Notebook as Resource

[valentinedwv]

Is your feature request related to a problem? Please describe. open a binder from a github url, pass a datafile to download, call dispatcher, call notebook

[smrgeoinfo]

So, what do we need in the dataset and application (jupyter notebook) metadata to make this work? The original DDS dispatcher was implemented to take a URL for an ISO metadata record, inspect the gmd:distributions to match formats with formats for input to an application.

The current geoCodes workflow uses :

1. application description for useful input formats, in resource registry

   "supportingData":{
   "@type":"DataFeed",
   "name":"Input Data Type specification", "position":"input",
   "encodingFormat":["http://www.opengis.net/doc/CS/las/1.4","application/octet-stream;type=ASPRS-LAS","text/plain; application=esri-asciigrid","application/vnd.esri-asciigrid","Point Cloud Data"]
   },

   and

2. Datasets should have a /distribution/encodingFormat that uses a format identifier that matches a supportingData/encodingformat for a registered application.

[MBcode]

We have had this for opening datasets for some time, and have it for opening collections, though the template/.. still needs to be taken from my working test version; We need to look at that next (medium difficulty) notebook test case. Much of the dispatch can be done in the notebook, as related 'tools' can be found by a call to the ec utils, and there are also utils to help marshal the data as well

[smrgeoinfo]

what are the use cases here, e.g.:

1. I want to do a particular kind of analysis; I find an application (notebook) to do that analysis, then I need to find data for the area/context I am interested in, get the data in a format the application can read, and do my science. Goal-- minimize user mouse clicks between finding the app and working with data they are interested in.
2. I find a data set and want to explore the content via graphics, or compare to other data . Goal-- minimize mouse clicks between finding data and having some useful visualization on the screen.

other user goals like getting data into a local environment to work on a problem of interest are not notebook related-- should implement through search UI and representation/distribution options.

[valentinedwv]

There is a folder where I have done some brain dumps. Expand, create a new scenario.

we need to do a review of what is being done in the community, before we run down a path... review the Earthcube Meeting notebooks, and see what we need to do help what the community is already doing.

The collection interface needs to expose the 'distributions' and let the user select which ones to use and what input for a tool that it should be associated with. already some functionality in the client api: https://github.com/earthcube/facetsearch/blob/f472029ba8d398eb0f21c5cc792b7875c3a72978/server/routes/index.js#L66

[MBcode]

Since the notebook/utils can do all the dataset-page matching, and more, incl. a sort of dispatch, which could include sub-notebooks after they are generalized and their descriptions are indexed; though it could be any unit of computation. lmk if we can get back to this