This page contains all the required steps to allow longer runtime on your desired Tool.

​
Requirements

​
Region and Authorization token

Follow the instruction on the API keys page to fetch your authorization token and your account region.

region = "REGION CODE AS INSTRUCTED in the API keys page"
authorization_token = "AUTHORIZATION TOKEN AS INSTRUCTED in the API keys page"

​
Studio ID

There are different ways to find the studio ID of a Tool. The easiest way is to use the URL. For example when on the Tool and Use tab, the component before use/app and after project ID is the studio ID. The string below shows the URL structure:

https://app.relevanceai.com/notebook/<Region>/<Project ID>/<Studio ID>/use/app

studio_id = "STUDIO ID AS INSTRUCTED ABOVE"

​
Tool inputs

A Get API call to the URL below, will provide you with the Tool information.

https://api-<Region>.stack.tryrelevance.com/latest/studios/list_public?filters=${JSON.stringify([{ filter_type: 'exact_match', field: 'title', condition_value: <ToolTitle>}])}
Tool title can be found on top of the page and s the name assigned to the Tool. 
import requests
import json

tool_title = "xxxxxxxxxx"
headers = {"Authorization": authorization_token}

filters = json.dumps([{ "filter_type": 'exact_match', "field": 'title', "condition_value": tool_title}])
url = f"https://api-{region}.stack.tryrelevance.com/latest/studios/list_public?filters={filters}"

tool_info = requests.get(url, headers=headers).json()["results"][0]
tool_inputs = tool_info["params_schema"]["properties"]

​
Start long running tool

Provide studio_id in the code below to trigger a task and check the progress:

  • Synchronous triggers last for 30 seconds
  • Asynchronous triggers last for 15 minutes
# Start long running tool
url = f"https://api-{region}.stack.tryrelevance.com/latest/studios/{studio_id}/trigger_async"
headers = {
    "Authorization": authorization_token,
    "Content-Type": "application/json"
}
payload = {"params":tool_inputs}
response = requests.post(url, headers=headers, json=payload)

# Extract the tools job id, so we can check its progress
job_id = response.json()['job_id']

# Every 5 seconds, check if the tool had finished by calling the poll endpoint
poll_url = f"https://api-{region}.stack.tryrelevance.com/latest/studios/{studio_id}/async_poll/{job_id}?ending_update_only=true"
tool_output = {}
for _ in range(100):
  poll_response = requests.get(poll_url,headers=headers).json()
  print(poll_response)
  if poll_response['type'] == "complete" or poll_response['type'] == 'failed':
    print(poll_response)
    tool_output = poll_response["updates"][0]["output"]["output"]
    break
  time.sleep(5)

# This is the tools output
print('Tool-output',tool_output)

Was this page helpful?

FAQGPT on my files