Remove support for Python 3.7 #1988

SimonYansenZhao · 2023-09-05T02:24:10Z

Description

This PR dropped the support for Python 3.7 since it'll reach the end of life on Jun 27, 2023.
This PR is moved from #1937, because of the sign-off requirement.

Update docs
Update Conda env script

Related Issues

[ASK] Whether to drop support for Python 3.6 #1638

References

Python endoflife.date

Checklist:

I have followed the contribution guidelines and code style for this project.
I have added tests covering my contributions.
I have updated the documentation accordingly.
This PR is being made to staging branch and not to main branch.

SimonYansenZhao · 2023-09-05T03:15:50Z

When using Python 3.11, some package requires numpy 1.19.4 which may not support Python 3.11. See https://github.com/recommenders-team/recommenders/actions/runs/6079503723/job/16492093035

miguelgfierro · 2023-09-05T09:18:43Z

When using Python 3.11, some package requires numpy 1.19.4 which may not support Python 3.11. See https://github.com/recommenders-team/recommenders/actions/runs/6079503723/job/16492093035

I see two options:

Start with 3.8-3.10, and leave 3.11 to a future PR
Consider adding different numpy versions depending on the python version as we did with cornac:

    "cornac>=1.1.2,<1.15.2;python_version<='3.7'",
    "cornac>=1.15.2,<2;python_version>='3.8'",

What do you think @SimonYansenZhao?

SimonYansenZhao · 2023-09-06T04:07:48Z

When using Python 3.11, some package requires numpy 1.19.4 which may not support Python 3.11. See https://github.com/recommenders-team/recommenders/actions/runs/6079503723/job/16492093035

I see two options:

Start with 3.8-3.10, and leave 3.11 to a future PR

Consider adding different numpy versions depending on the python version as we did with cornac:
    "cornac>=1.1.2,<1.15.2;python_version<='3.7'",
    "cornac>=1.15.2,<2;python_version>='3.8'",
What do you think @SimonYansenZhao?

@miguelgfierro I tried option 2, but failed, see https://github.com/recommenders-team/recommenders/actions/runs/6079732708/job/16496793669?pr=1988

I'll try python 3.8 - 3.10 in this PR and leave 3.11 in another future PR.

SimonYansenZhao · 2023-09-07T14:06:12Z

Some notes from the test results:

Some dependencies require numpy of some old version which does not support Python 3.10+, so numpy needs to be compiled against Python 3.10+ to build its binary file during setup of recommenders. So we need to find out which dependences require old version of numpy.
scrapbook is very old and is not actively developed now. The latest version was published in 2021. Some code of scrapbook uses DataFrame.append() (See https://github.com/recommenders-team/recommenders/actions/runs/6108846153/job/16578689728?pr=1988#step:3:2639) which is removed from pandas since pandas 2.0. So an alternative to scrapbook should be used, or we should bypass the code in scrapbook that uses DataFrame.append().

miguelgfierro · 2023-09-07T15:37:45Z

I asked ChatGPT how to replace papermill and scrapbook, and it looks that there is a dependency that we already have called nbconvert. This would be the code:

import nbformat
from nbconvert.preprocessors import ExecutePreprocessor

# Step 1: Prepare Your Jupyter Notebook

# Ensure your Jupyter Notebook is set up with parameters that can be modified and read.
# Use Markdown cells to specify parameters that need modification and code cells to set parameters that need to be read.

# Step 2: Modify Parameters Programmatically

# Load the Jupyter Notebook
notebook_path = "your_notebook.ipynb"
with open(notebook_path, "r") as notebook_file:
    notebook_content = nbformat.read(notebook_file, as_version=4)

# Define the parameters to modify
new_param1 = "New Value1"
new_param2 = "New Value2"

# Search for and replace parameter values in code cells
for cell in notebook_content.cells:
    if cell.cell_type == "code":
        cell.source = cell.source.replace("param1 = 'Default Value'", f"param1 = '{new_param1}'")
        cell.source = cell.source.replace("param2 = 'Default Value'", f"param2 = '{new_param2}'")

# Save the modified notebook
modified_notebook_path = "modified_notebook.ipynb"
with open(modified_notebook_path, "w", encoding="utf-8") as modified_notebook_file:
    nbformat.write(notebook_content, modified_notebook_file)

print("Parameters modified and saved to", modified_notebook_path)

