Add API handlers and API Docker image

[rchitale7]

Description

Initial implementation of Remote Vector Index Builder API handlers. The API handler is a Docker image, built from the base Core image. It contains the out the box functionality to integrate Faiss GPU accelerated index building with the k-NN vector engine.

Issues Resolved

Closes #42

By submitting this pull request, I confirm that my contribution is made under the terms of the Apache 2.0 license.
For more information on following Developer Certificate of Origin and signing off your commits, please check here.

[Rajrahane]

One comment with memory management.
I understand memory allocations are rounded up to nearest boundary, of say 256 bytes
using floats for multiplications would work, but we'll have to round up to the nearest boundary, and subtract that value from the total_memory used.

We might end up subtracting less and an incorrect request acceptance by the request queue would throw an out of memory error during the index build process

[rchitale7]

The memory estimations are already pretty rough - the formula we're using in calculate_memory_requirements isn't actually accurate. Improving the accuracy of memory calculations, along with more robust request store management features, should be taken as follow up issues/PRs, in my opinion.

[rchitale7]

Will add instructions for how to configure and run the API handler once this PR #46 is merged. Basically, i will condense whats talked about here opensearch-project/k-NN#2545 into easy to read instructions

[jed326]

I will make a second pass on the tests later, but the source code LGTM so approving

[Rajrahane]

LGTM