Quickstart

Prerequisites

API Credentials

To access Rune’s APIs, you will need to obtain API credentials. For multi-patient analyses, we recommended using user access tokens. These provide access to all the user’s allowed resources.

To create a new access token:

Log in to the Rune web portal

Click on the profile icon, in the top right corner.

Click on User Settings.

On the left sidebar, click on Access Tokens.

Click CREATE ACCESS TOKEN.

Copy the token ID and secret before closing the page. The secret will never be shown again.

See Configuration Setup for details about how to use these credentials with this library.

It is highly recommended that you rotate your access tokens every 3-6 months, by creating a new token and deactivating the old one. Store your access tokens securely, and do not share them.

Multiple Organizations

For users who are members of multiple organizations, note that user access tokens only operate within the context of the organization that is currently active. To switch your active organization, log in to the Rune web portal and click on the profile icon in the top right corner. If you are a member of multiple organizations, the profile icon’s dropdown menu will have an option to switch your organization.

Configuration Setup

runeq uses a YAML-formatted file to manage configuration settings (e.g. API credentials). The easiest way to set up this configuration is via the runeq command line tool, which is installed along with the Python library.

To get started, open a terminal and run the following command in a Python environment where runeq is installed. This command will prompt you to enter an access token ID and secret, and it will create a configuration file in the default location.

runeq configure setup

This command also provides options to get and set specific values in your config file. To see help documentation:

runeq configure --help

If you want to create or manage a configuration file manually, refer to the example config for the expected contents.

Once a configuration file exists, you won’t need to repeat this step (unless you’re rotating your access token, getting set up on a different computer, etc).

Initialization

To get started with the library, use initialize:

from runeq import initialize

initialize()

This loads credentials from your configuration file (see Configuration Setup).

Usage

Explore Metadata

After initializing the library, you can fetch metadata about various resources.

For example, you can get information about your user, based on the authentication credentials:

from runeq.resources.user import get_current_user

my_user = get_current_user()
print(my_user)
print('Active Org:', my_user.active_org_name)

You can also fetch metadata about all the patients you have access to:

from runeq.resources.patient import get_all_patients

patients = get_all_patients()

for patient in patients:
    print(patient)
    for device in patient.devices:
        print(' ', device)

    print('')

get_all_patients returns a PatientSet, which can be serialized as a list of dictionaries, e.g. to save the metadata to a file:

import json

with open('patients.json', 'w') as f:
    json.dump(patients.to_list(), f, indent=4)

You can also convert a PatientSet to a collection of devices (a DeviceSet). This may be more convenient for a columnar data format, like a pandas DataFrame.

import pandas as pd

devices = patients.devices
devices_df = pd.DataFrame(devices.to_list())

Similarly to fetching information about patients, you can fetch information about projects, and metadata related to the patients within projects (and cohorts).

You can find information about a single project:

from runeq.resources.project import get_project

project = get_project(project_id="example_id")
print(project.to_dict())

To view all the patients in a project, and their related project metrics you can use the following example:

from runeq.resources.project import get_project_patients

project_patients = get_project_patients(project_id="example_id")

for project_patient in project_patients:
    print(project_patient)
    for metric in project_patient.metrics:
        print(' ', metric)

    print('')

It may be easier to view a single project patient in a dataframe which you can do by:

from runeq.resources.project import get_project_patients

project_patients = get_project_patients(project_id="example_id")
target_patient_id = "patient_id_example"

df = project_patients[target_patient_id].get_patient_metadata_dataframe()

df

Fetch Timeseries Data

Use get_patient_stream_metadata to get a StreamMetadataSet with details about a particular patient’s data. If you’re interested in a more specific set of streams, the function accepts additional filters.

from runeq.resources.stream_metadata import get_patient_stream_metadata

patient_id = "c4bd060df1454aa0adc978985512c6e9"
patient_streams = get_patient_stream_metadata(patient_id)
print(f'Found {len(patient_streams)} streams')

Once you have a StreamMetadataSet, you can use the filter operation to get a more specific subset of streams:

# Filter for data collected from a particular device
device_id = "eb#8c31"
device_streams = patient_streams.filter(device_id=device_id)

# Filter by broad category
neural_streams = patient_streams.filter(category="neural")

# Specify multiple arguments to find streams that match
# all criteria
neural_device_streams = patient_streams.filter(
    category="neural",
    device_id=device_id,
)

# Use a custom filter function
import time

def in_last_two_weeks(stream) -> bool:
    """Return True if stream has data in the last two weeks"""
    two_weeks_ago = time.time() - 14*24*60*60
    return stream.max_time > two_weeks_ago

recent_vitals_streams = patient_streams.filter(
    category="vitals",
    filter_function=in_last_two_weeks
)

You can also combine multiple StreamMetadataSet s, using update:

from runeq.resources.stream_metadata import StreamMetadataSet

lfp_power_streams = patient_streams.filter(
    category="neural",
    measurement="lfp_trend_log_power",
)
tremor_streams = patient_streams.filter(
    category="symptom",
    measurement="tremor",
    stream_type_id="duration"
)

lfp_and_tremor_streams = StreamMetadataSet()
lfp_and_tremor_streams.update(lfp_power_streams)
lfp_and_tremor_streams.update(tremor_streams)

Using a StreamMetadataSet, you can fetch the availability of all or any of the streams:

availability_df = lfp_and_tremor_streams.get_batch_availability_dataframe(
    start_time=1662000000,
    end_time=1663123000,
    resolution=3600,
    batch_operation="any",
)

Note

The API for “batch availability” has a limit on the number of streams that it can process at a time. If you’re running the example code with a patient who has multiple devices, the snippet above may exceed the API limit. Try limiting the number of streams in the set using a custom filter function, to select for a few of those device IDs.

When you’re ready to fetch data, you can gather all the raw stream data into a pandas dataframe:

stream_df = lfp_and_tremor_streams.get_stream_dataframe(
    start_time=1662499000,
    end_time=1663123000,
)

You can also work directly with responses from the V2 Stream API. See stream and StreamMetadata for details.