Limit Karton analyses list to 10 elements, add pagination and optimize handling of object relationships

[psrok1]

Your checklist for this pull request

• I've read the contributing guideline.
• I've tested my changes by building and running the project, and testing changed functionality (if applicable)
• I've added automated tests for my change (if applicable, optional)
• I've updated documentation to reflect my change (if applicable)

What is the current behaviour?

Motivation is similar to #1023. When we are querying or uploading a very popular object with incrementally growing number of relationships and Karton analyses, all of these analyses are loaded by backend and returned which causes excessive use of resources and request timeouts.

In case of Karton analyses, Karton backend is additionally asked for status of each analysis which causes additional overhead. The result is usually worker killed with WORKER TIMEOUT.

What is the new behaviour?

• API returns only 10 Karton analyses at once. More analyses can be loaded using new older_than parameter.
• Backend avoids loading Object.analyses and Object.parents also during creation/removal of the relationships. Right now we observe that WORKER TIMEOUT happens also when such "popular" object is uploaded.
• Karton analyses returned by API were unordered. Now they're ordered from latest to oldest.
• I have also fixed some issues related with analysis id typing in the frontend part.

Test plan

API correctness should be already covered by automated tests. "Popular" object querying and upload was tested manually.

Closing issues

closes #1050