fix(imagery): paginate raw CMR pages within the record budget

Rebased onto current main with review follow-ups folded in.

Co-Authored-By: Claude Sonnet 5.5 <[email protected]>
This commit is contained in:
Rudy CelekliandClaude Sonnet 5.5 committed 2026-10-14 20:13:52 -04:00
1 parent 0265ad4c19
commit 360aa74f50
4 files changed
+72 -6

No files matched your search

+12 -6
View File
@@ -40,6 +40,7 @@ export async function fetchCmrPages({
}) {
const granules = [];
let hits = 0;
let readCount = 0;
let cursor = null;
for (;;) {
const init = { signal };
@@ -47,19 +48,24 @@ export async function fetchCmrPages({
const response = await fetchImpl(url, init);
if (!response?.ok)
throw new Error(`HTTP ${response?.status ?? 'error'} from CMR`);
const page = parseCmrUmm(await response.json(), product);
granules.push(...page.granules);
const json = await response.json();
const rawCount = Array.isArray(json?.items) ? json.items.length : 0;
const page = parseCmrUmm(json, product);
readCount += rawCount;
granules.push(
...page.granules.slice(0, Math.max(0, maxRecords - granules.length)),
);
hits = Math.max(hits, page.hits);
cursor = response.headers?.get?.(CMR_SEARCH_AFTER) ?? null;
if (
!cursor ||
page.granules.length < CMR_PAGE_SIZE ||
granules.length >= maxRecords ||
granules.length >= hits
rawCount < CMR_PAGE_SIZE ||
readCount >= maxRecords ||
readCount >= hits
)
break;
}
return { granules, truncated: hits > granules.length };
return { granules, truncated: hits > readCount };
}
function finiteOrNull(value) {
@@ -0,0 +1,56 @@
import test from 'node:test';
import assert from 'node:assert/strict';
import { fetchCmrPages } from './catalog.js';
const item = (id) => ({
meta: { 'concept-id': id },
umm: {
TemporalExtent: {
RangeDateTime: { BeginningDateTime: '2026-09-18T12:00:00Z' },
},
},
});
test('discarded rows do not turn a full CMR page into a final page', async () => {
let calls = 0;
const result = await fetchCmrPages({
product: 'S30',
url: 'fixture',
fetchImpl: async () => {
calls++;
const items =
calls === 1
? Array.from({ length: 200 }, (_, i) =>
i === 0 ? {} : item(String(i)),
)
: [item('last')];
return Response.json(
{ hits: 201, items },
{ headers: { 'CMR-Search-After': `cursor-${calls}` } },
);
},
});
assert.equal(calls, 2);
assert.equal(result.granules.length, 200);
assert.equal(result.granules.at(-1).id, 'last');
assert.equal(result.truncated, false);
});
test('a non-page-multiple limit never returns more admitted records than its budget', async () => {
let calls = 0;
const result = await fetchCmrPages({
product: 'S30',
url: 'fixture',
maxRecords: 250,
fetchImpl: async () => {
calls++;
return Response.json(
{
hits: 1000,
items: Array.from({ length: 200 }, (_, i) => item(`${calls}:${i}`)),
},
{ headers: { 'CMR-Search-After': `cursor-${calls}` } },
);
},
});
assert.equal(calls, 2);
assert.equal(result.granules.length, 250);
assert.equal(result.truncated, true);
});