…bute 'get'

python/importlib_metadata#406

This reverts commit a37f302.

Traceback (most recent call last):
  File "~/cihai/cihai-cli/.venv/bin/flake8", line 8, in <module>
    sys.exit(main())
  File "~/cihai/cihai-cli/.venv/lib/python3.7/site-packages/flake8/main/cli.py", line 22, in main
    app.run(argv)
  File "~/cihai/cihai-cli/.venv/lib/python3.7/site-packages/flake8/main/application.py", line 363, in run
    self._run(argv)
  File "~/cihai/cihai-cli/.venv/lib/python3.7/site-packages/flake8/main/application.py", line 350, in _run
    self.initialize(argv)
  File "~/cihai/cihai-cli/.venv/lib/python3.7/site-packages/flake8/main/application.py", line 330, in initialize
    self.find_plugins(config_finder)
  File "~/cihai/cihai-cli/.venv/lib/python3.7/site-packages/flake8/main/application.py", line 153, in find_plugins
    self.check_plugins = plugin_manager.Checkers(local_plugins.extension)
  File "~/cihai/cihai-cli/.venv/lib/python3.7/site-packages/flake8/plugins/manager.py", line 357, in __init__
    self.namespace, local_plugins=local_plugins
  File "~/cihai/cihai-cli/.venv/lib/python3.7/site-packages/flake8/plugins/manager.py", line 238, in __init__
    self._load_entrypoint_plugins()
  File "~/cihai/cihai-cli/.venv/lib/python3.7/site-packages/flake8/plugins/manager.py", line 254, in _load_entrypoint_plugins
    eps = importlib_metadata.entry_points().get(self.namespace, ())
AttributeError: 'EntryPoints' object has no attribute 'get'

…gured out

See also: python/importlib_metadata#406 (comment)

`importlib-metadata 5.0.0` breaks `flake8`, causing CI to fail.

`importlib` recommends constraining `importlib < 5` until flake8 adapts.

python/importlib_metadata#406 (comment)

* Tests fail on python3.7 because of a problem with the importlib_metadata package.
The issue was fixed using the workaround suggsted here: python/importlib_metadata#406 (comment)

- Causes xarray netcdf load error
- See python/importlib_metadata#406

* Updated readthedocs install method
* Pin importlib_metadata to < v5
  - Causes xarray netcdf load error
  - See python/importlib_metadata#406
* Update .readthedocs.yaml
* Update requirements.txt
* Update conf.py

See python/importlib_metadata#406

There is a compatibility issue with importlib-metadata and flake8 [1].

[1] python/importlib_metadata#406

There is a compatibility issue between importlib-metadata 5.0.0 and
flake8 under Python 3.7 [1].

[1] python/importlib_metadata#406

There is a compatibility issue between importlib-metadata 5.0.0 and
flake8 under Python 3.7 [1].

[1] python/importlib_metadata#406

There is a compatibility issue between importlib-metadata 5.0.0 and
flake8 under Python 3.7 [1].

[1] python/importlib_metadata#406

There is a compatibility issue between importlib-metadata 5.0.0 and
flake8 under Python 3.7 [1].

[1] python/importlib_metadata#406

Try installing older version of importlib_metadata for python/importlib_metadata#406

Try installing older version of importlib_metadata due to python/importlib_metadata#406

python/importlib_metadata#406

python/importlib_metadata#406

refer to:
python/importlib_metadata#406

Remove python 3.7 from test matrix, conflict in flake8 dependencies.

Refer to: python/importlib_metadata#406

Remove python 3.7 from test matrix, conflict in flake8 dependencies.

Refer to: python/importlib_metadata#406

Flake8 3.x uses a deprecated interface that was removed in
`importlib-metadata` 5 (python/importlib_metadata#406).

See: python/importlib_metadata#406

See python/importlib_metadata#406

See python/importlib_metadata#406 for explanation

* Cleanup whitespace in README.

* GitHub Actions: Add Py3.10, update Actions, more flake8 tests

"3.10" must be quoted in yaml https://dev.to/hugovk/the-python-3-1-problem-85g

* Test on Python 3.11 beta 5

* Reduce max line length in flake8 to the default of 79.

* Update authors and changelog.

* Update actions in codeql workflow.

