These are the core utilities needed to develop an open source artificial pancreas.
This is part of a series of tools to support a self-driven DIY implementation based on the OpenAPS reference design. The tools may be categorized as monitor (collecting data about environment, and operational status of devices and/or aggregating as much data as is relevant into one place), predict (make predictions about what should happen next), or control (enacting changes, and feeding more data back into the monitor).
By proceeding using these tools or any piece within, you agree to the copyright (see LICENSE.txt for more information) and release any contributors from liability.
Check out [the openaps docs] (https://openaps.gitbooks.io/building-an-open-artificial-pancreas-system/content/index.html) to help get you started.
Note: This is intended to be a set of tools to support a self-driven DIY implementation and any person choosing to use these tools is solely responsible for testing and implement these tools independently or together as a system. The [DIY part of OpenAPS is important] (http://bit.ly/1NBbZtO). While formal training or experience as an engineer or a developer is not required, what is required is a growth mindset to learn what are essentially "building blocks" to implement an OpenAPS instance. This is not a "set and forget" system; it requires diligent and consistent testing and monitoring to ensure each piece of the system is monitoring, predicting, and performing as desired. The performance and quality of your system lies solely with you.
Additionally, this community of contributors believes in "paying it forward", and individuals who are implementing these tools are asked to contribute by asking questions, helping improve documentation, and contribute in other ways.
This is not an artificial pancreas, but rather tools which independently allow:
-
monitor - Collect data about environment, and operational status of devices. Aggregate as much data relevant to therapy as possible into one place. We propose a tool,
openaps-useas a proof of concept. -
predict - Can make predictions about what should happen next.
-
control - Can enact changes in the world: emails, placing phone calls, SMS, issuing commands to pumps.
There are two ways to install openaps, from source, and as a python package via setuptools.
See GettingStarted for more important information about versions of software dependencies, but to install from source clone this repo, and issue:
sudo python setup.py develop
sudo easy_install -Z openaps
The following apt-get dependencies are required (they can be installed through variety of means, in debian/ubuntu and apt based systems the following packages are recommended/required:
sudo apt-get install python python-dev python-setuptools python-software-properties python-numpy
This installs openaps system wide.
Do not use openaps commands in the the openaps repo. Only use the
openaps directory for hacking on the core library, or for managing
upgrades through git. Running openaps inside of the openaps
source directory will error in the best case, and mess up your
openaps install in the worst case.
sudo easy_install -ZU openaps
usage: openaps [-h] [-c C C] [-C CONFIG] [--version] [command] ...
- command
- args
optional arguments:
-h, --help show this help message and exit
-c C C
-C CONFIG, --config CONFIG
--version show program's version number and exit
Utilities for developing an artificial pancreas system. openaps helps you manage and structure reports for various devices.
All of the device and report add and show commands modify
openaps.ini in the current working directory, which is assumed to be
a git repo explicitily dedicated to helping develop and configure a
DIY artificial pancreas system. This means openaps is an SDK for
an artificial pancreas system, not an artificial pancreas system.
See openaps init for setting up a brand new instance of your own
openaps, or see the notes below for details on how to convert an
existing git repo into an instance of openaps.
openaps init
openaps device <cmd>
Device commands allow you to match a device driver, with a name
and a configuration.
add - add device config to `openaps.ini`
remove - remove device from `openaps.ini`
show - print device uri, list all by default
openaps use [--format <json,stdout,text>]
[--output <filename>]
<device>
<use>
[use-args...]
For each device registered, the vendor implementation provides a
number of uses. This allows users to experiment with reports.
openaps report <cmd>
Reports match a device use to a format and filename.
add - add report config to `openaps.ini`
remove - remove report from `openaps.ini`
show - print report uri, list all by default
invoke - run and save report in file
Do not use openaps commands in the the openaps repo. Only use the
openaps directory for hacking on the core library, or for managing
upgrades through git. Instead change to a new directory, not managed
by git: cd ~/Documents.
Setup of new instance:
openaps init myopenaps - this creates an instance of openaps in a new
directory, called myopenaps
cd myopenaps - change directory to root of new repo
A valid instance of openaps is a git repo with a file called
openaps.ini present.
openaps will track configuration and some status information inside of
openaps.ini.
