medplum / medplum / 28139611743

92%

push

github

fix(server): paginate AWS Textract results via NextToken (#5879) (#9614)

AWS Textract paginates GetDocumentTextDetection results for large
documents: the first completed response carries a NextToken and the
caller must follow it to retrieve the remaining Blocks. The handler
stopped at the first response, so only the first page of text was
returned and a stale NextToken leaked into the output.

Follow NextToken after the job completes, accumulating Blocks across all
pages (matching the AWS pagination idiom in cloud/aws/config.ts) and
clearing the trailing token on the merged result. Add a regression test
covering the multi-page NextToken path.

Fixes #5879

Signed-off-by: Jake Iglesias <jacob.ds.iglesias.5@gmail.com>
Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

21000 of 23912 branches covered (87.82%)

Branch coverage included in aggregate %.

7 of 7 new or added lines in 1 file covered. (100.0%)

37879 of 40185 relevant lines covered (94.26%)

12567.93 hits per line