Pagination for List Entries

[lincmba]

succeeds #27

Problem:
When the FHIR gateway is used to fetch data from a FHIR server using the List endpoint with a specific ID ([GET] /List?_id=<some-id>), it currently fetches a bundle containing entry URLs. For certain resources, like Locations, with a large number of entries, the gateway faces timeout errors due to the volume of data.

Temporary Fix:
The temporary fix for the SID project was to increase the HTTP client timeout .

Proposed Solution:
Implement pagination consistent with the current FHIR responses. Introduce two parameters, _page (default set to 1) and _count (default set to 20), passed to the URL. These parameters will help limit the number of items fetched from the FHIR server.

Workflow:

• Gateway receives an initial request: [GET] /List?_id=<some-id>&_count=<page-size>&_sort=<some-sort>&<others>.
• Gateway receives subsequent requests: [GET] /List?_id=<some-id>&_page<page-number>&_count=<page-size>.
• Gateway fetches all relevant entries.
• Uses page number and offset to determine specific records to fetch from FHIR.
• Computes start and end: start = (_page - 1) * _count, end = start + _count.
• Generates previous and next page links.
• Fetches only the entries within the start and end limits from the FHIR server.
• Response is put in a bundle, including both previous and next page links and returned.
• No previous page link if it is the first page, and no next page links if it is the last page.

Acceptance Criteria:

• Pagination is implemented for the List endpoint.
• Fetching entries is limited based on the specified page and offset.
• Previous and next page links are generated appropriately.
• No previous page link for the first page, and no next page links for the last page.
• Tests
• Documentation on List mode entries in Read Me

[ndegwamartin]

We have to different handler methods for this functionality:

• one for when the request is /List/12345
• one for /List?_id=12345 - See line L251

The current implementation works for the first one. We need the second one to have the same implementation.