If you already have a git repo which you would like to become a valid openaps environent, in the root of your repo, run:
touch openaps.ini
git add openaps.ini
git commit -avm 'init openaps'
Now, wth a valid openaps environment, you can register devices for
use. A device is implemented by a vendor. openaps provides a
modular, language and process independent environment for creating
vendors and devices.
To register devices for use, see openaps device commands:
openaps device -h
openaps device add <name> <vendor> [opts...]
eg:
# register a medtronic device named pump
openaps device add pump medtronic 665455
# register a dexcom device named cgm
openaps device add cgm dexcom
Now that devices are known, and we have a variety of commands
available. We can explore how to produce reports by using devices
with the openaps use command:
openaps use <device-name> <use-name> [opts]
openaps use commands can only be used after devices have been added to
the openaps.ini config using openaps device add.
Eg:
openaps use pump -h - show available commands for the
device known as "pump"
openaps use pump iter_pump 100 - get last 100 pump history records
from the device called pump
openaps use cgm -h - show available commands for the
device known as "cgm"
openaps use cgm glucose
After experimenting with openaps use commands, users can save reports
using the openaps report commands.
openaps report commands map openaps use commands to filenames:
Adding a report means configuring a use command with a format and a
output, most commonly, a filename is used as the output.
openaps report add <report-name> <report-formatter> <device> <use> [opts]
# add a report, saved in a file called pump-history.json, which is
# JSON format, from device pump using use iter_pump.
openaps report add pump-history.json JSON pump iter_pump 100
# add a report, saved in a file called glucose.json, which is
# JSON format, from device cgm using use glucose.
openaps report add glucose.json JSON cgm glucose
Invoking a report means running a use command according to it's
configuration.
# invoke the report to create glucose.json
openaps report invoke glucose.json
# invoke the report to create pump-history.json
openaps report invoke pump-history.json
All commands support tab completion, and -h help options to help explore the live help system.
Assuming device is named pump:
usage: openaps-use pump [-h]
{Session, bolus, iter_glucose, iter_pump,
model, mytest, read_basal_profile_A,
read_basal_profile_B,
read_basal_profile_std, read_carb_ratios,
read_clock, read_current_glucose_pages,
read_current_history_pages,
read_glucose_data, read_history_data,
read_selected_basal_profile,
read_settings, read_status,
read_temp_basal, reservoir, resume_pump,
scan, set_temp_basal, settings, status,
suspend_pump}
...
positional arguments:
{Session, bolus, iter_glucose, iter_pump, model, mytest,
read_basal_profile_A, read_basal_profile_B,
read_basal_profile_std, read_carb_ratios, read_clock,
read_current_glucose_pages, read_current_history_pages,
read_glucose_data, read_history_data,
read_selected_basal_profile, read_settings, read_status,
read_temp_basal, reservoir, resume_pump, scan, set_temp_basal,
settings, status, suspend_pump}
Operation
Session session for pump
bolus Send bolus.
iter_glucose Read latest 100 glucose records
iter_pump Read latest 100 pump records
model Get model number
mytest Testing read_settings
read_basal_profile_A
Read basal profile A.
read_basal_profile_B
Read basal profile B.
read_basal_profile_std
Read default basal profile.
read_carb_ratios Read carb_ratios.
read_clock Read date/time of pump
read_current_glucose_pages
Read current glucose pages.
read_current_history_pages
Read current history pages.
read_glucose_data Read pump glucose page
read_history_data Read pump history page
read_selected_basal_profile
Fetch the currently selected basal profile.
read_settings Read settings.
read_status Get pump status
read_temp_basal Read temporary basal rates.
reservoir Get pump remaining insulin
resume_pump resume pumping.
scan scan for usb stick
set_temp_basal Set temporary basal rates.
settings Get pump settings
status Get pump status (alias for read_status)
suspend_pump Suspend pumping.
optional arguments:
-h, --help show this help message and exit
Some commands like read_glucose_data, read_history_data take a
page parameter, describing which page to fetch.
Some commands like bolus, set_temp_basal, take an input
parameter which may be - for stdin or a filename containing a json
data structure which represents the request.
All commands support -h and --help output.
usage: openaps-use cgm [-h] {glucose,iter_glucose,scan} ...
positional arguments:
{glucose,iter_glucose,scan}
Operation
glucose glucose (will pull all records)
iter_glucose <n> glucose ('n' for the number of records you want)
scan scan for usb stick
optional arguments:
-h, --help show this help message and exit