Python SDK

The Nextmv Python SDK can be used to bring your own Python decision model to Nextmv and interact with the Nextmv Platform. The Python SDK formalizes common patterns to standardize model options, input, logging, output, and statistics.

Export your API key as an environment variable:

Bash

export NEXTMV_API_KEY="<YOUR-API-KEY>"

Complete run workflow

Execute the complete run workflow (submit a run, poll for status, get the results). The output key will contain the actual output of the run. This is an aggregation of the other steps.

Python
Output (JSON)

import json
import os

from nextmv.cloud import Application, Client, PollingOptions

with open(os.getenv("INPUT_FILE")) as f:
    input = json.load(f)

client = Client(api_key=os.getenv("NEXTMV_API_KEY"))
app = Application(client=client, id=os.getenv("APP_ID"))
result = app.new_run_with_result(
    input=input,
    instance_id=os.getenv("INSTANCE_ID"),
    run_options={
        "solve.duration": "10s",
        "solve.iterations": "20",
    },
    polling_options=PollingOptions(),  # Customize polling options.
)
print(json.dumps(result.to_dict(), indent=2))  # Pretty print.

New run

Submit a new run. The run_id is returned.

Python
Output (Bash)

import json
import os

from nextmv.cloud import Application, Client

with open(os.getenv("INPUT_FILE")) as f:
    input = json.load(f)

client = Client(api_key=os.getenv("NEXTMV_API_KEY"))
app = Application(client=client, id=os.getenv("APP_ID"))
run_id = app.new_run(
    input=input,
    instance_id=os.getenv("INSTANCE_ID"),
    options={
        "solve.duration": "10s",
        "solve.iterations": "20",
    },
)
print(run_id)

Run metadata

Get a run metadata. The metadata.status key will contain the status of the run. Use the run_id obtained from submitting a run.

Python
Output (JSON)

import json
import os

from nextmv.cloud import Application, Client

client = Client(api_key=os.getenv("NEXTMV_API_KEY"))
app = Application(client=client, id=os.getenv("APP_ID"))
metadata = app.run_metadata(run_id=os.getenv("RUN_ID"))
print(json.dumps(metadata.to_dict(), indent=2))  # Pretty print.

Run result

Get a run result. The output key will contain the actual output of the run. Use the run_id obtained from submitting a run.

Python
Output (JSON)

import json
import os

from nextmv.cloud import Application, Client

client = Client(api_key=os.getenv("NEXTMV_API_KEY"))
app = Application(client=client, id=os.getenv("APP_ID"))
result = app.run_result(run_id=os.getenv("RUN_ID"))
print(json.dumps(result.to_dict(), indent=2))  # Pretty print.

The app.run_result_with_polling method is also available. It will poll for the run status and return the result when it is succeeded.

Cancel a run

To cancel a run, use the cancel_run method.

Python

import os

from nextmv.cloud import Application, Client

client = Client(api_key=os.getenv("NEXTMV_API_KEY"))
app = Application(client=client, id="<YOUR-APP-ID>")
app.cancel_run(run_id="<YOUR-RUN-ID>")

Account queue

Get the queue of runs for the account. A run is queued when the status_v2 in the metadata is queued. Use the Account.queue method.

Python
Output (JSON)

import json
import os

from nextmv.cloud import Client
from nextmv.cloud.account import Account

client = Client(api_key=os.getenv("NEXTMV_API_KEY_PROD"))
account = Account(client=client)
queue = account.queue()

print(json.dumps(queue.to_dict(), indent=2))  # Pretty print.

Reference for running an app remotely with the Python SDK.