When trying to fetch the Transcription Job from the previous node (AWS S3) I am unable to find it. The JSON is empty (output from AWS S3 2 in my flow) and is empty in the input data for the AWS Transcribe node. It is saying error 400 the key is empty but I cannot find a key or transcription job name anywhere to replace it with. I searched the forum and discord with no success.
Do you get a “job id” or some kind of identifier back in the response from AWS Transcribe 1? I see you have referenced $json['key'] in AWS Transcribe 2, but is that the job id? Does AWS send the Transcription “job id” back to you in the callback you get in the Wait node? It might be nested in $json.body.jobId (or something like that) instead.
The workflow is failing at AWS Transcribe 1 prior to the wait node because there is no “Key”. I have attached a screenshot of the AWS Transcribe 1 node so you can see the input data coming from the AWS S3 2 node.
Correct. You have expressions ($json['key']) pointing to something in the AWS S3 2 output where you would be using files found in the salesrecordingsmbiz bucket, but there apparently are none (the S3 “get” response item has nothing in it).
What does the AWS S3 2 output look like when there is something found in the bucket?
The ($json['key'] ) in AWS Transcribe 1 was autofilled. Not something I added. This screenshot shows you AWS S3 2. The nodes are succeeding up to this AWS Transcribe 1 node.
The reference to Key is part of a JSON object/item that is flowing out of the AWS S3 2 node and into the Transcribe Node. The problem is that the AWS S3 2 node did not fetch anything from S3, so that object is empty (i.e. there is no Key attribute, nor any other attributes). Here is approximately what it should look like, when it actually fetches some files from the S3 bucket.
I thought it was returning empty because maybe the file wasn’t uploading properly but the file is there in the bucket. Do you know what else may be casue the issue in node AWS S3 2 that it would output nothing? I attached a screenshot of the file in the bucket and then AWS S3 1 as well as AWS S3 2 nodes.
It looks like it should work. Possibly a timing thing. Maybe S3 takes a moment to re-build a bucket’s table of contents??? You could try adding a Wait node between the S3 Upload and the S3 Get Many.
Edit: Just noticed on the upload, you have nothing specified for File Content You would need to have a reference to a binary there, like file or data from the previous node.
Edit2: Just tested the Google Drive Trigger to see what it does for Changes involving a specific folder. You only get a list of metadata about items that changed, without the binary data. So, you’ll need to add a Google Drive - Download File step after that to fetch the binary, and then upload that to S3 (File Content would have the value data… literally… not an expression).
