member to fix it.

f2773c0

ddelange · 2024-08-17T20:42:40Z

@maartenbreddels ref #3 (comment) I'm going to need your support with:

the fingerprint issue, maybe related to pandas v1+ (required for recent python versions support, e.g. 3.11 was supported as of v1.5.0 ref and 3.12 as of v2.1.1 ref)
syntax on graphql-core v3 (required for recent python versions support ref)
AttributeError: 'csr_matrix' object has no attribute 'A'

maartenbreddels · 2024-08-23T14:42:46Z

We can do the fingerprint test like

x in ['dataframe-1c2f7e9c53dbd30220792e425418e343', 'dataframe-...']

The matrix issue can be fixed with (I cannot push to your branch):

diff --git a/packages/vaex-core/vaex/column.py b/packages/vaex-core/vaex/column.py
index 40de5aab..87938ce3 100644
--- a/packages/vaex-core/vaex/column.py
+++ b/packages/vaex-core/vaex/column.py
@@ -163,7 +163,10 @@ class ColumnSparse(Column):
 
     def __getitem__(self, slice):
         # not sure if this is the fastest
-        return self.matrix[slice, self.column_index].A[:,0]
+        mat = self.matrix[slice, self.column_index]
+        dense = mat.todense()
+        column_part = np.asarray(dense)[:,0]
+        return column_part
 
 
 class ColumnNumpyLike(Column):

We can drop support for graphql for older python versions, so we skip that test?

ddelange · 2024-08-23T22:17:20Z

the test does assert df_a.fingerprint() == df_b.fingerprint(), so I think the test actually catches something here which is making fingerprinting non-deterministic (and needs fixing)

ddelange · 2024-08-23T22:18:49Z

regarding the graphql v3 requirement, do you suggest to leave the test broken? maybe its just a simple syntax update to the query (just need to figure out which)

ddelange · 2024-08-24T12:58:01Z

I'll merge this into the main PR so you can take over again if that's OK 🙏

* fix[ml]: adjust tests to reflect latest apis of 3rd party libraries (xgboost, lightgbm) * Build wheels on pull_request * Maximize wheel build parallellization * git submodule update --remote --merge packages/vaex-core/vendor/pybind11 * Fix gcc error * Fix workflow syntax * Remove redefinition https://github.com/pybind/pybind11/blob/769fd3b889fef6cddb060f2a0be26aee62b4da05/include/pybind11/pytypes.h#L859 https://github.com/ddelange/vaex/actions/runs/3965609112/jobs/6795506653#step:6:2110 * Disable win32 https://github.com/ddelange/vaex/actions/runs/3965689146/jobs/6795667118#step:6:538 * Remove testing leftovers * Add upper cap on lightgbm microsoft/LightGBM#5196 (comment) * Migrate to mamba-org/setup-micromamba@v1 * Upload release assets on Github (pre)release * Add missing permission for release assets * Build cp312 wheels * Remove setuptools and wheel from pyproject.toml * Replace imp with importlib * chore: trigger ci * ci: upgrade xcode for brew install libomp * update requirements-ml to comply with the latest veex-ml expectations. * try to install lightgbm via pip * fix: only the mini taxi file is on s3 for cost savings * ci: pin dask<2024.2.0 to get the same hash keys This version gives different results, although not a problem in production (it will make your cache invalid though), for CI we test that we have stable keys (fingerprints) * fix: only the mini taxi file is on s3 for cost savings (2) * ci: skip notebooks that depend on the dataframe server * ci: skip ci steps in various old python versions * Bump micromamba and other actions * ci: specific test hangs on ci+osx * Bump cibuildwheel, use native arm64 mac runners * test: log is renamed to log loss * test: skip lightgbm tests on 36 and 37 * test: skip sklearn tests on 36 and 37 * Update packages/vaex-core/setup.py Co-authored-by: ddelange <[email protected]> * chore: drop python 3.6 and 3.7 support Co-authored-by: Ewout ter Hoeven <[email protected]> * test: skip a failing test for windows * ci: macOS 11 is retired as of June 28 * ci: always build wheels, but do not publish instead * ci: try with older micromamba, 1.5.6 seems to sometimes hang * Install setuptools in wheel.yml * Add sudo * Fix windows * Use sudo only on macos * Fix empty string evaluating false * Add .readthedocs.yaml * Pull submodules * Try editable rtd install * Try move editable install to requirements_rtd.txt * Allow newer sphinx * Autocancel previous runs in PRs * Autocancel all jobs * Sphinx sidebar fix https://github.com/dmwyatt/sphinx-book-theme/blob/2416ef6cebc3d83019dbc83f25c61b72719cad55/docs/customize/sidebar-primary.md#default-sidebar-elements * Remove autocancel, doesn't work from forks * Add cancel-in-progress keys * Amend to only cancel on PRs * Add CIBW_TEST_COMMAND * Disallow numpy v2 * Skip musllinux due to misding blake3 wheels * Install carco for blake3 source install * Fix CIBW_ENVIRONMENT * Add CIBW_TEST_SKIP * Build wheels with numpy v1 * run tests on 3.10-3.12 * upgrade micromamba * unpin pytest-asyncio * try different pin as <2.0 because mamba crashes on that * Fix dephell and CI hell (ddelange#3) * allow for multiple fingerprints related to package versions * explicit dtype=object * fix pandas issues * revert graphene upgrade * force uv to ignore Werkzeug (we do not use it, its a subdependency which has conflicts) * fix: work with modern scipy/sparse arrays * skip test for python 3.8 * skip test on windows * add missing import * skip some tests conditionally * tuple comparison fix * does osx run without sourcing .bash profile? * skip a few more tests * only run test on linux * Try building arm64 wheels on macos-13 * Switch to older Xcode * Fix readthedocs by manually unpinning werkzeug in the graphene-tornado v2.6.1 release # for now we leave graphene-tornado / graphene on v2 # ref #2356 (comment) # in top level vaex[ci] we added the following (additional) direct reference requirement which adds one commit (ddelange/graphene-tornado@d75f01f) on top of the 2.6.1 release to unpin werkzeug # "graphene-tornado @ https://github.com/ddelange/graphene-tornado/archive/refs/heads/2.6.1.unpin-werkzeug.zip", * Simplify * Revert "Simplify" This reverts commit 91263bc. * Add back tags event trigger for core wheel upload * Switch to trusted publishing * Add trusted publishing for vaex-core * Exclude windows 3.8 from pythonpackage.yml * Don't upload vaex-core source distribution * Disable cp38-win vaex-core wheel build * Add permission --------- Co-authored-by: Jovan Veljanoski <[email protected]> Co-authored-by: Maarten A. Breddels <[email protected]> Co-authored-by: Ewout ter Hoeven <[email protected]>

