Scheduled weekly dependency update for week 22

[pyup-bot]

Update sphinxcontrib-bibtex from 2.2.1 to 2.2.1.

Changelog

2.2.1

-------------------

* The LaTeX output now uses hyperlink instead of sphinxcite. This fixes
issues with double brackets and other mismatches between LaTeX and
HTML outputs (see issue 244 reported by zhi-wang).

* The setup function now also returns the version of the extension (see
issue 239 reported by lcnittl).

2.2.0

--------------------

* Support the ``:any:`` role (see issue 232).

* New natbib/biblatex inspired roles for textual and parenthetical
citation references (see issue 203 reported by matthew-brett).
For textual citation references, use ``:cite:t:``
and for parenthetical citation references, use ``:cite:p:``.
The old ``:cite:`` role is an alias for ``:cite:p:``.

* Use the ``s`` suffix to include the full author list
rather than abbreviating it with "et al.":
``:cite:ts:``, ``:cite:ps:``.

* For textual citation references,
use the ``c`` prefix to capitalize the first letter:
``:cite:ct:``, ``:cite:cts:``.

* New natbib inspired roles for citing
just the author, year, or label, optionally with brackets,
and optionally capitalizing the first letter of the author:
``:cite:author:``, ``:cite:authorpar:``,
``:cite:cauthor:``, ``:cite:cauthorpar:``
``:cite:year:``, ``:cite:yearpar:``,
``:cite:label:``, ``:cite:labelpar:``
(see issue 71 reported by bk322).

* To configure your referencing style,
an optional config setting ``bibtex_reference_style`` has been added.
If not specified, this defaults to the ``label`` style,
which will use the label to format citation references, matching the
referencing style as in previous versions.
The other style currently available is ``author_year``, for author-year
style referencing.

* Reference styles can be fully customized to your heart's desire
(see issue 203 reported by amichuda).
They are based on pybtex's template system, which was already used for
customizing bibliography styles.
Refer to the user documentation for examples, and to the API documentation
for full details.

* Other packages can register custom reference styles through entry points.
Refer to the user documentation for details.

* Propagate pybtex FieldIsMissing exception as a warning (see issue
235 reported by Zac-HD).

2.1.4

----------------------

* Fix ValueError exception when having citations from orphans (see issue 228,
reported by VincentRouvreau).

2.1.3

----------------------

* Sphinx 2.1 or later is now formally required (up from 2.0).

* Fix unresolved references when running the latex build immediately after
the html build, or when rerunning the html build after deleting the
generated html files without deleting the pickled doctrees/environment
(see issue 226, reported by skirpichev).

* No longer insert user defined header for bibliography directives if there are
no citations in it.

* Warnings now consistently provide source file and line number of where the
issue originated.

* Simpler and faster implementation of footcite and footbibliography.

* Improved type annotations throughout the API, now using forward
declarations where possible.

2.1.2

------------------------

* Fix KeyError exception when building documents with footbibliography
directives but without any footnotes needing to be generated for this
directive (see issue 223, reported by drammock).

2.1.1

------------------------

* Fix latex builder KeyError exception (see issue 221, reported by jedbrown).

* Fix citation references across documents in latex build.

2.1.0

------------------------

* The extension no longer relies on the ``bibtex.json`` method. Instead, the
extension now postpones identifying all citation cross-references to
Sphinx's consistency check phase.
The actual citation references and bibliography citations
are then generated in the resolve phase using post-transforms.
As a result, ``bibtex.json`` is no longer needed and thus
Sphinx no longer needs to run twice as in the past if the file did not exist
(closes issues 214 and 215).
*Thanks to everyone who chimed in on this, especially everyone who
made helpful suggestions to find better implementation approaches,
and everyone who helped with testing.*

* Citations with multiple keys will now reside in the same bracket
(closes issue 94).

* Consistent use of doctutils note_explicit_target to set ids, to ensure no
clashing ids.

* Improved and robustified test suite, using regular expressions to verify
generated html.

* The test suite now includes a patched version of the awesome but abandoned
sphinx-natbib extension, to help comparing and testing implementations and
features.
The long term intention is to fully support sphinx-natbib style citations.

* **BACKWARD INCOMPATIBLE**
The API has been refactored to accommodate the new design.
Refer to the API documentation for details.

2.0.0

------------------------

* There is a new ``footcite`` role and a new ``footbibliography``
directive, to allow easy and simple local (per document)
bibliographies through footnotes.
See issues 184 and 185.

* Parallel builds are now finally supported.
See issues 80, 96, and 164, as well as pull request 210.

