-
chore: allow more_itertools>=8.10 (
1b8366c
) -
chore: fix the release process (
c044bdd
) -
chore: remove unnecessary shebang lines (
adccaf6
) -
chore: fix typo in test error message (
fe495bd
)
-
feat: add a schema function to rename parameters (
de5a71e
) -
feat: save decorator arguments as function attributes (
367dff2
)
-
chore: upgrade to latest readthedocs config format (
74ba8ac
) -
chore: ignore deprecation warnings caused by datetime/arrow (
77b4216
) -
chore: switch to gitlint (
f5ea748
) -
chore: update ruff command (
1aac862
) -
chore: fix PyPI link to documentation (
4448eb7
) -
chore: add linting; update semantic release (
bbd5a87
)
- feat: improve error message (
e910232
)
- fix: handle class-level parametrization correctly
Fixes #21 (24f8f13
)
- 0.19.0 (
69ef36a
)
-
docs: update examples to use 'import parametrize_from_file as pff' (
794806b
) -
docs: fix typos (
7681380
) -
docs: fix syntax highlighting (
fbc1e14
) -
docs: make README example more intuitive (
277ba97
)
- feat: allow schema functions to add marks (
f1f0c2b
)
- fix: improve error messages (
3a92dbf
)
- Merge branch 'master' of github.com:kalekundert/parametrize_from_file (
c701520
)
- chore: replace contextlib2 with contextlib (#18)
contextlib.nullcontext was added in Python 3.7 which is the oldest supported version (9c68f20
)
- test: account for changed module name of
nt.load()
(1fb2b7b
)
- Merge branch 'master' of github.com:kalekundert/parametrize_from_file (
309056d
)
- feat: allow multiple cast functions for each field (
c026e8b
)
- feat: don't eval/exec the error() context managers (
9d8b2af
)
- feat: provide schema functions tailored for testing
-
Add the
cast
anddefaults
schema functions. -
Refactor
error
anderror_or
into standalone functions that don't depend on theNamespace
class. The correspondingNamespace
methods remain, but are now just thin wrappers. -
Interpret iterable schema arguments as pipelines.
-
Get rid of the voluptuous dependency.
Fixes #16 (275c79d
)
-
feat: allow loaders to be overridden locally (
466fa31
) -
feat: allow preprocess() to get additional contextual information (
a482725
)
- Merge branch 'master' of github.com:kalekundert/parametrize_from_file (
5cda235
)
-
docs: add a tutorial on temporary files (
09aa784
) -
docs: fix pytest cross-references (
33580bb
) -
docs: tweak wording (
f791460
)
- fix: improve error messages (
c1eaa78
)
- test: add pytest_tmp_files dependency (
bc71d5e
)
- feat: allow assertions on the direct causes of exceptions (
d4d2be9
)
- chore: treat all warnings as errors in CI (
97f215c
)
- feat: allow deferred eval/exec to be invoked by name
I think this makes the code easier to read. (08d0bf2
)
- fix: migrate to more_itertools.zip_broadcast()
This avoids a warning in python>=3.10
Fixes #13 (cdf37c3
)
- feat: add empty_ok() for voluptuous schema (
6fd446d
)
- feat: get multiple variables from executed snippets (
c43c57a
)
- fix: improve debug message (
08fa038
)
- feat: try to import the voluptuous submodule (
4d5cca4
)
- feat: allow exec() and eval() to be deferred
This commit also gets rid of exec_and_lookup(), which was basically just
doing one specific kind of deferral. Fixes #8. (b8cf9d5
)
- feat: make namespaces immutable, add more error checks
- Make namespaces immutable, fix #10.
- Provide more flexible exception assertions, fix #9.
- Add sensible
__bool__()
implementation for error context managers, fix #12. (8e52066
)
- test: make sure namespaces with unpickleable values can be copied (
1c61ee8
)
- fix: include voluptuous dependency (
62fc60b
)
- docs: tweak wording (
f3be56a
)
- feat: support eval/exec'ing mocks (
121995a
)
- feat: support parametrized fixtures
Fixes #6 (e9a95c4
)
- feat: support indirect parametrization
Fixes #7 (34308c4
)
- test: add cases that break mi.zip_broadcast()
See more-itertools/more-itertools#561 (980dd3b
)
- fix: correctly format error messages with braces (
7caaa31
)
-
docs: add examples of my projects that use this package (
f1e59f3
) -
docs: install dependencies (
90721fd
) -
docs: install dependencies (
6ca7ae8
)
-
feat: allow multiple paths/keys to be specified (
fe6696a
) -
feat: add the load_parameters() function
I don't have a specific use for this API, but it seems like a good thing
to have. It also forced me to refactor the code a bit to better
separate responsibilities. (5c8af43
)
- feat: add the preprocess argument (
219aa9f
)
- fix: check that the schema returns a dict (
85ecdf1
)
- Merge branch 'master' of github.com:kalekundert/parametrize_from_file (
7ef4818
)
- docs: slightly clarify error messages (
7393ee0
)
- feat: add the Namespace eval/exec/error helper (
e0c8523
)
- Merge branch 'master' of github.com:kalekundert/parametrize_from_file (
d9e4a06
)
- chore: remove stale imports (
707ac13
)
- feat: accept any callable as the schema
Fixes #1 (2f3d1e1
)
- feat: rename the 'key' argument (
25d7879
)
-
chore: add support for python 3.6 (
dc8f385
) -
chore: install the local version of this package (
e539e81
) -
chore: remove this package from the requirements file (
cf80457
) -
chore: try using a requirements file to satisfy ReadTheDocs (
fa569c8
) -
chore: specify which version of arrow to use when building docs (
4ca3325
) -
chore: add dateutil as doc dependency (
a016b2e
) -
chore: remove autoclasstoc doc dependency (
dc0aefc
)
- fix: don't apply the schema to 'id' and 'marks' (
e18e931
)
- refactor: the Params class is outside the scope of this module (
2b5a638
)
- Merge branch 'master' of github.com:kalekundert/parametrize_from_file (
8768ac2
)
- docs: initial documentation (
597abeb
)
- feat: give more helpful schema errors (
aac84a0
)
- chore: use 'toml' instead of 'rtoml'
'rtoml' is faster and more robust, but harder to get working in CI
environments because it requires rust to be installed. Since this
package is meant to be used for testing, I don't want to add that
hurdle. (fa96633
)