Re: [netcdfgroup] How to dump netCDF to JSON?

  • To: Chris Barker <chris.barker@xxxxxxxx>
  • Subject: Re: [netcdfgroup] How to dump netCDF to JSON?
  • From: John Readey <jreadey@xxxxxxxxxxxx>
  • Date: Thu, 20 Oct 2016 21:20:27 +0000
  • Authentication-results: spf=none (sender IP is ) smtp.mailfrom=jreadey@xxxxxxxxxxxx;
  • Spamdiagnosticmetadata: NSPM
  • Spamdiagnosticoutput: 1:99
Right, a reasonable approach would be to use the hdf5-json code to output json 
for datasets and attributes, but arrange it by groups in a tree hierarchy.  All 
I suspect that you’d want dimension scales treated as a distinct object type.

John

From: Chris Barker <chris.barker@xxxxxxxx>
Date: Thursday, October 20, 2016 at 1:48 PM
To: John Readey <jreadey@xxxxxxxxxxxx>
Cc: Pedro Vicente <pedro.vicente@xxxxxxxxxxxxxxxxxx>, netCDF Mail List 
<netcdfgroup@xxxxxxxxxxxxxxxx>, HDF Users Discussion List 
<hdf-forum@xxxxxxxxxxxxxxxxxx>
Subject: Re: [netcdfgroup] How to dump netCDF to JSON?

On Thu, Oct 20, 2016 at 12:02 PM, John Readey 
<jreadey@xxxxxxxxxxxx<mailto:jreadey@xxxxxxxxxxxx>> wrote:
So we came up with a scheme of Group, Dataset, and Datatype collections with a 
UUID to identify each object.  That way if you a reference to a specific UUID, 
you can always access the object regardless of what shenanigans may be 
happening with the links in the file.

It’s true that this makes path look ups a bit more cumbersome, but it’s a more 
general way of specify a directed graph (the HDF5 link structure) on a tree 
(the JSON hierarchy).

Hmm -- interesting. I hadn't realized that HDF was this flexible. For my part, 
I've only really used netcdf.

This is making me think that we may want a spec for netcdf-json that would be a 
subset of the hdf-json spec.

That way they can be as compatible as possible without "cluttering up" the 
netcdf spec too much.

-CHB





John

From: Pedro Vicente 
<pedro.vicente@xxxxxxxxxxxxxxxxxx<mailto:pedro.vicente@xxxxxxxxxxxxxxxxxx>>
Date: Tuesday, October 18, 2016 at 9:37 PM
To: John Readey <jreadey@xxxxxxxxxxxx<mailto:jreadey@xxxxxxxxxxxx>>, Chris 
Barker <chris.barker@xxxxxxxx<mailto:chris.barker@xxxxxxxx>>
Cc: netCDF Mail List 
<netcdfgroup@xxxxxxxxxxxxxxxx<mailto:netcdfgroup@xxxxxxxxxxxxxxxx>>, HDF Users 
Discussion List 
<hdf-forum@xxxxxxxxxxxxxxxxxx<mailto:hdf-forum@xxxxxxxxxxxxxxxxxx>>

Subject: Re: [netcdfgroup] How to dump netCDF to JSON?

@John

>> 1.       Complete fidelity to all HDF5 features
>> 2.       Support graphs that are not acyclic.

ok, understood.

In my case I needed a simple schema for a particular set of files.

But why didn't you start with the official HDF5 DDL

https://support.hdfgroup.org/HDF5/doc/ddl.html

and try to adapt to JSON?

Same thing for netCDF, there is already an official CDL, so any JSON spec 
should be "identical".



