Skip to content

Latest commit

 

History

History
132 lines (98 loc) · 5.23 KB

README.md

File metadata and controls

132 lines (98 loc) · 5.23 KB
Raw

vcs.js Build Status codecov

For documentation see the interface document.

To get started, first run npm install.

Useful npm commands:

npm run build                 # build the library to dist/vcs.js
npm run build-continuous      # build the library to dist/vcs.js and rebuild after changes
npm run test                  # run all tests against an actual server
npm run test:serve            # start test service (runs as files change)
npm run demo                  # serve the demo page at http://localhost:8080/demo/

Running the vcs.js demo

Executing the demo requires a running vtkweb service, which is assumed to exist at ws://localhost:9000/ws. A basic protocol implementation for demo purposes exists in this repository at scripts/vcs_server. You can start the server with the command PYTHONPATH=. vtkpython scripts/vcs-server -p 9000. Use the --help argument for a list of options. After starting the server, you can execute the demo using npm run demo and browsing to http://localhost:8080/demo.

If you get an error message refering to WAMP_FEATURES while building, try running the following:

npm run fix-autobahn

Running the vtkweb demo with vcs

To start the server run vtkpython demo1/vcs_server.py -p 1234. For runnig the client, use the same procedure as for vcs.js demo (npm run demo). If the client for vcs.js demo is already started, just browse to http://localhost:8080/demo1.

Running the vtkweb demo with VTK

To start the server run vtkpython demo1/vtk_web_cone.py -p 1234 To run the client use the same procedure as for vtkweb demo with vcs.

Running the vtkweb demo with ParaView

This is used to look at differences between vtkweb with ParaView server and with VTK server. To start the server run PV_ALLOW_BATCH_INTERACTION=1 ~/projects/ParaView/build/bin/pvbatch pv_server.py -p 1234 To run the client use the same procedure as for vtkweb demo with vcs.

Running the tests

In summary, tests are end-to-end and can be run in single shot or continuous mode, as long as we know where to find your uv-cdat conda environment. So either activate the environment (e.g. source activate 2.12) or just set the following environment variable:

export CONDA_PREFIX=/home/projects/miniconda3/envs/2.12

Then to run all the tests once you can run:

npm run test

Or to run in continuous mode for debugging you can run:

npm run test:serve

Some of the details

The testing framework uses karma, mocha, and chai, and provides an end-to-end testing solution to exercise the full path to the VCS backend, even comparing images delivered back to the client against known baselines. This approach has the benefite of exercising the vcs-js python code (rpc methods) as well as the client-side Javascript libraries.

For this to work, we must run a version of the python server with VCS modules available to import. Currently several npm commands are available to automate this entire process. All the tests can be run in a single shot, or can be run in continuous mode for debugging, using one of two npm scripts. The server and karma tests can also be started separately for more flexibility.

Debugging the tests

To do it all in a single command, you must first either activate the uv-cdat conda environment or just set the CONDA_PREFIX environment variable as described above. Then just run:

npm run test:serve

This will start a test python server and also start karma in watch mode, which brings up a Chrome browser. You can debug your tests by clicking the Debug button in the browser.

Currently coverage can be viewed in the terminal live, or by pointing your browser at the file system:

file:///home/projects/uvcdat/vcs-js/dist/coverage

and then navigating to the Chrome version folder.

Adding new tests/baselines

To add new tests, follow the existing examples in the project, which can be found at test/cases/vtkweb.js. Using the continuous karma testing approach, click on the Debug button to debug your new test. If it's an image comparison test, it should fail initially as there will be no baseline for it, but the karma debug tab will provide a link to save one.

Once you click the "Save baseline" link, copy it to the baselines folder, located at test/baselines/. Name it the same as the test, following the other examples there. If it is a new baseline for an existing test (i.e. one of multiple baselines), use the <testname>_01.png approach to name it.

Once you have copied the image, edit the test/baselines/index.js to include the new baseline. First you must import the baseline image, then add it to the baselineImages dictionary, following the other examples there.

At this point, the test should pass, assuming you did all of the above.