* **BACKWARD INCOMPATIBLE**
To enable parallel builds, a new mandatory
config setting ``bibtex_bibfiles`` has been added. This setting
specifies all bib files used throughout the project,
relative to the source folder.

* **BACKWARD INCOMPATIBLE**
The encoding of bib files has been moved to an optional
config setting ``bibtex_encoding``. The ``:encoding:``
option is no longer supported.

* Headers for ``bibliography`` and ``footbibliography`` directives
can be configured via the ``bibtex_bibliography_header`` and
``bibtex_footbibliography_header`` config setting.

* The ``bibliography`` directive no longer requires the bib files
to be specified as an argument. However, if you do, citations will
be constrained to those bib files.

* Support newlines/whitespace around cite keys when multiple keys are
specified.
Thanks to dizcza for help with testing.
See issue 205 and pull request 206.

* Improve citation ordering code (reported by ukos-git, see issue
182).

* The unresolved citations across documents issue has been resolved.
The extension stores all citation information in a ``bibtex.json`` file.
If it does not exist, the file will be created on
your first sphinx build, and you will have to rerun the build
to make use of it. The file is automatically kept up to date,
with a warning whenever you need to rerun the build.
Thanks to dizcza for help with testing.
See issues 197 and 204. Also see pull request 208.

* Migrate test suite to pytest, using sphinx's testing fixtures.

* **BACKWARD INCOMPATIBLE**
The API has been refactored.
Some functions have moved to different modules.
Refer to the API documentation for details.

* Drop Python 3.5 support.

* Add Python 3.9 support.

1.0.0

-------------------------

* Drop Python 2.7 and 3.4 support (as upstream sphinx has dropped
support for these as well).

* Add Python 3.8 support (contributed by hroncok).

* Update for Sphinx 2.x, and drop Sphinx 1.x support (as there is too
much difference between the two versions).

* Non-bibtex citations will now no longer issue warnings (fix
contributed by chrisjsewell).

* Switch to codecov for coverage reporting.

0.4.2

----------------------

* Drop Python 3.3 support, add Python 3.7 support.

* Work around issue with sphinx-testing on Fedora (reported by
jamesjer in issue 157, fix contributed by mitya57 in pull request
158).

0.4.1

------------------------

* Disable tinkerer test due to upstream bug.

* Remove crossref test due to changed upstream behaviour in pybtex.

* Fix latex test to match new upstream code generation.

* Fix documentation of encoding option (contributed by Kai Mühlbauer).

* Migrate to sphinx.util.logging in favour of old deprecated logging
method.

0.4.0

---------------------

* Remove latexcodec and curly bracket strip functionality, as this is
now supported by pybtex natively (see issue 127, reported by
erosennin).

* Fix tests failures with Sphinx 1.7 (see pull request 136, reported
and fixed by mitya57).

0.3.6

-------------------------

* Real fix for issue 111 (again reported by jamesjer).

* Fix test regressions due to latest Sphinx updates (see issues 115,
120, 121, and 122, reported by ndarmage and ghisvail).

* Fix test regressions on ascii locale (see issue 121, reported by
ghisvail).

* Support and test Python 3.6.

0.3.5

------------------------

* Fix extremely high memory usage when handling large bibliographies
(reported by agjohnson, see issue 102).

* Fix tests for Sphinx 1.5.1 (see issue 111, reported by jamesjer).

0.3.4

-------------------

* Document LaTeX workaround for ``:cite:`` in figure captions
(contributed by xuhdev, see issue 92 and pull request 93).

* Add ``bibtex_default_style`` config value to override the default
bibliography style (see issue 91 and pull request 97).

* Support Python 3.5 (see issue 100).

0.3.3

-----------------------

* Add per-bibliography key prefixes, enabling local bibliographies to
be used in isolation from each other (see issue 87, reported by
marscher).

* Documentation now points to new location of pybtex on bitbucket.

* Simplified testing code by using the new sphinx_testing package.

0.3.2

---------------------

* Document how to create custom label styles (see issue 77, reported
by tino).

* Disable parallel_read_safe for Sphinx 1.3 and later (see issue 80,
reported by andreacassioli).

0.3.1

--------------------

* Fix for ``type_.lower()`` bug: pybtex 0.18 expects type to be a
string (this fixes issue 68 reported by jluttine).

0.3.0

------------------

* **BACKWARD INCOMPATIBLE**
The alpha style is now default, so citations are labelled in a way
that is more standard for Sphinx. To get the old behaviour back, add
``:style: plain`` to your bibliography directives.

