CPM filterlist experiment

[muodov]

Task/Issue URL: https://app.asana.com/0/1201621853593513/1207940672701584/f
Tech Design URL:
CC:

Description:
This adds a CPM filterlist experiment as described in Asana

Steps to test this PR:

1. Check out this branch and the corresponding BSK branch with the same name
2. Build
3. Enable Autoconsent logs in Debug menu, make sure that CPM feature is enabled in user settings
4. Point to a custom config with the new CPM subfeature:

        "autoconsent": {
            "state": "enabled",
            "features": {
                "filterlistExperiment": {
                    "state": "enabled",
                    "rollout": {
                        "steps": [
                            {
                                "percent": 100
                            }
                        ]
                    }
                }
            }
        },

5. Search for "buy shoes" on DDG. Verify in logs that the experiment cohort is picked ("fc" for control, or "ft" for test)
6. Click on some ads, verify that m_mac_ad_click_detected pixel has the corresponding parameter value ("1" for test, "0" for control)
7. In Debug menu, select Reset Data -> Reset CPM Experiment Cohort to clear the cohort. Restart the app.
8. Repeat resetting until you get a different cohort. Verify that pixel parameters change accordingly.
9. Disable the subfeature in config, restart the app. Verify that the pixel parameter is NOT set in this case.

Definition of Done:

• Does this PR satisfy our Definition of Done?

---

Internal references:

Pull Request Review Checklist
Software Engineering Expectations
Technical Design Template
Pull Request Documentation

[federicocappelli]

@muodov could you please remove the local BSK and point the package to the specific branch and commit it?

[github-actions]

	Warnings
⚠️	PR has more than 500 lines of code changing. Consider splitting into smaller PRs if possible.

Generated by 🚫 dangerJS against 808f6c3

[federicocappelli]

Can you explain these changes? Seems strange to modify the CI in main

[muodov]

This CI check is verifying that the autoconsent JS bundle matches the version pinned in the dependencies. This is mainly to catch when we update package.json, but do not actually commit the recompiled JS bundle.
In this case, we're actually doing exactly that: shipping a custom version that is not released as a proper library version. I removed this check so that it doesn't fail every build while the experiment is active

[federicocappelli]

Understood, do we have a task scheduled for re-activating this check?

[muodov]

Yup, https://app.asana.com/0/72649045549333/1208032179159556/f

[federicocappelli]

Fantastic thanks, then the PR LGTM if we agree that the monitoring part is enough.
You could merge BSK, release a new version (patch version bump I think) and update the branch here in the meantime.

[muodov]

Thanks @federicocappelli! It's great to have this change approved. Once we the JS review and the monitoring discussion are resolved, I'll do a final merge.

[muodov]

@federicocappelli I've added the parameter to the breakage pixel too. Could you check if it looks ok?

[federicocappelli]

The use of singletons is not how we want to do this but it's ok for an experiment.
LGTM