[GSK-1504] Integration with W&B

[rabah-khalek]

Integration

Prerequisites

1. Create a wandb account here.
2. Open your docker app
3. Install python package and server

   pip install wandb
   wandb login --relogin # input the API key you get from the website
   wandb server start --upgrade # this will download the docker images if they're not already downloaded

Features

import giskard, wandb
# [...] wrap model and dataset with giskard
scan_results = giskard.scan(giskard_model, giskard_dataset)
test_suite_results = scan_results.generate_test_suite().run()
shap_results = giskard.explain_with_shap(giskard_model, giskard_dataset)

wandb.login()
giskard_dataset.to_wandb() # log dataset
scan_results.to_wandb() # log scan results
test_suite_results.to_wandb() # log test suite results
shap_results.to_wandb() # log shap results as plots 

Todo

• log scan results (@rabah-khalek)
• log test suite results (@AbSsEnT)
• log dataset artifact (@AbSsEnT)
• think of other objects we can log (@AbSsEnT and @rabah-khalek)
  • log shapely values as interactive graphs (@AbSsEnT)
• write tests (@rabah-khalek and @AbSsEnT)
• write doc (@rabah-khalek and @AbSsEnT)
  • write 1 demo notebook (credit scoring) (@AbSsEnT)
• write article once PR is accepted (@AbSsEnT and @rabah-khalek)

[linear]

GSK-1505 Exploring the tool

• mykyta go through https://docs.wandb.ai/quickstart and explore a bit the platform
• mykyta go through this example: https://docs.deepchecks.com/stable/general/usage/exporting_results/auto_examples/plot_exports_output_to_wandb.html#exporting-a-suite-s-output-suiteresult-to-wandb
• once you're familiar with it, let's discuss https://docs.wandb.ai/guides/integrations

[linear]

GSK-1504 Integration with W&B

[andreybavt]

Generally speaking looks good to me, I left a few non-major comments

[andreybavt]

LGTM

[sonarqubecloud]

Kudos, SonarCloud Quality Gate passed!   

0 Bugs
0 Vulnerabilities
0 Security Hotspots
1 Code Smell

91.6% Coverage
0.0% Duplication