I am trying to implement a simple Google Workflow that will grab a PDF from an existing bucket, call an LLM via an API and output a summary of the document. However, the Workflow gets stopped at the getDocument stage because it cannot find the bucket, even though I am certain I have created it and input the right information. I have tried to consult the internet, but haven’t had any success.
I am quite new at this, so sorry if this is super basic.
Below is the YAML of the workflow:
main:
params: [args]
steps:
-
init:
assign: -
project: “predict-able”
-
location: “us-central1”
-
model: “text-bison”
-
method: “predict”
-
llm_api_endpoint: ${“https://” + location + “-aiplatform.googleapis.com” + “/v1/projects/” + project + “/locations/” + location + “/publishers/google/models/” + model + “:” + method}
-
sumoutputs: {}
-
getDocument:
call: googleapis.storage.v1.objects.get
args:
bucket: args.bucket
object: args.object
alt: “media”
result: documentContent -
ask_llm:
call: http.post
args:
url: ${llm_api_endpoint}
auth:
type: OAuth2
body:
instances:
-prompt: ‘${“Summarize the following document” + documentContent.data}’
parameters:
temperature: 0.1
maxOutputTokens: 2048
topP: 0.1
topK: 40
result: llm_response
my inputs are: {“bucket”: “testbuckety”, “object”: “testbuckety/WorkdayFY24”}