ezgh ocr batches download <batch> (--dir out/ | --zip out.zip | --jsonl out.jsonl) [flags]Save a finished batch’s outputs, streamed to disk (never printed on a terminal):
--dir d each document's output in d, named after the document (receipt.pdf's is
receipt.txt, or receipt.json for json-layout), in the subdirectories it was
uploaded from. An output already there is kept; one cut off (<name>.part) is
resumed where it stopped. Documents that failed have no output.
--zip f a zip of every output, named by document ID.
--jsonl f results.jsonl: a line per document with its status, error and page counts.--zip - and --jsonl - write to stdout, when it isn’t a terminal. Files are readable only by you. Outputs are kept for 30 days after the batch finishes.
Part of ezgh ocr batches.
Examples
ezgh ocr batches download obt_k3f9a0x2m7qp --dir out/
ezgh ocr batches download march --zip march.zip
ezgh ocr batches download march --jsonl - | jq -r 'select(.status != "succeeded") | .name'Flags
| Flag | Type | Default | Description |
|---|---|---|---|
--dir |
string | save each document’s output in this directory | |
--jsonl |
string | save results.jsonl to this file (- for stdout) | |
--zip |
string | save a zip of every output to this file (- for stdout) |
This command also takes the global flags.