DiPhyxDiPhyx
Workspace
The idea

A filter takes dataand returns data

That is the entire contract. Filters chain, nothing is modified in place, and the pipeline you build is the record of what you did.

They chain

A filter's output is another filter's input, and the browser on the left is the chain drawn out.

Nothing is destroyed

The source stays as it was. Delete a branch and everything above it is untouched.

They re-execute

Change a parameter upstream and everything below recomputes. This is why order affects speed so much.

Cutting

Slice, Clipand Extract Subset

Three ways to look at less. They are not interchangeable, and choosing wrongly is the usual reason a picture is confusing.

SliceAn infinitely thin plane through the data. You get a 2D surface — good for a clean cross-section.
ClipThrows away everything on one side of a plane and keeps the volume of the rest. Use it to see inside.
Extract SubsetKeeps a box of cells, or every Nth cell. The right first move when a dataset is too slow to work with.
Finding shape

Contour, Thresholdand Stream Tracer

The filters that turn a field of numbers into something with an outline.

Contour

An isosurface at a chosen value. Needs point data — run Cell Data To Point Data if it is greyed out.

Threshold

Keeps cells inside a value range. Unlike Contour it keeps volume, so it is both analysis and reduction.

Glyph

An arrow or a sphere per point, scaled and oriented by an array. Reduce the point count first.

Stream Tracer

Integrates paths through a vector field. The seed placement matters more than any other setting.

Contour needs point data, not cell data. If it is greyed out, run Cell Data To Point Data first — this catches nearly everyone once, and the interface does not explain why. Stream Tracer needs a vector array and a seed source; the default line seed is rarely where you want it, and moving it is usually the difference between a useful picture and spaghetti.

New fields, and order

Compute what you mean,then reduce early

The filter that derives the quantity you actually want, and the habit that decides how long the whole pipeline takes.

Python Calculator
# vorticity magnitude from a velocity field
mag(curl(inputs[0].PointData['U']))
CalculatorArithmeticsmall expressions over arrays
Python CalculatorNumPygradients, conditions, anything
GradientBuilt invorticity and Q-criterion directly

Calculator handles arithmetic on arrays with a small expression language. Python Calculator gives you NumPy over the same arrays, which is what you want for gradients, magnitudes and anything conditional. Gradient Of Unstructured DataSet computes vorticity and Q-criterion directly, and is faster than doing it by hand.

Cut before you computeA Calculator over a clipped region does a fraction of the work of one over the whole volume.
Build against a subsetExtract Subset first, get the pipeline right, then change the input back to the full dataset at the end.
Save the pipelineA .pvsm state file, or a traced Python script. Both let you point the same pipeline at the next run.
On a machine

ParaView isalready built

Nothing here needs compiling. Pick a machine, and these two commands put it on there.

hub://paraview

$ dxflow workflow create --identity paraview hub://paraview

$ dxflow workflow start paraview

Start a machineWhat ParaView needs
Read it, then run itPrepaid and by the hour, on a machine that is yours about a minute after you ask.
Start a machineAll notes