OVis is a desktop application that allows you to visualize the data you've recorded using OpenMDAO's SQLite recorder. To use, simply open OVis, select your SQLite database file, and use the N² diagram to view model hierarchy and plots to see iteration history.

OVis is available for Windows, Mac, and Ubuntu.

1. Go to the OVis's releases and download the correct installer for your operating system from the most recent release.

   • Windows: OVis.Setup.x.y.z.exe
   • Mac: OVis-x.y.z.dmg
   • Ubuntu: OVis-x.y.z-x86_64.AppImage

2. Once downloaded, double-click to run the installer. Note that, for Windows, we do not yet sign the installer. As a result, opening the file will likely result in a "Windows protected your PC" popup. Click "More info" then "Run anyway" to proceed with installation.

3. Follow the installer's instructions. Once complete, you should be able to run the OVis executable.

OVis is written in a combination of Node.js, JavaScript, HTML, and CSS using the Electron framework. Testing uses Mocha, Chai, and Spectron with babel-node and isparta for code coverage. To begin developing OVis:

1. Install Node.js and npm
2. Run git clone https://github.com/OpenMDAO/ovis.git
3. Go to the cloned folder with cd ovis
4. Run npm install to install all application dependencies.
   • Note: I've noticed there can be problems installing the sqlite3 library on macOS. If you run into this problem, try running the command again with Python 2 active.

If all goes well, you should now be able to start OVis with npm start.

Electron applications run in two processes: the main process and the renderer process, with communication between these being performed using a built-in IPC module. For OVis, the entirety of the logic for the main process can be found in main.js, which is responsible for application startup, file selection, and automatic updating. The renderer process handles plotting, N² diagram logic, layout, and interfacing with the SQLite file.

Note: Direction of arrows indicate the direction of calls/requests (the presentation layer calls methods in the logic layer, the logic layer makes requests of the data layer, etc.)

The renderer process can be split into a Node.js backend and a HTML/CSS/JavaScript frontend (formerly the python HTTP server).

The backend - found in the src/data_server folder - implements an explicit three-tier architecture. The backend's data layer queries the database, the logic layer transforms the data into a more useful form, and the presentation layer presents an API for requesting data. The interface between the backend and frontend can be found in server.js.

The frontend is primarily implemented in vanilla JavaScript and includes plotting, layout, and the N² diagram.

As mentioned above, this application uses the Mocha JavaScript test framework with Chai for assertions and isparta/babel-node for code coverage. Spectron is used in end-to-end testing to run the application.

• Run tests with npm test
• Run tests and coverage with npm run test-cov

Currently, coverage isn't reported for the vanilla JS frontend tested using the end-to-end Spectron tests. Rather than going through the effort to get it to report coverage for for those files, it would probably be best to convert them to Node.js and add unit tests.

To ensure that new test files are run with the testing command, add them directly or in a sub-directory of the test folder.

OVis takes advantage of the built-in electron-builder and electron-updater for build and updating, respectively, and GitHub's Releases for release.

Building OVis is as simple as running npm run build. This command builds the installers and associated files for all supported operating systems and stores them in the dist folder.

If you intend to build a Windows installer on macOS or Linux, make sure you install Wine first. You may also need to install the Wine dependency XQuartz if you're building on a mac.

• To build for macOS only: npm run build:m
• To build for Windows only: npm run build:w
• To build for Ubuntu only: npm run build:l

Note: Due to a bug in electron-builder, the AppImage produced on Windows and macOS will fail. The Ubuntu AppImage must be produced on a Linux machine.

Note: Similar to installing on macOS, there can be problems building the sqlite3 module. If you run into this problem, try running the command again with python2 active.

Note: macOS requires that the installer be signed using verified certificates. If the certificates are imported into the keychain on your mac, electron-builder will automatically sign the application as a part of the build process.

Releases are handled on the OpenMDAO OVis repository's Releases on GitHub.

1. Navigate to the OVis repository's releases
2. Click "Draft a new release"
3. Add a tag version and attach the files in the dist directory that were generated when building
   • Do not attach the mac, win-unpacked, or linux-unpacked folders. These are intermediary files that can be discarded
4. Click "Publish release" when you're finished

OVis implements automatic updating so that users don't need to manually download new versions. Nothing needs to be done to make this work with each release.

A python HTTP server (see github-release-http-server.py) running on WebFaction pings GitHub every hour for new releases. If a new version is available, this server downloads it from GitHub. OVis pings this python server on startup to determine if a new version is available and, if it is, downloads the update in the background and installs it when the application is closed.