Difference between pygplates.TopologicalSnapshot and pygplates.resolve_topologies

Q&A
1

Hello everyone:

May I ask what are the differences between pygplates.TopologicalSnapshot and pygplates.resolve_topologies? Specifically, can you please give examples of when I create a pygplates.TopologicalSnapshot versus when I call function pygplates.resovle_topologies?

Many thanks.

2

Hi Lavie,

You can use either in any situation. It’s just that pygplates.TopologicalSnapshot is a more object-oriented approach, which has some benefits.

To compare, with pygplates.resolve_topologies() you could resolve some plate polygons to 10Ma and export to a Shapefile with something like:

# Resolve topologies and export in one go.
pygplates.resolve_topologies(
    'dynamic_plate_polygons.gpml',
    'rotations.rot',
    'resolved_plate_polygons_10Ma.shp',
    reconstruction_time=10)

…whereas the equivalent with pygplates.TopologicalSnapshot would be:

# First resolve topologies (at 10Ma).
snapshot_10 = pygplates.TopologicalSnapshot(
    'dynamic_plate_polygons.gpml',
    'rotations.rot',
    10)

# Then export (using the 'snapshot_10' object).
snapshot_10.export_resolved_topologies('resolved_plate_polygons_10Ma.shp')

But if you wanted to do more, then you might have to call pygplates.resolve_topologies() twice (both calls at 10Ma). In the following example, we want to export resolved networks to a Shapefile and return them as a Python list so we can query/process them further. This requires resolving the topologies twice (at 10Ma):

# Resolve topologies and export only resolved networks in one go.
pygplates.resolve_topologies(
    'plate_polygons_and_networks.gpml',
    'rotations.rot',
    'resolved_networks_10Ma.shp',
    reconstruction_time=10,
    resolve_topology_types=pygplates.ResolveTopologyType.network)

# Resolve topologies again and return only resolved networks (in a list) in one go.
resolved_networks = []
pygplates.resolve_topologies(
    'plate_polygons_and_networks.gpml',
    'rotations.rot',
    resolved_networks,
    reconstruction_time=10,
    resolve_topology_types=pygplates.ResolveTopologyType.network)

…whereas with pygplates.TopologicalSnapshot we only need to resolve topologies once (at 10Ma):

# Resolve topologies only once.
snapshot_10 = pygplates.TopologicalSnapshot(
    'plate_polygons_and_networks.gpml',
    'rotations.rot',
    10)

# Then export (only resolved networks).
snapshot_10.export_resolved_topologies(
    'resolved_networks_10Ma.shp',
    pygplates.ResolveTopologyType.network)

# Then return resolved networks (in a list).
resolved_networks  = snapshot_10.get_resolved_topologies(
    pygplates.ResolveTopologyType.network)

Furthermore, multiple pygplates.TopologicalSnapshot’s (one for each reconstruction time) can be generated from a single pygplates.TopologicalModel, like this:

# Topological model can be used for any reconstruction time.
topological_model = pygplates.TopologicalModel(
    'plate_polygons_and_networks.gpml',
    'rotations.rot')

# Iterate over reconstruction times [0, 1, ..., 100].
for reconstruction_time in range(100+1):

    # Generate a snapshot at the current reconstruction time.
    topological_snapshot = topological_model.topological_snapshot(reconstruction_time)

    # Then export snapshot (only resolved networks).
    snapshot.export_resolved_topologies(
        'resolved_networks_{}Ma.shp'.format(reconstruction_time),
        pygplates.ResolveTopologyType.network)

    # Then get resolved networks (in a list) from snapshot.
    resolved_networks  = snapshot.get_resolved_topologies(
        pygplates.ResolveTopologyType.network)
3

Thank you very much for your detailed response.