* **BACKWARD INCOMPATIBLE**
:meth:`~sphinxcontrib.bibtex.cache.Cache.is_cited` has been removed.
Use :meth:`~sphinxcontrib.bibtex.cache.Cache.get_cited_docnames` instead,
which will return an empty list for keys that are not cited.

* Improved support for local bibliographies (see issues 52, 62, and
63; test case provided by Boris Kheyfets):

- New ``docname`` and ``docnames`` filter identifiers.

- Filter expressions now also support set literals and the operators
 ``in``, ``not in``, ``&``, and ``|``.

See documentation for details.

* Multiple comma-separated citation keys per cite command (see issue
61, suggested by Boris Kheyfets).

* Add support for pypy and Python 3.4.

* Drop support for Python 2.6 and Python 3.2.

* Drop 2to3 and instead use six to support both Python 2 and 3 from a
single code base.

* Simplify instructions for custom styles.

* Various test suite improvements.

0.2.9

----------------------

* Upgrade to the latest pybtex-docutils to produce more optimal html output
(specifically: no more nested ``<span>``\ s).

* Remove latex codec code, and rely on latexcodec package instead.

* :class:`FilterVisitor` has been removed from the public API.
Use :meth:`~sphinxcontrib.bibtex.cache.Cache.get_bibliography_entries`
instead.

* Fix upstream Sphinx bug concerning LaTeX citation hyperlinks
(contributed by erikb85; see pull request 45).

* Fix most pylint warnings, refactor code.

0.2.8

---------------------

* Use pybtex-docutils to remove dependency on pybtex.backends.doctree.

0.2.7

---------------------

* Integrate with coveralls.io, first release with 100% test coverage.

* Minor bug fixes and code improvements.

* Remove ordereddict dependency for Python 2.7 and higher (contributed
by Paul Romano, see pull requests 27 and 28).

* New ``:filter:`` option for advanced filtering (contributed by
d9pouces, see pull requests 30 and 31).

* Refactor documentation of advanced features.

* Document how to create custom pybtex styles (see issues 25, 29,
and 34).

* Code is now mostly pep8 compliant.

0.2.6

--------------------

* For unsorted styles, citation entries are now sorted in the order
they are cited, instead of following the order in the bib file, to
reflect more closely the way LaTeX handles unsorted styles
(addresses issue 15).

* Skip citation label warnings on Sphinx [source] links (issue 17,
contributed by Simon Clift).

0.2.5

-----------------------

* Duplicate label detection (issue 14).

* New ``:labelprefix:`` option to avoid duplicate labels when having
multiple bibliographies with a numeric label style (addresses
issue 14).

0.2.4

----------------------

* New options for the bibliography directive for rendering the
bibliography as bullet lists or enumerated lists: ``:list:``,
``:enumtype:``, and ``:start:``.

* Minor latex codec fixes.

* Turn exception into warning when a citation cannot be relabeled
(fixes issue 2).

* Document LaTeX encoding, and how to turn it off (issue 4).

* Use pybtex labels (fixes issue 6 and issue 7).

* Cache tracked citation keys and labels, and bibliography enumeration
counts (fixes issues with citations in repeated Sphinx runs).

* Bibliography ids are now unique across documents (fixes issue that
could cause the wrong bibliography to be inserted).

* The plain style is now the default (addresses issue 9).

0.2.3

--------------------

* Document workaround for Tinkerer (issue 1).

* Use tox for testing.

* Full 2to3 compatibility.

* Document supported versions of Python (2.6, 2.7, 3.1, and 3.2).

0.2.2

-------------------

* Documentation and manifest fixes.

0.2.1

--------------------

* First public release.

Links

• PyPI: https://pypi.org/project/sphinxcontrib-bibtex
• Changelog: https://pyup.io/changelogs/sphinxcontrib-bibtex/
• Repo: https://github.com/mcmtroffaes/sphinxcontrib-bibtex

Update six from 1.16.0 to 1.16.0.

Changelog

1.16.0

------

- Pull request 343, issue 341, pull request 349: Port _SixMetaPathImporter to
Python 3.10.

1.15.0

------

- Pull request 331: Optimize `six.ensure_str` and `six.ensure_binary`.

1.14.0

------

- Issue 288, pull request 289: Add `six.assertNotRegex`.

- Issue 317: `six.moves._dummy_thread` now points to the `_thread` module on
Python 3.9+. Python 3.7 and later requires threading and deprecated the
`_dummy_thread` module.

- Issue 308, pull request 314: Remove support for Python 2.6 and Python 3.2.