# Step 3: Execute the Modified Notebook with nbconvert

# Load the modified notebook
with open(modified_notebook_path, "r") as modified_notebook_file:
    modified_notebook_content = nbformat.read(modified_notebook_file, as_version=4)

# Create an execution preprocessor
execute_preprocessor = ExecutePreprocessor(timeout=600, kernel_name="python3")

# Execute the notebook
executed_notebook, _ = execute_preprocessor.preprocess(modified_notebook_content, {"metadata": {"path": "./"}})

# Save the executed notebook
executed_notebook_path = "executed_notebook.ipynb"
with open(executed_notebook_path, "w", encoding="utf-8") as executed_notebook_file:
    nbformat.write(executed_notebook, executed_notebook_file)

print("Notebook execution complete. Results saved to", executed_notebook_path)

# Step 4: Read Parameters Programmatically

# Load the executed notebook
with open(executed_notebook_path, "r") as executed_notebook_file:
    executed_notebook_content = nbformat.read(executed_notebook_file, as_version=4)

# Extract parameter values from code cells
param3 = None
for cell in executed_notebook_content.cells:
    if cell.cell_type == "code":
        if "param3" in cell.source:
            # Extract param3 value from code cell
            exec(cell.source)
            break

# Now you can use param3 in your workflow
if param3 is not None:
    print("Value of param3:", param3)
else:
    print("param3 not found in the executed notebook.")

what do you think @SimonYansenZhao?

miguelgfierro · 2023-09-07T17:57:06Z

@SimonYansenZhao @anargyri I'm trying to replicate papermill miguelgfierro/pybase#73 if I do the basic functionality, we can remove 2 deps, and will depend only on nbconvert, which is a library supported by Jupyter, and re. Still WIP.

SimonYansenZhao · 2023-09-08T09:48:32Z

@miguelgfierro I think using nbconvert is doable. But now could you help me with the errors in the test result: https://github.com/recommenders-team/recommenders/actions/runs/6116241630/job/16601154561?pr=1988#step:3:2654

NameError: name 'XDeepFMModel' is not defined

SimonYansenZhao · 2023-09-08T10:25:34Z

And I find ploomber may be used to replace scrapbook. Ploomer is built on top of papermill.

anargyri · 2023-09-08T13:31:33Z

scrapbook is very old and is not actively developed now. The latest version was published in 2021. Some code of scrapbook uses DataFrame.append() (See https://github.com/recommenders-team/recommenders/actions/runs/6108846153/job/16578689728?pr=1988#step:3:2639) which is removed from pandas since pandas 2.0. So an alternative to scrapbook should be used, or we should bypass the code in scrapbook that uses DataFrame.append().

But you have pandas < 1.6 in setup.py, don't you? Or I am missing something?

SimonYansenZhao · 2023-09-09T01:14:53Z

3. scrapbook is very old and is not actively developed now. The latest version was published in 2021. Some code of scrapbook uses DataFrame.append() (See https://github.com/recommenders-team/recommenders/actions/runs/6108846153/job/16578689728?pr=1988#step:3:2639) which is removed from pandas since pandas 2.0. So an alternative to scrapbook should be used, or we should bypass the code in scrapbook that uses DataFrame.append().

But you have pandas < 1.6 in setup.py, don't you? Or I am missing something?

@anargyri after I found pandas 2.0 didn't work for scrapbook, I tried to set pandas < 1.6 in setup.py to see if pandas < 1.6 would work. Now scrapbook works with pandas < 1.6, but we cannot use pandas 2.0+ in the future if scrapbook won't upgrade.

miguelgfierro · 2023-09-09T05:51:19Z

I'm looking into this:

@miguelgfierro I think using nbconvert is doable. But now could you help me with the errors in the test result: https://github.com/recommenders-team/recommenders/actions/runs/6116241630/job/16601154561?pr=1988#step:3:2654

