Skip to content
This repository has been archived by the owner on Feb 2, 2022. It is now read-only.

Latest commit

 

History

History
207 lines (174 loc) · 8.33 KB

install_notes.md

File metadata and controls

207 lines (174 loc) · 8.33 KB
Raw

Installation notes for DVH Analytics

Before you start

Please note this software does not work "out-of-the-box." The smoothest way to run this software is with Docker and DVH Analytics Docker. But if you prefer running the source code, here are some broad strokes:

  • Must have a PostgreSQL DB setup prior to use. PostgreSQL Notes
  • Must initialize import_settings.txt and sql_connection.cfg
    $ dvh settings or $ dvh settings_simple
  • Must import data via the admin GUI or $ dvh import prior to running the main Bokeh app.

DVH Analytics relies on Bokeh to create web servers/pages for the GUIs, so it's best to follow this guide as this app may not behave in the way you might expect from a traditional python app.

Pre-requisites

If any of these are not installed, see Additional Notes.

Setting up your python environment

DVH Analytics is available via pip install:

$ pip install dvh-analytics

If you're on Mac or Linux, you may need:

$ sudo pip install dvh-analytics

And for Linux only, you may need to install freetype prior to pip install dvh-analytics:

$ sudo apt-get install libfreetype6-dev

Define directories and connection settings

For a web-based UI, type in the following to access the Settings:

$ dvh settings

or the following for a command-line based setup:

$ dvh settings_simple

Processing test files

To verify all of your settings are valid and installation was successful, type:

$ dvh test

This will process all dicom files located in the test files folder. If this test passes, then all dependencies are successfully installed (e.g., PostgreSQL, Python libraries), import directories are valid, and communication with your SQL DB has been established.

Importing your own data

To import your own data, you can click "Import all from inbox" in the Database Editor tab of the settings GUI

Alternatively, to import your own data via command line for more options:

$ dvh import

Assuming all of the previous steps were successful, all dicom files anywhere within your inbox will be imported into the SQL database, then organized by patient ID and moved to the specified imported folder. Note that by default, the code will not import dicom files with Study Instance UIDs that are already in the database.

You may force the import regardless of UID conflicts by typing:

$ dvh import --force-update

If you'd like to import from a directory other than the one in your settings:

$ dvh import --start-path /some-absolute-directory-name

ROI Name Manager, DB Editor, and Backup & Restore

Using the 'settings' command will start a Bokeh server and open your default browser to access the ROI Name Manager, DB Editor, and Backup & Restore modules. Be sure to stop the server when you're done by press ctrl + c in the command line window.

$ dvh admin

Main DVH Analytics view

Type the following to start the Bokeh server:

$ dvh run

If Chrome or Firefox is not your default browswer, you'll need to copy and paste the url into Chrome or Firefox. From within the active terminal, press ctrl + c to stop the Bokeh server.

If you would like to specify an IP or port other than the default 127.0.0.1:5006, use the following when starting the Bokeh server. You may be interested in this if you are running from a computer with a static IP and would like to access this Bokeh server from across your network.

$ dvh run --allow-websocket-origin <new IP:port>

$ dvh run --port <some other port number>

These two features work with any of the commands that start web-based UI.

Additional Notes

Python 2.7

(Windows only, Xcode command line tools for mac and Ubuntu both include Python 2.7)
Download Python 2.7 (not 3.x) from: https://www.python.org/downloads/windows/

Be sure to include pip in the installation and let the installer update your environment variables.

Xcode Command Line Tools (Mac Only)

Make sure Xcode command line tools are installed. If the full Xcode package is installed (from the Mac App Store), enter the following into a terminal window:

$ xcode-select --install

PostgreSQL

If you are familiar with PostgreSQL and have access to a PostgreSQL DB, you simply need to fill in the login information by running:

$ dvh settings_simple --sql

If you need PostgreSQL, here are some options for each OS.

Mac OS
Simply download the PostgreSQL app: http://postgresapp.com/

  • Open the app
  • Click "Start"
  • Double-click "postgres" with the cylindrical database icon
  • Type the following in the SQL terminal:
create database dvh;

Then quit by typing:

\q

NOTE: You may replace dvh with any database name you wish, but you must update dbname in settings to reflect what database name you chose.

Ubuntu
You probably already have PostgreSQL installed, but if you don't, type the following in a terminal:

$ sudo apt-get install postgresql postgresql-client postgresql-contrib libpq-dev
$ sudo apt-get install pgadmin3

Upon successful installation, open type 'pgadmin3' in the terminal to open the graphical admin.
Then, create a user and database of your choice (same instructions found below for Windows)

Windows
Download the installer for BigSQL: https://www.bigsql.org/postgresql/installers.jsp/

  • Be sure to include pgAdmin3 LTS
  • After installation, launch pgAdmin3 LTS from the Windows Start Menu.
    • Right-click localhost and then click connect.
    • Right-click Login Roles and then click New Login Role.
    • Fill in Role name (e.g., dvh), click OK
    • Right-click Databases then click New Database
    • Fill in Name (e.g., dvh), set owner to the Role name you just created. Click OK.

PIP (python package management):

From a terminal window, type:

Mac

$ easy_install pip

Ubuntu

$ sudo apt-get install python-pip

Windows
Download get-pip.py from here: https://pip.pypa.io/en/stable/installing/
Then compile:

$ python get-pip.py

Security Concerns

Obviously, data intended for an application like this may be sensitive and require HIPPA compliance. The end user is entirely liable for setting up an appropriately secure environment. Bokeh provides some help here with regards to a reverse proxy, so that HTTPS may be implemented.

DVH Analytics does provide a loose framework for authentication. There is a parameter 'auth_user_req' located in options.py; by default this is set to False. If set to True, user name and password fields will be provided to the user when accessing any of the Bokeh servers in this app. These credentials will be passed to the check_credentials function in auth.py. By default, this function simply returns True. The end user must supply their own code for authentication. An example code using python-ldap is provided in auth.py, but requires some editing for each end user's own implementation. And please remember, you need to setup HTTPS on your own for this authentication to make any sense, otherwise user name and passwords will not be encrypted inbetween the web-user and the Bokeh server.

Your SQL DB password is stored as plain text in preferences/sql_connection.cnf. I realize this is not ideal, but if OS user authentication is implemented, you don't need a password. Alternatively, you could change permissions on this file so only you or root can access it, just be sure to run your bokeh serve with sudo if needed.

Customizations

See options.py if you would like to customize some of the plots in the main app, e.g., colors, font sizes, line dash, etc. These will be editable in settings.py in a future update. Currently, you need to relaunch your Bokeh server after editing options.py.