Create test cases for Port au Prince road data

[bwitham]

@mschicker Have you made any of these tests yet? If not and you still want to work on test 31 some more, I can work on this.

[mschicker]

@bwitham - I have a stack of items to make tests out of, based on conversations with Curran. Do you know if I can add the tests to the cases dir on github, or does that data need to go into hoot-tests?

[bwitham]

Ok, cool. If you already done work on this, then I'll work on something else.

hoot-tests is just for regression tests and any tests that reference FOUO data. There is no mechanism currently to run case tests from within hoot-tests, so if any those do reference FOUO data, you're probably out of luck for the time being. I'm guessing they don't though...Curran would know.

[mschicker]

So the word I'm getting is that the data I was hoping to use is not appropriate for github... Did you have any datasets in mind for this ticket?

[bwitham]

I only created this in response to what users and Curran reported as bad conflation on PAP data. So, the data is FOUO? I swear I saw an issue opened in VGI that referenced PAP data in github, but that was a long time ago so maybe I'm mistaken. If FOUO, only options would be to create a regression test out of it or write scripts in hoot-tests that execute conflate case tests.

[mschicker]

I just asked Curran, and he said that his understanding was that the MGCP PaP data isn't approved for "public release" - which means no to github.

HOWEVER, now that you mention it, I seem to recall that VGI issue. Maybe I'll go track it down.

[bwitham]

That's fine. Maybe just use as regression test data then. There could be a difference between posting a screenshot and actually putting the data in github. Another option to use the data in case tests would be to recreate similar geometries in brand new case test data without the original attribution and in a different AOI...that may be more trouble than its worth, though.

[curranMapper]

@bwitham @mschicker To my understanding those MGCP cells that are labled as unclassified (like in PoP) are still not to be added to github as they are public facing. Someone please correct me if that statement is not correct.

[bwitham]

Yeah, no problem. That's fine. We'll create tests and work with that data in the private hoot-tests repo.

[curranMapper]

Ah, yea, reading the above comms, let me get some further clarity, I may be still confused....

[drew-bower]

I believe all the Haiti MGCP was released for disaster support.

[mattjdnv]

Micah and I have been chatting and I think the Haiti datasets are the same ones as the GAIT Test data sets and these are approved for public release.

[drew-bower]

At one point you could download from esri.com.

[mattjdnv]

@drew-bower You are right. The FGDB is available on the ESRI site.

Also, I think we are OK with the GAIT data. From the README on another copy of the datasets: NOTE: I took out Georges email and phone number.

All these data sets have been publicly released.

The "GIFD" data set consists of NGA GIFD data for Montenegro accompanied by SRTM Level 1 elevation data.

The "MGCP" data sets are draft and final versions of MGCP cell W073N18 produced by Canada over Haiti.
Production of the draft version was expedited and publicly released to support earthquake relief efforts.
Final version has been through QC/QA and accepted in the International Geospatial Warehouse.  Normally, MGCP data is LIMDIS.  This is a unique exception. It is  accompanied by SRTM Level 1 elevation data.

The "LTDS" data set is a prototype produced and released by ESRI in Arc 10 geodatabase format. It 
is accompanied by a topo graphic (pdf) and USGS National Elevation Data (NED).

I can be reached at XXXXXX@XXXXXXXX or 703-XXX-XXXX
if there are questions.

George XXXXXXX
Institute for Defense Analyses

[mschicker]

Merged #1619