ddelange force-pushed the fix-ci branch 26 times, most recently from 410c757 to db0dd6c Compare July 19, 2024 21:21

Fix CI using uv instead of micromamba

6ebe094

ddelange force-pushed the fix-ci branch from db0dd6c to 6ebe094 Compare July 19, 2024 21:30

ddelange added 2 commits July 20, 2024 01:01

Update 04-run-test-suite.sh

5002236

Update 04-run-test-suite.sh

317292e

ddelange added 5 commits July 20, 2024 15:26

Update transformations.py

a4cd7b8

Update setup.py

271497f

Update setup.py

8f4267c

Update tensorflow_test.py

06ca095

Update agg_test.py

3df5fe7

ddelange mentioned this pull request Jul 20, 2024

Support CPython 3.11, 3.12, and aarch64 processors vaexio/vaex#2331

Merged

ddelange changed the title ~~Fix CI using uv instead of micromamba~~ Fix dephell and CI hell Jul 20, 2024

Update strings_test.py

25c09a0

maartenbreddels reviewed Aug 1, 2024

View reviewed changes

ddelange commented Aug 1, 2024

View reviewed changes

Update pd.Timedelta initialize

68e689b

ddelange added 4 commits August 2, 2024 08:06

Fix windows CI

fee017a

Downgrade macos

62ef75f

Fix weekofyear deprecation

febe688

Run ruff format on vaex-core setup.py

ff974b0

Fix keras model save error

f2773c0

ddelange merged commit 0482887 into build-matrix Aug 24, 2024
0 of 15 checks passed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Fix dephell and CI hell #3

Fix dephell and CI hell #3

ddelange commented Jul 19, 2024 •

edited

Loading

ddelange commented Jul 25, 2024

maartenbreddels left a comment

maartenbreddels Aug 1, 2024

ddelange Aug 1, 2024

maartenbreddels Aug 1, 2024

maartenbreddels Aug 1, 2024

ddelange Aug 1, 2024

maartenbreddels Aug 1, 2024

maartenbreddels commented Aug 1, 2024

ddelange commented Aug 1, 2024

ddelange Aug 1, 2024 •

edited

Loading

ddelange Aug 1, 2024

ddelange commented Aug 1, 2024

maartenbreddels commented Aug 1, 2024

ddelange commented Aug 2, 2024 •

edited

Loading

ddelange commented Aug 2, 2024

ddelange commented Aug 17, 2024 •

edited

Loading

maartenbreddels commented Aug 23, 2024

ddelange commented Aug 23, 2024 •

edited

Loading

ddelange commented Aug 23, 2024

ddelange commented Aug 24, 2024

Fix dephell and CI hell #3

Fix dephell and CI hell #3

Conversation

ddelange commented Jul 19, 2024 • edited Loading

ddelange commented Jul 25, 2024

maartenbreddels left a comment

Choose a reason for hiding this comment

maartenbreddels Aug 1, 2024

Choose a reason for hiding this comment

ddelange Aug 1, 2024

Choose a reason for hiding this comment

maartenbreddels Aug 1, 2024

Choose a reason for hiding this comment

maartenbreddels Aug 1, 2024

Choose a reason for hiding this comment

ddelange Aug 1, 2024

Choose a reason for hiding this comment

maartenbreddels Aug 1, 2024

Choose a reason for hiding this comment

maartenbreddels commented Aug 1, 2024

ddelange commented Aug 1, 2024

ddelange Aug 1, 2024 • edited Loading

Choose a reason for hiding this comment

ddelange Aug 1, 2024

Choose a reason for hiding this comment

ddelange commented Aug 1, 2024

maartenbreddels commented Aug 1, 2024

ddelange commented Aug 2, 2024 • edited Loading

ddelange commented Aug 2, 2024

ddelange commented Aug 17, 2024 • edited Loading

maartenbreddels commented Aug 23, 2024

ddelange commented Aug 23, 2024 • edited Loading

ddelange commented Aug 23, 2024

ddelange commented Aug 24, 2024

ddelange commented Jul 19, 2024 •

edited

Loading

ddelange Aug 1, 2024 •

edited

Loading

ddelange commented Aug 2, 2024 •

edited

Loading

ddelange commented Aug 17, 2024 •

edited

Loading

ddelange commented Aug 23, 2024 •

edited

Loading