@Chris

{
"dset1" : ["dataset", "STAR_INT32", 2, [3, 4], [1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 
11, 12]]
}

>> * Do you need "rank"?

sometimes a bit of redundancy is useful, to make it visually clear

>> BTW, is a "dataset" in HDF the same thing as a "variable" in netcdf?)

yes

>>It would be really great to have this become an "official" spec -- if you 
>>want to get it there, you're probably going to need to develop it more out in 
>>the open with a wider community. These lists are the way to get that started, 
>>but I suggest
>>1) put it up somewhere that people can collaborate on it, make suggestions, 
>>capture the discussion, etc. gitHub is one really nice way to do that. See, 
>>for example the UGRID spec project:


ok, anyone interested send me an off list  email


-Pedro



----- Original Message -----
From: John Readey<mailto:jreadey@xxxxxxxxxxxx>
To: Chris Barker<mailto:chris.barker@xxxxxxxx> ; Pedro 
Vicente<mailto:pedro.vicente@xxxxxxxxxxxxxxxxxx>
Cc: netCDF Mail List<mailto:netcdfgroup@xxxxxxxxxxxxxxxx> ; Charlie 
Zender<mailto:zender@xxxxxxx> ; HDF Users Discussion 
List<mailto:hdf-forum@xxxxxxxxxxxxxxxxxx> ; David 
Pearah<mailto:David.Pearah@xxxxxxxxxxxx>
Sent: Tuesday, October 18, 2016 11:15 PM
Subject: Re: [netcdfgroup] How to dump netCDF to JSON?

Hey,

The hdf5-json code is here: https://github.com/HDFGroup/hdf5-json and docs are 
here:  http://hdf5-json.readthedocs.io/en/latest/.

The package is both a library of HFD5 <-> JSON conversion functions and some 
simple scripts for converting HDF5 to JSON and vice-versa.  E.g.
$ python h5tojson.py –D <hdf5-file>
outputs JSON minus the dataset data values.

While it may not be the most elegant JSON schema, it’s designed with the 
following goals in mind:

1.       Complete fidelity to all HDF5 features (i.e. the goal is that you 
should be able to take any HDF5 files, convert it to JSON, convert back to HDF5 
and wind up with a file that is semantically equivalent to what you started 
with.

2.       Support graphs that are not acyclic.  I.e. a group structure like 
<root> links with A, and B.  And A and B links to C.  The output should only 
produce one representation of C.
Since NetCDF doesn’t use all these features, it’s certainly possible to come up 
with something simpler for just netCDF files.

Suggestions, feedback, and pull requests are welcome!

Cheers,
John

From: Chris Barker <chris.barker@xxxxxxxx<mailto:chris.barker@xxxxxxxx>>
Date: Friday, October 14, 2016 at 12:32 PM
To: Pedro Vicente 
<pedro.vicente@xxxxxxxxxxxxxxxxxx<mailto:pedro.vicente@xxxxxxxxxxxxxxxxxx>>
Cc: netCDF Mail List 
<netcdfgroup@xxxxxxxxxxxxxxxx<mailto:netcdfgroup@xxxxxxxxxxxxxxxx>>, Charlie 
Zender <zender@xxxxxxx<mailto:zender@xxxxxxx>>, John Readey 
<jreadey@xxxxxxxxxxxx<mailto:jreadey@xxxxxxxxxxxx>>, HDF Users Discussion List 
<hdf-forum@xxxxxxxxxxxxxxxxxx<mailto:hdf-forum@xxxxxxxxxxxxxxxxxx>>, David 
Pearah <David.Pearah@xxxxxxxxxxxx<mailto:David.Pearah@xxxxxxxxxxxx>>
Subject: Re: [netcdfgroup] How to dump netCDF to JSON?

Pedro,

When I first started reading this thread, I thought "there should be a spec for 
how to represent netcdf in JSON"

and then I read:

1) The specification to convert netCDF/HDF5 to "a" JSON format (note the "a" 
here)

Awesome -- that's exactly what we need -- as you say there is not one way to 
represent netcdf data in JSON, and probably far more than one "obvious" way.

Without looking at your spec yet, I do think it should probably look as much 
like CDL as possible -- we are all familiar with that.

