flow sdk
pip install convect-flow-sdk
# set up environment variables
# FLOW_WORKSPACE_ID: your flow workspace id
# FLOW_API_TOKEN: your flow api token from the workspace
# FLOW_HOST: flow host, default to https://flow.convect.ai
from convect_flow_sdk import FlowAlgo
from pprint import pprint
# this is an example of how to use the flow_algo_sdk
flow_algo = FlowAlgo()
# this algo_id is will be the one defined in the algo project
algo_id = '2f1a1fa9-2958-4afe-bb48-4239a960986d'
pprint(flow_algo.list_algos())
pprint(flow_algo.list_algo_runs(algo_id))
# this submit will zip all the files in path_to_input folder and submit to flow
run_id =flow_algo.submit(algo_id,
"weekly_run",
{
"input_file": "input.csv",
"output_file": "output.csv",
"predict_start_week": "202348",
"predict_end_week": "202348",
"algo": "v03-percentile"
},
"path_to_input")
# flow_algo.terminate(run_id)
flow_algo.check_status(run_id)
flow_algo.log(run_id)
# this gather will download the output.tar.gz from flow and extract to path_to_output folder
flow_algo.gather(run_id,"./output")
# flow_algo_sdk.clear_local_algo_cache()
# clear local algo cache will delete the local history of submitted runs
# set up environment variables
# FLOW_WORKSPACE_ID: your flow workspace id
# FLOW_API_TOKEN: your flow api token from the workspace
# FLOW_HOST: flow host
# you need to make sure the workspace has all the active apps
# you can run the following command to run the regression test
flow-cli regression-test