How do I determine if all the documents within the batch have been processed?
Semantria guarantees that all queued documents will be processed on our side regardless of whether or not they were pushed through in batches or one-by-one. A batch is simply a collection of documents that allow the users to queue a configured number of documents in one API call.
Semantria does not offer an option to check the batch’s processing status. This feature is available for documents only. If the document is not processed on Semantria, due to some reason (ie. too many characters per sentence or to many non-alphanumeric characters), it will be returned to the client without any analysis data, a FAILED status, and appropriate message.
Semantria will return an entire batch as FAILED if it doesn't meet the minimum requirements (ie. one document within the batch is empty). Please ensure there is source text for every given unique ID.
Can Semantria handle concurrent batches of documents being sent to its engine?
Semantria is completely asynchronous service so you can queue your batches concurrently if needed. All batches will be processed on multi-threading environment so they won't be delivered in the same order as queued.
Is there a limit to the number of documents that can be sent per batch?
The batch size is different and customized for each Semantria account. We can allow any size of batch for your account but this value should be consistent with the allowed document size. For example, if we consider a document size of 8KB, 100 documents will be a 800KB batch, which is large amount of text. Take a note that Semantria output will be at least 2-3 times larger, so extremely large batches are undesirable.
Sending the documents in batches is more preferable than sending them one-by-one because it guarantees faster processing due to internal Semantria specifics.