ShotGrid provides a simple Python-based API for accessing ShotGrid and integrating with other tools. This is the official API that is maintained by ShotGrid Software (https://knowledge.autodesk.com/contact-support)
The latest version can always be found at http://github.com/shotgunsoftware/python-api
- ShotGrid server v2.4.12+.
- Python v2.7 or v3.7
Tutorials and detailed documentation about the Python API are available at http://developer.shotgridsoftware.com/python-api).
Some useful direct links:
You can see the full history of the Python API on the documentation site.
The API comes with a copy of the httplib2
inside the shotgun_api3/lib
folder. To update the copy to a more recent version of the API, you can run the update_httplib2.py
script at the root of this repository like this:
python update_httplib2.py vX.Y.Z
where vX.Y.Z
is a release found on httplib2
's release page.
python-api should remain compatible with both Python 2, and 3. To make this easier, we use six. When adding code that works with types that have changed between Python 2 and 3, notably strings and files, it's advisable to use the six
types for casting and comparisons. Be sure to follow Python 2 and 3 compatible conventions in code, especially when raising or capturing exceptions and printing. While we don't use future
, this page contains a fairly comprehensive list of Python 2/3 compatibility sticking points to look out for.
Additionally, the python-modernize tool can be helpful when updating Python 2 code for Python 3 compatibility.
Python 2:
if isinstance(my_variable, str):
Python 2/3:
if isinstance(my_variable, six.string_types):
Python 2:
except SomeExceptionType, e:
print "I like to swallow exceptions!"
Python 2/3:
from __future__ import print_function
except SomeExceptionType as e:
print("I like to swallow exceptions!")
Python 2:
print "My spoon is too big!"
Python 2/3:
from __future__ import print_function
print("My spoon is too big!")
Additionally, when testing locally, tests should be run for both python 2 and python 3 to ensure changes won't break cross-compatibility.
Integration and unit tests are provided.
- All tests require:
- The nose unit testing tools,
- The nose-exclude nose plugin
- (Note: Running
pip install -r tests/ci_requirements.txt
will install this package)
- A
tests/config
file (you can copy an example fromtests/example_config
). - Tests can be run individually like this:
nosetests --config="nose.cfg" tests/test_client.py
- Make sure to not forget the
--config="nose.cfg"
option. This option tells nose to use our config file. This will exclude python 2- and 3-specific files in the/lib
directory, preventing a failure from being reported by nose for compilation due to incompatible syntax in those files.
- Make sure to not forget the
test_client
andtests_unit
use mock server interaction and do not require a ShotGrid instance to be available (no modifications totests/config
are necessary).test_api
andtest_api_long
do require a ShotGrid instance, with a script key available for the tests. The server and script user values must be supplied in thetests/config
file. The tests will add test data to your server based on information in your config. This data will be manipulated by the tests, and should not be used for other purposes.- To run all of the tests, use the shell script
run-tests
.
- Update the Changelog in the
HISTORY.rst
file- Add bullet points for any changes that have happened since the previous release. This may include changes you did not make so look at the commit history and make sure we don't miss anything. If you notice something was done that wasn't added to the changelog, hunt down that engineer and make them feel guilty for not doing so. This is a required step in making changes to the API.
- Try and match the language of previous change log messages. We want to keep a consistent voice.
- Make sure the date of the release matches today. We try and keep this TBD until we're ready to do a release so it's easy to catch that it needs to be updated.
- Make sure the version number is filled out and correct. We follow semantic versioning.
- Ensure any changes or additions to public methods are documented
- Ensure that doc strings are updated in the code itself to work with Sphinx and are correctly formatted.
- Examples are always good especially if this a new feature or method.
- Think about a new user to the API trying to figure out how to use the features you're documenting.
- Update the version value in
python-api/setup.py
to match the version you are packaging. This controls what version users will get when installing via pip. - Update the
__version__
value inshotgun_api3/shotgun.py
to the version you're releasing. This identified the current version within the API itself. - Commit these changes in master with a commit message like
packaging for the vx.x.x release
. - Create a tag based off of the master branch called
vx.x.x
to match the version number you're releasing. - Push master and your tag to Github.
- Update the Releases page with your new release.
- The release should already be there from your tag but if not, create a new one.
- Add more detailed information regarding the changes in this release. This is a great place to add examples, and reasons for the change!
Post a message in the Pipeline Community channel and send an email to shotgun-dev with a link to the community post.
- Update the
__version__
value inshotgun_api3/shotgun.py
to the next version number with.dev
appended to it. For example,v3.0.24.dev
- Add a new section to the Changelog in the
HISTORY.rst
file with the next version number and a TBD date
**v3.0.24 - TBD**
+ TBD
- Commit the changes to master with a commit message like
Bump version to v3.0.24.dev
- Push master to Github