Any jobs you create with a single request are grouped into a batch. When you create the jobs, we return a batch_id in the response. The batch_id allows you to retrieve the batch object, which contains information on the current status of the batch. You can also retrieve the IDs for all the jobs in the batch.
The job batch object can be inspected using Get job batch.
Job batch object attributes
| Attribute | Type | Description |
|---|---|---|
app_uuid | string | Unique identifier for the data program for which this batch was created |
batch_id | string | Unique identifier for the job batch |
batch_total | integer | The total number of jobs included in the batch |
created | ISO 8601 date | UTC timestamp of batch creation |
invalid_index | integer | The position within the array of inputs where the first invalid input appears |
total_processed | integer | The total number of jobs within the batch that have been processed |
valid_count | integer | The total number of jobs within the batch that have passed validation and are available for labeling |
batch_processed | boolean | Whether the entire batch has been processed or not |
error_message | string | An indication of what caused the first input to fail during processing |
errors_with_filename | list[string] | List of CSV tuples containing "{filename}, {errror_message}" |
{
"app_uuid": "123abc45-6789-01de-f2g3-45h67ijk8901",
"batch_id": "a1b23456-7890-1234-c56d-7890123e4f56",
"batch_total": 6,
"created": "2020-06-08T14:27:36.862Z",
"invalid_index": 0,
"total_processed": 0,
"valid_count": 0,
"batch_processed": true,
"error_message": "string",
"errors_with_file": [
"\"https://cdn.super.ai/non_existing_file.pdf\",\"Error occurred while fetching the document:https://cdn.super.ai/non_existing_file.pdf, maximum retries exceeded.\""
],
Sample Error Messages
In case you exceeded the maximum payload size:
Document size exceeds the maximum size allowed: {doc_size} > 50 MB
In case you exceeded the maximum number of pages:
The number of pages in the document exceeds the maximum number of pages allowed: {num_pages} > 100
File specific error messages
Sometimes only a subset of documents in a batch is containing errors. In those cases, the rest of the documents get submitted as usual (when using non-strict mode). To debug the issues with the remaining files, the API returns file specific rejection reasons in the errors_with_filename field. The format of that field is a list[str] where each string object is a comma separated tuple with {filename}, {error_message}.