(recommenders) u@unicorn:~/MS/recommenders$ python --version
Python 3.9.16
(recommenders) u@unicorn:~/MS/recommenders$ pip list | grep pandas
pandas                       1.5.3
(recommenders) u@unicorn:~/MS/recommenders$ pytest tests/unit/recommenders/models/test_deeprec_model.py::test_xdeepfm_component_definition --disable-warnings
=========================================================================== test session starts ===========================================================================
platform linux -- Python 3.9.16, pytest-7.3.2, pluggy-1.0.0
rootdir: /home/u/MS/recommenders
configfile: pyproject.toml
plugins: typeguard-4.0.0, cov-4.1.0, hypothesis-6.79.1, mock-3.11.1, anyio-3.7.0
collected 1 item

tests/unit/recommenders/models/test_deeprec_model.py .                                                                                                              [100%]

===================================================================== 1 passed, 20 warnings in 3.70s ======================================================================

(recommenders) u@unicorn:~/MS/recommenders$ pip install pandas --upgrade
(recommenders) u@unicorn:~/MS/recommenders$ pip list | grep pandas
pandas                       2.1.0
(recommenders) u@unicorn:~/MS/recommenders$ pytest tests/unit/recommenders/models/test_deeprec_model.py::test_xdeepfm_component_definition --disable-warnings
=========================================================================== test session starts ===========================================================================
platform linux -- Python 3.9.16, pytest-7.3.2, pluggy-1.0.0
rootdir: /home/u/MS/recommenders
configfile: pyproject.toml
plugins: typeguard-4.0.0, cov-4.1.0, hypothesis-6.79.1, mock-3.11.1, anyio-3.7.0
collected 1 item

tests/unit/recommenders/models/test_deeprec_model.py .                                                                                                              [100%]

===================================================================== 1 passed, 20 warnings in 3.82s ======================================================================
(recommenders) u@unicorn:~/MS/recommenders$

(reco_gpu) u@unicorn:~/MS/recommenders$ python --version
Python 3.8.13
(reco_gpu) u@unicorn:~/MS/recommenders$ pip list | grep pandas
pandas                        1.4.2
(reco_gpu) u@unicorn:~/MS/recommenders$ pytest tests/unit/recommenders/models/test_deeprec_model.py::test_xdeepfm_component_definition --disable-warnings
=========================================================================== test session starts ===========================================================================
platform linux -- Python 3.8.13, pytest-7.1.2, pluggy-1.0.0
rootdir: /home/u/MS/recommenders, configfile: pyproject.toml
plugins: hypothesis-6.46.3, rerunfailures-10.2, cov-3.0.0, mock-3.7.0
collected 1 item

tests/unit/recommenders/models/test_deeprec_model.py .                                                                                                              [100%]

===================================================================== 1 passed, 11 warnings in 14.74s =====================================================================
(reco_gpu) u@unicorn:~/MS/recommenders$ pip install pandas --upgrade
(reco_gpu) u@unicorn:~/MS/recommenders$ pip list | grep pandas
pandas                        2.0.3
(reco_gpu) u@unicorn:~/MS/recommenders$ pytest tests/unit/recommenders/models/test_deeprec_model.py::test_xdeepfm_component_definition --disable-warnings
=========================================================================== test session starts ===========================================================================
platform linux -- Python 3.8.13, pytest-7.1.2, pluggy-1.0.0
rootdir: /home/u/MS/recommenders, configfile: pyproject.toml
plugins: hypothesis-6.46.3, rerunfailures-10.2, cov-3.0.0, mock-3.7.0
collected 0 items / 1 error