* Fix INDICATOR keyword

* docs: Add a PiPY package badge

* Acknowledge support for Python 3.10

A subset of andialbrecht#645 
Follow on to andialbrecht#661

* DIV is Operator

* added some builtin types that can appear in Spark create table statements.

* Don't make slice copies in TokenList._token_matching().

Since we are working with indexes anyway, don't bother calling
enumerate() with a slice from self.tokens (which requires copying
memory).  Instead, just generate the indexes using range() and use
normal indexing to access the desired tokens.

The old behavior resulted in quadratic runtime with respect to the
number of tokens, which significantly impacted performance for
statements with very large numbers of tokens.

With the new behavior, the runtime is now linear with respect to the
number of tokens.

* Refactor to reduce redundant code.

* Add regex pattern to identify REGEXP as a Compasion token

* Make tzcast grouping function less eager

* Update authors and changelog.

* Add DISTINCTROW keyword (fixes andialbrecht#677).

* Update Changelog.

* CREATE TABLE tbl AS SELECT should return get_alias() for its column

* Update Changelog.

* Fixed bad parsing of create table statements that use lower case

* Update CHANGELOG.

yeah... I promise to make fewer commits like that in the future...

* add backtick to remove_quotes character list

* Update Changelog and authors.

* Add tests for utils.remove_quotes.

* Add docstring and comments.

* Allow any unicode character as identifier name (fixes andialbrecht#641).

* Bump version.

* Switch back to development mode.

* Revert "add regex pattern to identify IN as a Compasion token"

This reverts commit 28c4d40.

See andialbrecht#694. The expectation is that IN is primarily recognized as a keyword,
although it acts as a comparison operator. This also matches the definition of
IN in most SQL syntax references where it is listed as a reserved
keyword (PostgreSQL:
https://www.postgresql.org/docs/current/sql-keywords-appendix.html, MySQL:
https://dev.mysql.com/doc/refman/8.0/en/keywords.html, for example).

* Update changelog.

* Update workflow runner.

* Update tested Python versions in workflow.

* Setup a nightly build, even without changes in the module itself.

* Fix schedule trigger syntax.

* Switch to pyproject.toml (fixes andialbrecht#685).

* configurable syntax

* test configurable syntax

* remove type annotations for python 3.5 compatibility

* test for changing the regex

* lexer documentation

* flake8

* additional documentation

* change singleton behavior

* Fix get_type with comments between WITH keyword

* Cleanup regex for detecting keywords (fixes andialbrecht#709).

* CI: Use codecov action.

codecov module is deprecated and was removed from PyPI in
favor of the github action.

* Remove unnecessary parts in regex for bad escaping.

The regex tried to deal with situations where escaping in the
SQL to be parsed was suspicious.

* Testing branch

* Removed test file

* Bump version.

* Update Changelog.

* Switch back to development mode.

* Update python-app.yml: Try with 3.12-beta1

* Update python-app.yml: Revert to 3.12-dev and add check-latest

* Add classifier for Python 3.11 (fixes andialbrecht#726).

* Remove outdated and unused editorconfig.

* Drop support for Python 3.5.

* Get tox running again.

We have to pin versions for tox and virtualenv because newer versions
don't support Python 3.6 anymore.

* Don't pin virtualenv in dev section.

The requirements in dev section are more general (and may require newer versions).

* Add reminder for github release (fixes andialbrecht#732).

* Update issue templates

* Update bug_report.md

* Add link to discussion when creating issues.

* Update config.yml.

* Update test action.

* Add Code of Conduct.

* Add contributing guide.

* Create SECURITY.md

* Add Pull request template.

* Update Python version in test action.

* Updated too early... switching back to 3.12.0-beta.4.

* CI: Try to fix importlib issue with flake8.

See python/importlib_metadata#406

* Bump GitHub Actions

* Add support for Python 3.12

* Update changelog and authors.

* Ignore attributes starting with dunder in _TokenType (fixes andialbrecht#672).

This issue came up, when trying to deepcopy a parsed statement.
deepcopy uses getattr(obj, '__deepcopy__', None) to get a method
for copying an object. Before this change a new attribute
'__deepcopy__' was created as a new instance of _TokenType (a tuple).

* Simplify regex.

* Code cleanup.

