uw-macrostrat / gplates-reconstruct Goto Github PK

In reference to paleobiodb/data_service#11, the data service should return back modern coordinates as part of the GPlates service for Holocene rotations.

Hi,

Is it possible to change the default rotation model from Wright(2013) to Scotese(2016)?
The 2016 update is big change from the 2013 model:
UpdateRotationModel_MacrostratReconstruct.pptx

Here are the links to the 2016 model:
"url": "https://www.earthbyte.org/paleomap-paleoatlas-for-gplates/",
"data_url": "ftp://ftp.earthbyte.org/earthbyte/PaleoAtlas_Scotese_v3.zip",

Thanks!
Ananya

We will want to provide parallel paths (essentially a clone of this service) for the Seton EarthByte model and other EarthByte models as they are released.

I have been communicating with @jonhusson about adding additional rotation models such as the Torsvik and Cocks (2017) model:

http://www.earthdynamics.org/earthhistory/Data_Software.html
http://www.earthdynamics.org/earthhistory/bookdata/CEED6.zip

In this model, however, 000 represents the mantle reference frame rather than the spin axis reference frame. 001 is the spin axis reference frame (i.e. the reference frame one would wish to reconstruct to using gplates-reconstruct). The pygplates.reconstruct() function takes an anchor_plate_id argument (default is 0).

One solution for this would be to add the default anchor_plate_id to models.json. The .json for Torsvik and Cocks could look like:

{
  "name": "torsvik2017",
  "citation": "Torsvik, T. and Cocks, L.R.M. 2017. Earth History and Palaeogeography. Cambridge University Press",
  "url": "http://www.cambridge.org/9781107105324",
  "data_url": "http://www.earthdynamics.org/earthhistory/bookdata/CEED6.zip",
  "date_acquired": "2017-10-30",
  "isbn": "9781107105324",
  "min_age": 0,
  "max_age": 500,
  "coverage": "continental",
  "models": [{
      "min_age": 0,
      "max_age": 500,
      "rotation_file": "Torsvik_Cocks_HybridRotationFile.rot",
      "anchor_plate_id": 1,
  }]
}

that anchor_plate_id could be extracted in models.py and then fed into the rotate function within reconstruct.py

def rotate(features, age, model):
    if len(models[model]['models']) == 1:
        rotation_model = models[model]['models'][0]['rotation']
        anchor_plate  =  models[model]['models'][0]['anchor_plate_id']

    for rot in models[model]['models']:
        if rot['min_age'] <= age and rot['max_age'] >= age:
            rotation_model = rot['rotation']

    rotated = []
    pygplates.reconstruct(features, rotation_model, rotated, age, anchor_plate)
    return rotated

Only use the 0-550 model and plates.

The data service cannot process larger features, and will return an error that the request is too long.

Example GeoJSON:
https://macrostrat.org/api/places?format=geojson_bare&name=Wisconsin&placetype=region

Example Rotation Attempt:
curl -X POST -F my.geojson -F -o rotated_my.geojson -- https://dev.macrostrat.org/reconstruct

Error Type:
Sometimes it will return the textfile rotate_my.geojson with the following error message.

The server encountered an internal error and was unable to complete your request. Either the server is overloaded or there is an error in the application.

Other times Bash will return a curl error that the argument list is too long.

I tried to transfer a geoJSON file using the "reconstruct an entire file" code and it's not working. Can anyone tell why I got the error that my "argument list is too long"?

The links to the original gplates files in the README are dead/non-existent.