This document serves as a style guide for all jwql
software development. Any requested contribution to the jwql
code repository should be checked against this guide, and any violation of the guide should be fixed before the code is committed to
the master
or develop
branch. Please refer to the accompanying example.py
script for an example of code that abides by this style guide.
It is assumed that the reader of this style guide has read and is familiar with the following:
- The PEP8 Style Guide for Python Code
- The PEP257 Docstring Conventions Style Guide
- The
numpydoc
docstring convention
All software development for the jwql
project should follow a continuous integration workflow, described in the git
& GitHub workflow for contributing. Before committing any code changes, use flake8
to check the code against PEP8
standards. Also check that your code is conforming to this style guide.
Any changes pushed to the master
branch should be tagged with a version number. The version number convention is x.y.z
, where
x = The main version number. Increase when making incompatible API changes.
y = The feature number. Increase when change contains a new feature with or without bug fixes.
z = The hotfix number. Increase when change only contains bug fixes.
The following items should never be committed in the jwql
source code or GitHub issues/pull requests:
- Account credentials of any kind (e.g. database usernames and passwords)
- Internal directory structures or filepaths
- Machine names
- Proprietary data
If jwql
code needs to be aware of this information, it should be stored in a configuration file that is not part of the jwql
repository.
Additionally, developers of this project should be mindful of application security risks, and should adhere to the OWASP Top 10 as well as possible.
jwql
code shall adhere to the PEP8
conventions save for the following exceptions:
- Lines of code need not be restricted to 79 characters. However, it is encouraged to break up obnoxiously long lines into several lines if it benefits the overall readability of the code
Additionally, the code shall adhere to the following special guidelines:
- Function and class definitions should be placed in alphabetical order in the module
- It is encouraged to annotate variables and functions using the
typing
module (see PEP 483, PEP 484, and PEP 526). In addition, it is recommended that code be type-checked usingmypy
before a pull request is submitted.
jwql
code shall adhere to the PEP257
and numpydoc
conventions. The following are further recommendations:
- Each module should have at minimum a description, and
Authors
andUse
sections. - Each function/method should have at minimum a description, and
Parameters
(if necessary) andReturns
(if necessary) sections.
jwql
employs standards for logging monitoring scripts. See the logging guide
for further details.
To the extent possible, jwql
shall define frequently-used variable types/values consistently. A list of adopted standards is provided below:
-
JWST instrument names: In all internal references and structures (e.g. dictionaries) instrument names shall be all lower-case strings, e.g. one of
fgs
,miri
,niriss
,nircam
,nirspec
. When variations are required for interfaces, e.g.Nircam
for MAST,NIRCam
orNIRCAM
for SIAF, etc. these should be defined as dictionaries injwql/utils/constants.py
and imported from there. -
Program/proposal identifiers: JWST program IDs shall be stored and referred to internally as integers and parsed to strings only when needed. For example, the inputs
"001144"
and"1144"
shall both be converted to an integer variable with value1144
.
jwql
strives to create web pages and code that are as accessibile as possible for all users. See the accessibility guidelines
page for further details.
argparse
for parsing command line argumentsbokeh
for interactive plotting