Add external demo website based on public docker image

[Mletter1]

once #740 is done we will be able to complete this

[Mletter1]

alright our image is ready see https://hub.docker.com/r/slycat/slycat-developer/

[alexsielicki]

I just finished putting an external demo slycat site: https://myslycat.com (slycat.com was already taken)

It's currently still showing a certificate warning and will keep doing that until my new certificate arrives. In the meantime, feel free to play around and let me know if you find any problems.

[alexsielicki]

FYI, Slycat needs about 1GB of RAM. I couldn't get it to work on a 500MB VM. Got the following error on CouchDB startup:

ll_alloc: Cannot allocate 494927872 bytes of memory (of type "port_tab"). Aborted (core dumped)

Looks like Couch wants 500MB memory just for itself.

[alexsielicki]

The new certificate is now installed and seems to be working. You should no longer be getting warnings from the browser.

[pjcross]

The example CCA model for TAIS that we provide needs to be changed to make the inputs be ajo, bjo, njo, and beta. The outputs need to be ndrf_last and ndhr_last.

[pjcross]

We should designate the above variables as input/output columns in the Parameter Space model as well. The color-coding helps to understand the data, which will be unfamiliar to users. Also I'd like to have the description for the Parameter Space TAIS model be as follows:

Taylor Anvil Impact Scenario (TAIS) SIERRA/SM CATALYST DAKOTA. SIERRA/SM CATALYST DAKOTA ENSEMBLE ANALYSIS of an Oxygen Free High Conductivity (OFHC) Copper Cylinder of Length 2.54 cm and Diameter 0.762 cm with initial velocity of 190 m/sec impacts a rigid wall. A Johnson-Cook inelastic constitutive law was used with the finite element analysis code, SIERRA/SM (i.e., "presto"), to numerically predict the response and final deformed shape. The Dakota parameter sensitivity study considered four parameters of the Johnson-Cook constitutive law, AJO, BJO, NJO, and BETA (the fraction of plastic work converted to heat).

[pjcross]

In the cars Parameter Space example, inputs are cylinders, displacement, weight, year, and origin. Outputs are MPG, Horsepower, and Acceleration. Model is neither since it is really just a label. Since cylinders, year and origin have small, fixed numbers of value choices, I want to make them be categorical so that we have an example of buttons in the filtering.

[pjcross]

I tried creating a bookmark for the cars Parameter Space model, but it darkened the screen and never brought up the dialog. The same thing happened when I tried to rename a model. I could see a black line (like the edge of a dialog box) and the affordance to pull the corner, which allowed me to make a large white square from the line. However, none of the fields were present to do anything with it. I could exit this state by backing up the browser.

[pjcross]

We should either remove the help button ("?") or add a real email address (currently it goes to nobody@example.com).

[alexsielicki]

Which email should we use for the help button?

[pjcross]

Why don’t you put my email. The slycat-help address is limited to Sandia, so we can’t use it.

[alexsielicki]

Creating bookmarks and renaming models work for me from home on Safari and Firefox on macOS. Info works as well. Perhaps it's a Windows issue or Sandia network?

[alexsielicki]

I swapped in your email address, @pjcross

I'm using mine for the error emails.

[Mletter1]

907b99a as of this commit we have a new docker image, we just need to add some example models to it then we can put it up on the website and commit the image to dockerhub as well