Make pagination more generic and not Franklin specific

[pomadchin]

Overview

This PR makes pagination more generic and not Franklin specific.

Checklist

• Changelog updated (please use chan)

// add "com.softwaremill.sttp.client3" %% "async-http-client-backend-cats-ce2" % "3.3.15" % Test
package com.azavea.stac4s.api.client

import com.azavea.stac4s.StacItem
import cats.effect.IO
import cats.syntax.either._
import io.circe.parser
import sttp.client3.asynchttpclient.cats.AsyncHttpClientCatsBackend
import sttp.client3.UriContext

import org.scalatest.BeforeAndAfterAll
import org.scalatest.funspec.AnyFunSpec
import org.scalatest.matchers.should.Matchers

import scala.concurrent.ExecutionContext

class ItSpec extends AnyFunSpec with Matchers with BeforeAndAfterAll {
  val filters = parser
    .parse("""{
          "collections": [
          "mcd43a4"
        ],
        "bbox": [
          34.74406064953656,
          -4.950752223991549,
          35.321110216900706,
          -4.573783231966013
        ],
        "datetime": "2017-01-01T00:00:00.757Z/2017-12-31T00:00:00.635Z"
      }""")
    .flatMap(_.as[SearchFilters])
    .valueOr(throw _)
  
  describe("should fetch items with pagination") {
    it("request with limit no merge") {
      implicit val cs = IO.contextShift(ExecutionContext.global)
      val realitems: List[StacItem] = AsyncHttpClientCatsBackend
        .resource[IO]()
        .use { backend =>
          SttpStacClient(backend, uri"https://eod-catalog-svc-prod.astraea.earth/")
            .search(filters)
            .compile
            .toList
        }
        .unsafeRunSync()

      realitems.length shouldBe 364
    }
  }
}

Closes #409
Closes #410
Closes #411

[pomadchin]

That is probably not a part of this PR and could be a more broad discussion. Good for now!

[jisantuc]

Eh, I think this is fine -- we don't really know anything about "semantics" of a pagination token, it might not even be a non-empty string in some cases. We should revisit if this becomes insufficient but I think this is pretty much all we know about pagination tokens right now

[jisantuc]

I'm good with this -- I think when we were talking about pagination originally I wanted stac4s to treat the token as completely opaque, but we went with copy-pasta instead to unblock ourselves, so I'm happy this is getting genericized