High-level API overview

The Python programming interface gives you programmatic access to most of OVITO’s program features. Using Python scripts, you can perform many of the things you know from the interactive user interface (and even more):

  • Import from and export data to simulation files

  • Use modifiers to set up a data processing pipeline

  • Render images or movies of the three-dimensional scene

  • Access result data computed by an analysis pipeline

  • Write your own data modification functions

OVITO’s data pipeline architecture

If you have worked with OVITO’s graphical user interface before, you should already be familiar with the data pipeline concept: After loading a simulation file into the program, you typically build up a data pipeline by applying one or more modifiers that act on the data in some way. The result of this sequence of modifiers is dynamically computed and displayed in the interactive viewports of OVITO.

The Python interface of OVITO lets you construct data pipelines in the same way by setting up a sequence of modifiers and configuring their parameters. The data pipeline is represented by an instance of the Pipeline Python class, which manages the list of applied modifiers. Furthermore, a data pipeline always has some type of data source, which is a separate object providing or generating the input data that will be passed to the pipeline’s first modifier. Typically, the data source is an instance of the FileSource class, which is responsible for loading the data from an external simulation file.


If you want to visualize the results of a Pipeline, you need to place the pipeline into the scene, i.e. the three-dimensional space visible in rendered images. Only pipelines that have explicitly been inserted into the current scene by calling the add_to_scene() method will show up in rendered images or OVITO’s interactive viewports. Note that this happens automatically in the graphical user interface of OVITO, but you need to do it explicitly when using the Python scripting interface.

The scene, including all explicitly inserted pipelines and other state information that would get saved along in a .ovito file, is represented by the Scene Python class. A script always runs in the context of one current Scene instance, which is accessible through the ovito.scene global variable.

Importing data from disk

You typically create a new Pipeline by importing a simulation data file from disk using the ovito.io.import_file() function:

from ovito.io import import_file
pipeline = import_file("simulation.dump")

This high-level function creates a new Pipeline instance and wires it to a FileSource, which will take care of loading the data from the given input file into memory. The FileSource object is accessible through the pipeline’s source field:

>>> print(pipeline.source)
<FileSource at 0x7f9ea70aefb0>

The FileSource may be reconfigured to load a different input file if desired, allowing you to replace the original input of the pipeline with new data. This is useful if you intend to batch-process a number of simulation files, reusing the same processing pipeline. The File I/O section gives more details on how to import and export data using the scripting interface.

Applying modifiers

Initially, the Pipeline created by the import_file() function contains no modifiers. That means it will simply pass through the original, unmodified data loaded by the FileSource from disk. We can change this by inserting some modifiers into the pipeline’s modifiers list:

from ovito.modifiers import *
pipeline.modifiers.append(ColorCodingModifier(property = 'Potential Energy'))
pipeline.modifiers.append(SliceModifier(normal = (0,0,1)))

Modifiers are constructed by instantiating one of the built-in modifier classes, which are all contained in the ovito.modifiers Python module. Note how a modifier’s parameters can be initialized in two different ways:


When creating a new object such as an OVITO modifier, you can directly initialize its parameters by passing keyword arguments to the constructor function. Thus

pipeline.modifiers.append(CommonNeighborAnalysisModifier(cutoff=3.2, only_selected=True))

is equivalent to assigning the values to the parameter fields one by one after constructing the object:

modifier = CommonNeighborAnalysisModifier()
modifier.cutoff = 3.2
modifier.only_selected = True

Obviously, the first method of initializing the parameters is more convenient and is the recommended way whenever the parameter values are known at construction time of the object.

Keep in mind that it is possible to change the parameters of modifiers in a pipeline at any time, or to remove modifiers from a pipeline again. This option is useful if you want to sequentially process the input data in multiple different ways. Alternatively, you can also set up multiple data pipelines, all sharing the same data source or even some of the modifiers. The Data pipelines section provides more information on working with data pipelines and modifiers.

Exporting data to disk

Once a Pipeline has been set up, you can pass it to the ovito.io.export_file() function to let OVITO compute the result of the pipeline and write it to an output file in the given format:

from ovito.io import export_file
export_file(pipeline, "outputdata.dump", "lammps/dump",
    columns = ["Position.X", "Position.Y", "Position.Z", "Structure Type"])

The export_file() function takes the output filename and the desired format as arguments, in addition to the Pipeline generating the data to be exported. Furthermore, depending on the selected format, additional keyword arguments such as the list of particle properties to export must be provided. See the documentation of the export_file() function and this section for more information.

Accessing computation results

Instead of directly piping the computation results to an output file, you can also request the pipeline to return a DataCollection object, which represents the output data leaving the pipeline:

data = pipeline.compute()

The compute() method performs two things: It first requests the input data from the pipeline’s data source. Then, it lets all modifiers of the pipeline act on the data, one by one. The final data state is returned to the caller as a DataCollection, which essentially is a heterogeneous container storing a set of data objects that each represent different parts of a dataset:

>>> data.objects
[SimulationCell(), Particles(), AttributeDataObject(), AttributeDataObject()]

In the example above, the DataCollection.objects list contains a SimulationCell object, a Particles object, and several global attribute objects, which were either loaded from the source data file of the pipeline or which were generated on the fly by modifiers in the pipeline.

The DataCollection class provides various fields for accessing particular kinds of data objects. For example, the cell field returns the SimulationCell object storing the simulation cell vectors and position of the cell origin as a matrix:

>>> print(data.cell[...])
[[ 148.147995      0.            0.          -74.0739975 ]
 [   0.          148.07200623    0.          -74.03600311]
 [   0.            0.          148.0756073   -74.03780365]]

The particles field returns the Particles container object, which manages all particle properties:

>>> print(data.particles.positions[...])
[[ 73.24230194  -5.77583981  -0.87618297]
 [-49.00170135 -35.47610092 -27.92519951]
 [-50.36349869 -39.02569962 -25.61310005]
 [ 42.71210098  59.44919968  38.6432991 ]
 [ 42.9917984   63.53770065  36.33330154]
 [ 44.17670059  61.49860001  37.5401001 ]]

The Particles section provides more information on this topic.

Accessing a pipeline’s input data

Sometimes it may be necessary to access the original data that enters a pipeline. The input data is read from the input file by the pipeline’s FileSource. This object also provides a compute() method returning a DataCollection:

input_data  = pipeline.source.compute()
output_data = pipeline.compute()

Rendering images and movies

As mentioned earlier, it is necessary to add the Pipeline to the three-dimensional scene to visualize data. This is done by invoking its add_to_scene() method, typically right after creating a new pipeline:

pipeline = import_file("simulation.dump")

Furthermore, to render an image or a movie, a Viewport object is required, which defines the viewpoint from which the scene is seen:

from ovito.vis import Viewport
vp = Viewport()
vp.type = Viewport.Type.Perspective
vp.camera_pos = (-100, -150, 150)
vp.camera_dir = (2, 3, -3)
vp.fov = math.radians(60.0)

The Viewport’s parameters control the position and orientation of the virtual camera, the type of projection (perspective or parallel), and the field of view (FOV) angle. To automatically position the camera such that all objects in the scene are fully visible, you may call the viewport’s zoom_all() method. Finally, the Viewport.render_image() method must be called to render an image and save it to disk:

vp.render_image(filename="myimage.png", size=(800,600))

Note that as part of the rendering process, all pipelines inserted into the current scene will automatically be evaluated.

Further reading

The following sections provide more information on various aspects of OVITO’s Python scripting interface: