vaeterchenfrost / tdvisu Goto Github PK

Documentation of API and other things.

1. Added git files
2. Added LICENSE+COPYING
3. Added README
4. Added Repo Description and Tags
5. Removed previous Urls/Dates from files
6. Added sourcecode
7. Added test folder
8. Added CHANGELOG
9. Added environment.yml for conda, frozen stable-requirements and conda_packages_list
10. Added setup.py
11. Removed all previous build-directories.

Use more property-based automated test-creation:
Example: https://github.com/SethMMorton/fastnumbers/blob/master/tests/test_fastnumbers.py

Docs: https://hypothesis.readthedocs.io/en/latest/usage.html

Arguments from the API are ready, but svgjoin only handles single arguments, no iterators for multiple joins. would be more convenient if handled logic there.

#6

Add python tests (pytest)

https://www.psycopg.org/docs/install.html

See https://www.psycopg.org/docs/install.html#build-prerequisites

for example on linux might need before completion:

sudo apt install libpq-dev

Right now incidence (+primal, dual) and general-graph are only either there or not.
Make the program more extensible with lists of bipartite, general graphs.

See zuo/unittest_expander#5
Fine for Python 3.8 still.

For e.g. MinVc it would be easier to see the subgraph if it would get an extra layout within an image.

Add a flag to the API to indicate that images for the subgraphs should be generated.

Clean up the logging config for the different modules.
Most up2date approach seems to be
https://docs.python.org/3/howto/logging.html#configuring-logging with YAML
https://docs.python.org/dev/library/logging.config.html#dictionary-schema-details

because we flagged the repo with 'Operating System :: OS Independent', we should make sure and document how to test it on the most common OS.

Yaml as of 1.2 is a full superset of JSON.
Give the user the choice, which file/data-format is best in each situation.

The dependency (PyYAML) is already specified in our python-benedict.

https://stackoverflow.com/questions/1773805/how-can-i-parse-a-yaml-file-in-python

Add test-coverage of python code for tdvisu/tdvisu.

Can regression-test more intermed steps.

Use functionality in folder \test

See https://changhsinlee.com/pytest-mock/ for mocking in pytest (testing database e.g.)

See also VaeterchenFrost/gpusat-VISU#37

Replacing unittest & unittest_expander
https://docs.pytest.org/en/latest/
https://docs.pytest.org/en/latest/reference.html#pytest-mark-parametrize

See also https://www.slant.co/versus/9148/9149/~unittest_vs_pytest

Called by: construct_dpdb_visu.py, visualization.py
Using: utilities.logging_cfg

Should be unified in the method.
Should not catch BaseException
Should provide better error handling

Update or add docstrings
Catch Errors early
README updates
CHANGELOG updates

Currently when creating output the invisible nodes are still considered as contributing to image size.

Should be possible to crop directly w/o post-processing.

Better splitting and order of translations is needed.

Unscaled:

• Moving center of image

After scaling:

• All translations get scaled too (as expected)

Todo: Sort translations after / before scaling or consider scaling in all translations after.

Use/add exposed parameters and a call to join images with svgjoin.