================================================================================= ERRORS ==================================================================================
__________________________________________________ ERROR collecting tests/unit/recommenders/models/test_deeprec_model.py __________________________________________________
tests/unit/recommenders/models/test_deeprec_model.py:7: in <module>
    from recommenders.datasets import movielens
recommenders/datasets/movielens.py:36: in <module>
    import pandera as pa
../../anaconda/envs/reco_gpu/lib/python3.8/site-packages/pandera/__init__.py:48: in <module>
    from . import errors, pandas_accessor, typing
../../anaconda/envs/reco_gpu/lib/python3.8/site-packages/pandera/typing/__init__.py:9: in <module>
    from . import dask, fastapi, geopandas, modin, pyspark
../../anaconda/envs/reco_gpu/lib/python3.8/site-packages/pandera/typing/dask.py:9: in <module>
    import dask.dataframe as dd
../../anaconda/envs/reco_gpu/lib/python3.8/site-packages/dask/dataframe/__init__.py:3: in <module>
    from dask.dataframe import backends, dispatch, rolling
../../anaconda/envs/reco_gpu/lib/python3.8/site-packages/dask/dataframe/backends.py:20: in <module>
    from dask.dataframe.core import DataFrame, Index, Scalar, Series, _Frame
../../anaconda/envs/reco_gpu/lib/python3.8/site-packages/dask/dataframe/core.py:29: in <module>
    from dask.dataframe import methods
../../anaconda/envs/reco_gpu/lib/python3.8/site-packages/dask/dataframe/methods.py:21: in <module>
    from dask.dataframe.utils import is_dataframe_like, is_index_like, is_series_like
../../anaconda/envs/reco_gpu/lib/python3.8/site-packages/dask/dataframe/utils.py:18: in <module>
    from dask.dataframe import (  # noqa: F401 register pandas extension types
../../anaconda/envs/reco_gpu/lib/python3.8/site-packages/dask/dataframe/_dtypes.py:3: in <module>
    from dask.dataframe.extensions import make_array_nonempty, make_scalar
../../anaconda/envs/reco_gpu/lib/python3.8/site-packages/dask/dataframe/extensions.py:6: in <module>
    from dask.dataframe.accessor import (
../../anaconda/envs/reco_gpu/lib/python3.8/site-packages/dask/dataframe/accessor.py:190: in <module>
    class StringAccessor(Accessor):
../../anaconda/envs/reco_gpu/lib/python3.8/site-packages/dask/dataframe/accessor.py:275: in StringAccessor
    @derived_from(pd.core.strings.StringMethods)
E   AttributeError: module 'pandas.core.strings' has no attribute 'StringMethods'
========================================================================= short test summary info =========================================================================
ERROR tests/unit/recommenders/models/test_deeprec_model.py - AttributeError: module 'pandas.core.strings' has no attribute 'StringMethods'
============================================================================ 1 error in 1.36s =============================================================================
ERROR: not found: /home/u/MS/recommenders/tests/unit/recommenders/models/test_deeprec_model.py::test_xdeepfm_component_definition
(no name '/home/u/MS/recommenders/tests/unit/recommenders/models/test_deeprec_model.py::test_xdeepfm_component_definition' in any of [<Module test_deeprec_model.py>])

(reco_gpu) u@unicorn:~/MS/recommenders$

@SimonYansenZhao @anargyri it seems that with python 3.8, we need pandas<2, while with 3.9 we can have pandas>1.5

Trying different pandas versions with different python -> https://github.com/recommenders-team/recommenders/actions/runs/6129325061
[UPDATE] In the logs of python 3.8 I see pandas 1.5.3 and it fails.
In local it works:

(reco_gpu) u@unicorn:~/MS/recommenders$ python --version
Python 3.8.13
(reco_gpu) u@unicorn:~/MS/recommenders$ pip list | grep pandas
pandas                        1.5.3
(reco_gpu) u@unicorn:~/MS/recommenders$ pytest tests/unit/recommenders/models/test_deeprec_model.py::test_xdeepfm_component_definition --disable-warnings
=========================================================================== test session starts ===========================================================================
platform linux -- Python 3.8.13, pytest-7.1.2, pluggy-1.0.0
rootdir: /home/u/MS/recommenders, configfile: pyproject.toml
plugins: hypothesis-6.46.3, rerunfailures-10.2, cov-3.0.0, mock-3.7.0
collected 1 item

tests/unit/recommenders/models/test_deeprec_model.py .                                                                                                              [100%]

===================================================================== 1 passed, 11 warnings in 6.88s ======================================================================
(reco_gpu) u@unicorn:~/MS/recommenders$

So it seems that the error doesn't come from pandas.

@SimonYansenZhao have you been able to replicate the error NameError: name 'XDeepFMModel' is not defined in local?

miguelgfierro · 2023-09-09T09:15:42Z

I looked at LightGBM https://github.com/microsoft/LightGBM, and they have support for Python 3.7. Wouldn't it be easier if we keep 3.7?