- Issue 250, issue 165, pull request 251: `six.wraps` now ignores missing
attributes. This follows the Python 3.2+ standard library behavior.

1.13.0

------

- Issue 298, pull request 299: Add `six.moves.dbm_ndbm`.

- Issue 155: Add `six.moves.collections_abc`, which aliases the `collections`
module on Python 2-3.2 and the `collections.abc` on Python 3.3 and greater.

- Pull request 304: Re-add distutils fallback in `setup.py`.

- Pull request 305: On Python 3.7, `with_metaclass` supports classes using PEP
560 features.

1.12.0

------

- Issue 259, pull request 260: `six.add_metaclass` now preserves
`__qualname__` from the original class.

- Pull request 204: Add `six.ensure_binary`, `six.ensure_text`, and
`six.ensure_str`.

1.11.0

------

- Pull request 178: `with_metaclass` now properly proxies `__prepare__` to the
underlying metaclass.

- Pull request 191: Allow `with_metaclass` to work with metaclasses implemented
in C.

- Pull request 203: Add parse_http_list and parse_keqv_list to moved
urllib.request.

- Pull request 172 and issue 171: Add unquote_to_bytes to moved urllib.parse.

- Pull request 167: Add `six.moves.getoutput`.

- Pull request 80: Add `six.moves.urllib_parse.splitvalue`.

- Pull request 75: Add `six.moves.email_mime_image`.

- Pull request 72: Avoid creating reference cycles through tracebacks in
`reraise`.

1.10.0

------

- Issue 122: Improve the performance of `six.int2byte` on Python 3.

- Pull request 55 and issue 99: Don't add the `winreg` module to `six.moves`
on non-Windows platforms.

- Pull request 60 and issue 108: Add `six.moves.getcwd` and
`six.moves.getcwdu`.

- Pull request 64: Add `create_unbound_method` to create unbound methods.

1.9.0

-----

- Issue 106: Support the `flush` parameter to `six.print_`.

- Pull request 48 and issue 15: Add the `python_2_unicode_compatible`
decorator.

- Pull request 57 and issue 50: Add several compatibility methods for unittest
assertions that were renamed between Python 2 and 3.

- Issue 105 and pull request 58: Ensure `six.wraps` respects the *updated* and
*assigned* arguments.

- Issue 102: Add `raise_from` to abstract out Python 3's raise from syntax.

- Issue 97: Optimize `six.iterbytes` on Python 2.

- Issue 98: Fix `six.moves` race condition in multi-threaded code.

- Pull request 51: Add `six.view(keys|values|items)`, which provide dictionary
views on Python 2.7+.

- Issue 112: `six.moves.reload_module` now uses the importlib module on
Python 3.4+.

1.8.0

-----

- Issue 90: Add `six.moves.shlex_quote`.

- Issue 59: Add `six.moves.intern`.

- Add `six.urllib.parse.uses_(fragment|netloc|params|query|relative)`.

- Issue 88: Fix add_metaclass when the class has `__slots__` containing
`__weakref__` or `__dict__`.

- Issue 89: Make six use absolute imports.

- Issue 85: Always accept *updated* and *assigned* arguments for `wraps()`.

- Issue 86: In `reraise()`, instantiate the exception if the second argument is
`None`.

- Pull request 45: Add `six.moves.email_mime_nonmultipart`.

- Issue 81: Add `six.urllib.request.splittag` mapping.

- Issue 80: Add `six.urllib.request.splituser` mapping.

1.7.3

-----

- Issue 77: Fix import six on Python 3.4 with a custom loader.

- Issue 74: `six.moves.xmlrpc_server` should map to `SimpleXMLRPCServer` on Python
2 as documented not `xmlrpclib`.

1.7.2

-----

- Issue 72: Fix installing on Python 2.

1.7.1

-----

- Issue 71: Make the six.moves meta path importer handle reloading of the six
module gracefully.

1.7.0

-----

- Pull request 30: Implement six.moves with a PEP 302 meta path hook.

- Pull request 32: Add six.wraps, which is like functools.wraps but always sets
the __wrapped__ attribute.

- Pull request 35: Improve add_metaclass, so that it doesn't end up inserting
another class into the hierarchy.

- Pull request 34: Add import mappings for dummy_thread.

- Pull request 33: Add import mappings for UserDict and UserList.

- Pull request 31: Select the implementations of dictionary iterator routines
at import time for a 20% speed boost.

1.6.1

-----

- Raise an AttributeError for six.moves.X when X is a module not available in
the current interpreter.

1.6.0

-----

- Raise an AttributeError for every attribute of unimportable modules.