(why Python? HDF5 developer tools should be all about writing in C/C++)

Because Python is an excellent language with which to "drive" C/C++ libraries 
like HDF5 and netcdf4. If I were to do this, I'd sure use Python. Even if you 
want to get to a C++ implementation eventually, you'd probably benefit from 
prototyping and working out the kinks with a Python version first.

But whoever is writing the code....


The specification is here

http://www.space-research.org/

Just took a quick look -- nice start.

I've only used HDF through the netcdf4 spec, so there may be richness needed 
that I'm missing, but my first thought is to a make greater use of "objects" in 
JSON (key-value structures, hash tables, dicts in python), rather than array 
position for heterogeneous structures. For instance, you have:

 a dataset

{
"dset1" : ["dataset", "STAR_INT32", 2, [3, 4], [1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 
11, 12]]
}

I would perhaps do that as something like:

{
...
"dset1":{"object_type": "dataset",
         "dtype": "INT32"
         "rank": 2,
         "dimensions": [3,4],
         "data": [[1,2,3,4],
                  [5,6,7,8],
                  [9,10,11,12]]
         }
...
}

NOTES:

* I used nested arrays, rather than flattening the 2-d array -- this maps 
nicely to things like numpy arrays, for example -- not sure about the C++ 
world. (you can flatten and un-flatten numpy arrays easily, too, but this seems 
like a better mapping to the structure) And HDF is storing this all in chunks 
and who knows what -- so it's not a direct mapping to the memory layout anyway.

* Do you need "rank"? -- can't you check the length of the dimensions array?

* Do you  need "object_type" -- will it always be a dataset? Or you could have 
something like:

{
...
"datasets": {"dset1": {the actual dataset object},
             "dset2": {another dataset object},
 ....
}

Then you don't need object_type or a name


(BTW, is a "dataset" in HDF the same thing as a "variable" in netcdf?)

I would like to make this some kind of "official" netCDF/HDF5 JSON format for 
the community, so I encourage anyone to read the specification

If you see any flaw in the design or anything in the design that you would like 
to have change please let me know now

done :-)

It would be really great to have this become an "official" spec -- if you want 
to get it there, you're probably going to need to develop it more out in the 
open with a wider community. These lists are the way to get that started, but I 
suggest:

1) put it up somewhere that people can collaborate on it, make suggestions, 
capture the discussion, etc. gitHub is one really nice way to do that. See, for 
example the UGRID spec project:

  https://github.com/ugrid-conventions/ugrid-conventions

(NOTE that that one got put on gitHub after there was a pretty complete draft 
spec, so there isn't THAT much discussion captured. But also note that that is 
too bad -- there is no good record of the decision process that led to the spec)

At the moment it only (intentionally) uses common generic features of both 
netCDF and HDF5, which are the numeric atomic types and strings.

Good plan.

-Chris


--

Christopher Barker, Ph.D.
Oceanographer

Emergency Response Division
NOAA/NOS/OR&R            (206) 526-6959<tel:%28206%29%20526-6959>   voice
7600 Sand Point Way NE   (206) 526-6329<tel:%28206%29%20526-6329>   fax
Seattle, WA  98115       (206) 526-6317<tel:%28206%29%20526-6317>   main 
reception

Chris.Barker@xxxxxxxx<mailto:Chris.Barker@xxxxxxxx>



--

Christopher Barker, Ph.D.
Oceanographer

Emergency Response Division
NOAA/NOS/OR&R            (206) 526-6959   voice
7600 Sand Point Way NE   (206) 526-6329   fax
Seattle, WA  98115       (206) 526-6317   main reception

Chris.Barker@xxxxxxxx<mailto:Chris.Barker@xxxxxxxx>
  • 2016 messages navigation, sorted by:
    1. Thread
    2. Subject
    3. Author
    4. Date
    5. ↑ Table Of Contents
  • Search the netcdfgroup archives: