Fix ImagePane tests

[da-h]

Description

This PR fixes the failing ImagePane cypress tests.

Motivation and Context

Tests started to fail after #900 & #902 were merged, although both were succeeding when tested separately.

• One problem was that in the previous version, the click-positions were always off by one pixel. The new version has fixed this. I went the extra few steps to also check for correct pixel positions in the test, which has not been done before.
• Second, the image dragging cypress tests did not work as expected. After some tinkering, I best fixed the test using the @4tw/cypress-drag-drop cypress plugin. I also added to test the combination of zooming + panning.

How Has This Been Tested?

• Locally with cypress against the current master + this PR
• on my fork on top of #898
• on my fork on top of #901

Types of changes

• [x] Test changes (non-breaking change which fixes an issue)
• [ ] Bug fix (non-breaking change which fixes an issue)
• [ ] New feature (non-breaking change which adds functionality)
• [ ] Breaking change (fix or feature that would cause existing functionality to not work as expected)
• [ ] Code refactor or cleanup (changes to existing code for improved readability or performance)

Checklist:

• [ ] I adapted the version number under py/visdom/VERSION according to Semantic Versioning
• [ ] My code follows the code style of this project.
• [ ] My change requires a change to the documentation.
• [ ] I have updated the documentation accordingly.

[da-h]

Note: The "Initialize Visual Regression Test" will fail here, although no visual changes should occur as this PR changes only functional tests.

The test is failing here, because

• this PR adds a npm module (@4tw/cypress-drag-drop) to the tests.
• for the initializaton-phase of the visual regression tests, there is a checkout of the PR's test-folder (excluding package.json and yarn.lock).
• Thus, in this particular case, the visual regression initialization will always fail.

I should probably submit a PR for that someday. :sweat_smile: ~Note to future-contributor or future-self: this requires to checkout only the devDependencies of package.json not the actual dependencies as we want to test against the previous version, e.g. when there is a npm update.~ I am not sure how to tackle this, though. Only installing devDependencies is not a good solution, as it would fail to detect bugs or other changes due to an update in the testing framework. (Like the one we had not long time ago, where a cypress update lead to text being rendered a bit differently).

[da-h]

Resolved any conflict (in file yarn.lock) by rebasing onto master.

The failing (polling) test suggests that I need to look into this once more, but more thoroughly. :sweat_smile:

[da-h]

LGTM now. It seems it has been a timing issue with the polling variant. I've just added a short delay in the test before the second click in the callback-test.