- Issue 56: Make the fake modules six.moves puts into sys.modules appear not to
have a __path__ unless they are loaded.

- Pull request 28: Add support for SplitResult.

- Issue 55: Add move mapping for xmlrpc.server.

- Pull request 29: Add move for urllib.parse.splitquery.

1.5.2

-----

- Issue 53: Make the fake modules six.moves puts into sys.modules appear not to
have a __name__ unless they are loaded.

1.5.1

-----

- Issue 51: Hack around the Django autoreloader after recent six.moves changes.

1.5.0

-----

- Removed support for Python 2.4. This is because py.test no longer supports
2.4.

- Fix various import problems including issues 19 and 41. six.moves modules
are now lazy wrappers over the underlying modules instead of the actual
modules themselves.

- Issue 49: Add six.moves mapping for tkinter.ttk.

- Pull request 24: Add __dir__ special method to six.moves modules.

- Issue 47: Fix add_metaclass on classes with a string for the __slots__
variable.

- Issue 44: Fix interpretation of backslashes on Python 2 in the u() function.

- Pull request 21: Add import mapping for urllib's proxy_bypass function.

- Issue 43: Add import mapping for the Python 2 xmlrpclib module.

- Issue 39: Add import mapping for the Python 2 thread module.

- Issue 40: Add import mapping for the Python 2 gdbm module.

- Issue 35: On Python versions less than 2.7, print_ now encodes unicode
strings when outputting to standard streams. (Python 2.7 handles this
automatically.)

1.4.1

-----

- Issue 32: urllib module wrappings don't work when six is not a toplevel file.

1.4.0

-----

- Issue 31: Add six.moves mapping for UserString.

- Pull request 12: Add six.add_metaclass, a decorator for adding a metaclass to
a class.

- Add six.moves.zip_longest and six.moves.filterfalse, which correspond
respectively to itertools.izip_longest and itertools.ifilterfalse on Python 2
and itertools.zip_longest and itertools.filterfalse on Python 3.

- Issue 25: Add the unichr function, which returns a string for a Unicode
codepoint.

- Issue 26: Add byte2int function, which complements int2byte.

- Add a PY2 constant with obvious semantics.

- Add helpers for indexing and iterating over bytes: iterbytes and indexbytes.

- Add create_bound_method() wrapper.

- Issue 23: Allow multiple base classes to be passed to with_metaclass.

- Issue 24: Add six.moves.range alias. This exactly the same as the current
xrange alias.

- Pull request 5: Create six.moves.urllib, which contains abstractions for a
bunch of things which are in urllib in Python 3 and spread out across urllib,
urllib2, and urlparse in Python 2.

1.3.0

-----

- Issue 21: Add methods to access the closure and globals of a function.

- In six.iter(items/keys/values/lists), passed keyword arguments through to the
underlying method.

- Add six.iterlists().

- Issue 20: Fix tests if tkinter is not available.

- Issue 17: Define callable to be builtin callable when it is available again
in Python 3.2+.

- Issue 16: Rename Python 2 exec_'s arguments, so casually calling exec_ with
keyword arguments will raise.

- Issue 14: Put the six.moves package in sys.modules based on the name six is
imported under.

- Fix Jython detection.

- Pull request 4: Add email_mime_multipart, email_mime_text, and
email_mime_base to six.moves.

1.2.0

-----

- Issue 13: Make iterkeys/itervalues/iteritems return iterators on Python 3
instead of iterables.

- Issue 11: Fix maxsize support on Jython.

- Add six.next() as an alias for six.advance_iterator().

- Use the builtin next() function for advance_iterator() where is available
(2.6+), not just Python 3.

- Add the Iterator class for writing portable iterators.

1.1.0

-----

- Add the int2byte function.

- Add compatibility mappings for iterators over the keys, values, and items of a
dictionary.

- Fix six.MAXSIZE on platforms where sizeof(long) != sizeof(Py_ssize_t).

- Issue 3: Add six.moves mappings for filter, map, and zip.

1.0

----------

- Reworked six into one .py file.  This breaks imports.  Please tell me if you
are interested in an import compatibility layer.

1.0.0

-----

- Issue 2: u() on Python 2.x now resolves unicode escapes.

- Expose an API for adding mappings to six.moves.

Links

• PyPI: https://pypi.org/project/six
• Changelog: https://pyup.io/changelogs/six/
• Repo: https://github.com/benjaminp/six

[coveralls]

Coverage remained the same at 3.976% when pulling a7861d2 on pyup-scheduled-update-2021-05-31 into bed56ab on master.

[pyup-bot]

Closing this in favor of #129