Thoughts @SimonYansenZhao @anargyri

SimonYansenZhao · 2023-09-11T10:01:33Z

@miguelgfierro I cannot replicate the error NameError: name 'XDeepFMModel' is not defined, which is why I am always waiting for the test results in DevOps.

miguelgfierro · 2023-09-11T15:37:24Z

setup.py

        "Operating System :: POSIX :: Linux",
    ],
    extras_require=extras_require,
    keywords="recommendations recommendation recommenders recommender system engine "
    "machine learning python spark gpu",
    install_requires=install_requires,
    package_dir={"recommenders": "recommenders"},
-    python_requires=">=3.6, <3.10",
+    python_requires=">=3.8, <=3.9",


Suggested change

python_requires=">=3.8, <=3.9",

python_requires=">=3.6",

@SimonYansenZhao we had a discussion in the reco meeting. There are a couple of things we are proposing:

leave python requires with >= 3.6 so people can install recommenders with any version of python. Right now, for example, people can't install the pypi package on 3.10 Then we say that we officially only support 3.8-3.10. With that we avoid issues like [ASK] No module named 'recommenders' after installing recommenders in Colab #1928 where people can't install recommenders on versions that are not in the requires, and at the same time, we officially only tests the versions we support.

Maybe it would be easier to split this PR into 2, first we remove the support of 3.7, and then in another PR we add 3.10.

What do you think?

Signed-off-by: Simon Zhao <[email protected]>

is_sequence() is deleted in TensorFlow 2.13.0. See https://github.com/tensorflow/tensorflow/releases/tag/v2.13.0 Signed-off-by: Simon Zhao <[email protected]>

Signed-off-by: Simon Zhao <[email protected]>

…der.tokenize_text() Signed-off-by: Simon Zhao <[email protected]>

Signed-off-by: Simon Zhao <[email protected]>

SimonYansenZhao · 2023-09-14T05:19:31Z

@miguelgfierro @anargyri I find ./tools/generate_conda_file.py is out of sync with setup.py. Is it still needed?

As discussed in #1988 (comment), I only dropped the support for Python 3.7 in this PR, and will create another PR for adding support for 3.10.

miguelgfierro · 2023-09-14T08:48:01Z

@miguelgfierro @anargyri I find ./tools/generate_conda_file.py is out of sync with setup.py. Is it still needed?

No, we can drop it

miguelgfierro

I think we should change python_requires=">=3.6", as of this comment: #1988 (comment) or if we don't like 3.6, we can do 3.7+

Apart from that, this is great

miguelgfierro reviewed Sep 11, 2023

View reviewed changes

SimonYansenZhao changed the title ~~Remove support for Python 3.7, add Python 3.10 and 3.11~~ Remove support for Python 3.7 Sep 12, 2023

SimonYansenZhao added 9 commits September 12, 2023 14:14

Upgrade dependencies, remove support for python 3.7 and add python 3.11

284a9ff

Signed-off-by: Simon Zhao <[email protected]>

Use numpy>=1.23.3 for Python 3.11

73a47c2

Signed-off-by: Simon Zhao <[email protected]>

Remove support for Python 3.11

bfd9336

Signed-off-by: Simon Zhao <[email protected]>

Remove constraint on hypothesis

3ce6f03

Signed-off-by: Simon Zhao <[email protected]>

Try only Python 3.8 and 3.9

a870c4f

Signed-off-by: Simon Zhao <[email protected]>

Set atol and rtol

e7447f9

Signed-off-by: Simon Zhao <[email protected]>

Set pandas<1.6

418f4b0

Signed-off-by: Simon Zhao <[email protected]>

Replace is_sequence() with is_nested()

18c89aa

is_sequence() is deleted in TensorFlow 2.13.0. See https://github.com/tensorflow/tensorflow/releases/tag/v2.13.0 Signed-off-by: Simon Zhao <[email protected]>

Revert dependency versions

1c6a1ff

Signed-off-by: Simon Zhao <[email protected]>

SimonYansenZhao force-pushed the simonz-dep-upgrade-20230905 branch from 229a3bf to 1c6a1ff Compare September 12, 2023 07:03

