flow_example.rst

Advecting Flow Paths¶

Vapor can advect and render flow paths through your data.

In [1]:

import example_utils
from vapor import session, renderer, dataset, camera

ses = session.Session()
data = example_utils.OpenExampleDataset(ses)

Warning: sysroot "/Applications/Xcode_12.4.app/Contents/Developer/Platforms/MacOSX.platform/Developer/SDKs/MacOSX.sdk" not found (ignoring for now).

Vapor 3.9.2
Python 3.9.19 (/Users/pearse/miniconda3/envs/readTheDocs)
OpenGL 4.1 Metal - 83.1

Default advection¶

When you create a Flow renderer, the default advection will create a regularly dispersed grid of seed points and simulate a streamline advection in 2D or 3D, depending on the provided variables.

In [2]:

# Use first 2 2D variables as our U,V for the flow advection
dimension = 2
U,V = data.GetDataVarNames(dimension)[0:2]

ren:renderer.FlowRenderer = data.NewRenderer(renderer.FlowRenderer)
ren.SetFieldVariableNames([U, V])
ses.GetCamera().ViewAll()
ses.Show()

Rendering techniques¶

The rendering of the flow paths can be configured. A few examples are shown below. Call help(renderer.FlowRenderer) to see additional options.

In [3]:

ren.SetRenderType(ren.RenderType.RenderTypeStream)
ren.SetRenderRadiusScalar(3)
ren.SetRenderGeom3D(True)
ren.SetColorMapVariableName(U)
ses.Show()

Rather than rendering the flow lines, you can also render the sample points that are taken along the flow path individually.

In [4]:

ren.SetRenderType(ren.RenderType.RenderTypeSamples)
ses.Show()

In [5]:

ren.SetRenderGlyphType(ren.GlpyhType.GlpyhTypeArrow)
ren.SetRenderRadiusScalar(7)
ses.Show()

In [6]:

# Reset the rendering style for the next section
ren.SetRenderRadiusScalar(3)
ren.SetRenderType(ren.RenderType.RenderTypeStream)

Seeding the advection¶

The seeds from which the advected particles start can be configured in a variety of manners. Call help(renderer.FlowRenderer) to see additional options.

Uniform Distribution¶

This is the default. It creates a uniformly distributed grid of seed points over the area/volume of the advected region.

In [7]:

ren.SetSeedGenMode(ren.FlowSeedMode.UNIFORM)
rake = ren.GetRakeRegion()
defaultRakeExtents = rake.GetExtents()
rake.SetExtents((20, 20), (40, 40))
print(f"Seeding a {ren.GetGridNumOfSeeds()} grid over {rake.GetExtents()}")
ses.Show()
rake.SetExtents(*defaultRakeExtents)

Seeding a [5, 5] grid over ([20.0, 20.0, 0.0], [40.0, 40.0, 0.0])

Random Distribution¶

In [8]:

ren.SetSeedGenMode(ren.FlowSeedMode.RANDOM)
print(f"Seeding {ren.GetRandomNumOfSeeds()} random points over {rake.GetExtents()}")
ses.Show()

Seeding 50 random points over ([0.0, 0.0, 0.0], [63.0, 63.0, 63.0])

Biased Distribution¶

This generates a random list of seed points however they are biased by a given variable. The number of seed points can be higher or lower based on the value of the RakeBiasVariable. You can set the RakeBiasStrength to a negative value to invert the bias.

In [9]:

ren.SetSeedGenMode(ren.FlowSeedMode.RANDOM_BIAS)
ren.SetRakeBiasVariable(V)
ren.SetRakeBiasStrength(1)
print(f"Seeding {ren.GetRandomNumOfSeeds()} random points biased by {ren.GetRakeBiasVariable()} over {rake.GetExtents()}")
ses.Show()

Seeding 50 random points biased by V10 over ([0.0, 0.0, 0.0], [63.0, 63.0, 63.0])

Manual List¶

You can pass in a manually created list of seeds using a basic text file format. The code below generates an example seed file and passes it to Vapor. The full documentation for the seed file format can be found on Vapor's website.

In [10]:

with open("flow_seeds.txt", "w") as f:
    print("# X, Y, Z, T (optional)", file=f)
    print("10, 10, 0", file=f)
    print("20, 20, 0", file=f)

ren.SetSeedGenMode(ren.FlowSeedMode.LIST)
ren.SetSeedInputFilename("flow_seeds.txt")
ses.Show()

Outputting Flow Lines¶

The advected flow lines can be saved to a .csv file

In [11]:

ren.SetFlowlineOutputFilename("flow_output.csv") # Needs to be set once per renderer
ren.SetFlowOutputMoreVariables([U, V]) # Optionally specify additional variables to sample at each point

ren.SetNeedFlowlineOutput(True) # Needs to be called before each render call that should output the computed flow in a .csv file
_ = ses.RenderToImage() # This returns a PIL image variable that is not displayed by default and can be ignored if only saving flow output

In [12]:

import pandas as pd
pd.read_csv("flow_output.csv", sep=",").head(10)

Out[12]:

	# ID	X-position	Y-position	Z-position	U10	V10
0	0	10.000000	10.000000	0.0	1.612162	-0.012178
1	0	10.399255	9.996881	0.0	1.632684	-0.013170
2	0	10.803591	9.993517	0.0	1.653455	-0.014174
3	0	11.316035	9.988953	0.0	1.677380	-0.015548
4	0	11.966468	9.982625	0.0	1.705838	-0.017370
5	0	12.793562	9.973661	0.0	1.735683	-0.019941
6	0	13.845867	9.960633	0.0	1.766247	-0.023478
7	0	15.182375	9.941173	0.0	1.791113	-0.028411
8	0	16.870373	9.911378	0.0	1.800304	-0.035164
9	0	18.972748	9.864971	0.0	1.776475	-0.043920