---------

Co-authored-by: Andi Albrecht <[email protected]>
Co-authored-by: Christian Clauss <[email protected]>
Co-authored-by: osmnv <[email protected]>
Co-authored-by: Hao Xin <[email protected]>
Co-authored-by: Aki Ariga <[email protected]>
Co-authored-by: Simon Heisterkamp <[email protected]>
Co-authored-by: Daniel Harding <[email protected]>
Co-authored-by: JavierPan <[email protected]>
Co-authored-by: Erik Cederstrand <[email protected]>
Co-authored-by: Long Le Xich <[email protected]>
Co-authored-by: Simon Heisterkamp <[email protected]>
Co-authored-by: Shikanime Deva <[email protected]>
Co-authored-by: Kevin Stubbings <[email protected]>
Co-authored-by: Hugo van Kemenade <[email protected]>

Summary:

Update flake8 version to avoid a dependency error (see python/importlib_metadata#406). Also fix any outstanding flake8 issues.

The broken cpu/gpu tests are being fixed in D50850008 and D50707779.

Differential Revision: D50849894

Summary:
Pull Request resolved: #162

Update flake8 version to avoid a dependency error (see python/importlib_metadata#406). Also fix any outstanding flake8 issues.

The broken cpu/gpu tests are being fixed in D50850008 and D50707779.

Reviewed By: JKSenthil

Differential Revision: D50849894

fbshipit-source-id: 1bd8e961d42d24a5faaf9934343f52b0e2c70e80

…data#406

…age for ubuntu-latest

bump flake version to work around importlib issue python/importlib_metadata#406

Due to using annotations we can only support python versions from 3.7 onwards so it's possible to allow all pytest versions >= 7.1.0.
Step up flake8 to at least 4.0 to overcome the issue desribed at below link:
python/importlib_metadata#406

With Python 3.12 and 3.13, flake8 fails:

```
   eps = importlib_metadata.entry_points().get(self.namespace, ())
          ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
AttributeError: 'EntryPoints' object has no attribute 'get'
```

This is somehow related to PyCQA/flake8#1701
and python/importlib_metadata#406. We can't
use the former `import_metadata < 5` workaround any more, as flake8
imports this library from Python's stdlib (`importlib.metadata`), so the
external module gets ignored.

I don't know how to fix that. As the Python version that flake8 runs
with isn't very interesting, pin it to 3.11 for the time being.

* ci: Disable fast failing

This is annoying, as it serializes the errors of different python
versions.

* ci: Ignore https://data.safetycli.com/v/75180/f17/

We control which wheels we install (which is only setuptools), but we
don't control the pip version for the test, so ignore this. This
unbreaks the Python 3.6 test.

* ci: Drop support for Python 3.8

RHEL 8 has Python 3.6, RHEL 9 has Python ≥ 3.9, currently supported
Fedoras have ≥ 3.12. So none of the platforms we currently support have
3.8.

Dropping 3.8 is helpful because that is really hard to support wrt. the
`pkg_resources` → `importlib` migration. 3.8 doesn't yet have
`importlib.resources.files()`, which we need for that. Littering the
whole code with pylint and mypy overrides for this unknown API would be
error prone and ugly.

* fix: Prefer importlib.resources over pkg_resources

`pkg_resources` was deprecated in Python 3.12 and removed from 3.13 [1].
It still lives on in the `setuptools` package, but let's just move to
the current `importlib.resources` API, see the migration guide [2].

Still keep the `pkg_resources` code path as a fallback for Python 2.

[1] https://setuptools.pypa.io/en/latest/pkg_resources.html
[2] https://importlib-resources.readthedocs.io/en/latest/migration.html

* ci: Test Python 3.12 and 3.13

As the pkg_resources module is deprecated, the pylint env now needs an
explicit setuptools dependency to be able to check it.

Signed-off-by: Martin Pitt <[email protected]>

* ci: HACK: Pin flake8 python to 3.11

With Python 3.12 and 3.13, flake8 fails:

```
   eps = importlib_metadata.entry_points().get(self.namespace, ())
          ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
AttributeError: 'EntryPoints' object has no attribute 'get'
```

This is somehow related to PyCQA/flake8#1701
and python/importlib_metadata#406. We can't
use the former `import_metadata < 5` workaround any more, as flake8
imports this library from Python's stdlib (`importlib.metadata`), so the
external module gets ignored.

I don't know how to fix that. As the Python version that flake8 runs
with isn't very interesting, pin it to 3.11 for the time being.

---------

Signed-off-by: Martin Pitt <[email protected]>