SimonYansenZhao added 5 commits September 12, 2023 15:45

Lower version of category-encoders

7b2d9ff

Signed-off-by: Simon Zhao <[email protected]>

Loose dependency upper versions

96bdaa3

Signed-off-by: Simon Zhao <[email protected]>

Use newer version of transformers

67da0b5

Signed-off-by: Simon Zhao <[email protected]>

Correct min_df in recommenders.models.tfidf.tfidf_utils.TfidfRecommen…

12f1eda

…der.tokenize_text() Signed-off-by: Simon Zhao <[email protected]>

Set fastai<2

97deed1

Signed-off-by: Simon Zhao <[email protected]>

SimonYansenZhao mentioned this pull request Sep 13, 2023

[FEATURE] Upgrade fastai to v2 #1992

Closed

SimonYansenZhao added 3 commits September 13, 2023 17:42

Revert dependency versions

1371084

Signed-off-by: Simon Zhao <[email protected]>

Update

659ac17

Signed-off-by: Simon Zhao <[email protected]>

Only drop the support for Python 3.7

edfa101

Signed-off-by: Simon Zhao <[email protected]>

SimonYansenZhao marked this pull request as ready for review September 14, 2023 04:58

SimonYansenZhao requested review from gramhagen, anargyri, loomlike and wutaomsft as code owners September 14, 2023 04:58

Update versions

31a8b3d

Signed-off-by: Simon Zhao <[email protected]>

miguelgfierro approved these changes Sep 15, 2023

View reviewed changes

python 3.6+ in requires

4f99875

miguelgfierro mentioned this pull request Sep 15, 2023

Delete conda and requirements generator files not used anymore #1993

Merged

4 tasks

miguelgfierro merged commit 8d67ef5 into staging Sep 15, 2023

miguelgfierro deleted the simonz-dep-upgrade-20230905 branch September 15, 2023 11:51

miguelgfierro mentioned this pull request Oct 8, 2023

[FEATURE] Alternative to scrapbook to execute notebooks programmatically for tests #2012

Closed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Remove support for Python 3.7 #1988

Remove support for Python 3.7 #1988

SimonYansenZhao commented Sep 5, 2023 •

edited

Loading

SimonYansenZhao commented Sep 5, 2023

miguelgfierro commented Sep 5, 2023

SimonYansenZhao commented Sep 6, 2023

SimonYansenZhao commented Sep 7, 2023 •

edited

Loading

miguelgfierro commented Sep 7, 2023

miguelgfierro commented Sep 7, 2023

SimonYansenZhao commented Sep 8, 2023

SimonYansenZhao commented Sep 8, 2023 •

edited

Loading

anargyri commented Sep 8, 2023

SimonYansenZhao commented Sep 9, 2023

miguelgfierro commented Sep 9, 2023 •

edited

Loading

miguelgfierro commented Sep 9, 2023

SimonYansenZhao commented Sep 11, 2023

miguelgfierro Sep 11, 2023

SimonYansenZhao Sep 12, 2023

SimonYansenZhao commented Sep 14, 2023

miguelgfierro commented Sep 14, 2023

miguelgfierro left a comment •

edited

Loading

Remove support for Python 3.7 #1988

Remove support for Python 3.7 #1988

Conversation

SimonYansenZhao commented Sep 5, 2023 • edited Loading

Description

Related Issues

References

Checklist:

SimonYansenZhao commented Sep 5, 2023

miguelgfierro commented Sep 5, 2023

SimonYansenZhao commented Sep 6, 2023

SimonYansenZhao commented Sep 7, 2023 • edited Loading

miguelgfierro commented Sep 7, 2023

miguelgfierro commented Sep 7, 2023

SimonYansenZhao commented Sep 8, 2023

SimonYansenZhao commented Sep 8, 2023 • edited Loading

anargyri commented Sep 8, 2023

SimonYansenZhao commented Sep 9, 2023

miguelgfierro commented Sep 9, 2023 • edited Loading

miguelgfierro commented Sep 9, 2023

SimonYansenZhao commented Sep 11, 2023

miguelgfierro Sep 11, 2023

Choose a reason for hiding this comment

SimonYansenZhao Sep 12, 2023

Choose a reason for hiding this comment

SimonYansenZhao commented Sep 14, 2023

miguelgfierro commented Sep 14, 2023

miguelgfierro left a comment • edited Loading

Choose a reason for hiding this comment

SimonYansenZhao commented Sep 5, 2023 •

edited

Loading

SimonYansenZhao commented Sep 7, 2023 •

edited

Loading

SimonYansenZhao commented Sep 8, 2023 •

edited

Loading

miguelgfierro commented Sep 9, 2023 •

edited

Loading

miguelgfierro left a comment •

edited

Loading