| Inheritance |
|
You can pass this object to pdg.GraphContext._cook() to configure how the
graph cooks.
Methods ¶
checkpointingEnabled
: bool
Property
Enables writing a JSON-based checkpoint file while the graph cooks, which contains the work item and dependency data for the graph
checkpointPath
: str
Property
The file path that the checkpoint data should be written to. If the path
ends with the .gz extension, for example $HIP/checkpoint.json.gz, the file
will be compressed using GZip.
checkpointRate
: int
Property
The rate at which the checkpoitn file is updated. This rate is specified as the number of work items that should be queued in the checkpoint manager before flushing it to disk – it isn’t a time value.
cookTime
: float
Property
The time that the graph is evaluating at.
Defaults to 0. When a PDG graph is cooked from TOPs, the value is set
based on the Evaluation Time parameter on the TOP network.
cookType
: pdg.cookType
Property
The type of cook to perform.
Defaults to pdg.cookType.Full.
nodeNames
: list
of str
Property
The names of the nodes to cook, as an array of strings. Used when the
cookType value is set to pdg.cookType.Node or pdg.cookType.GenerateNode.
saveScene
: bool
Property
When this property is set to True, PDG will attempt to save the scene
before cooking, if any of the nodes being cooked need the scene file. For
example, the ROP Fetch TOP node creates work items that load the current
scene file when they cook.
Defaults to False.
staticFilter
: pdg.Filter
Property
The set of the static items that should be cooked.
Defaults to all static items.
Methods from pdg.EvaluationOptions ¶
autoRecook
: bool
Property
When this option is set to True, the graph will automatically recook
at the interval specified in the autoRecookInterval property.
Defaults to False.
autoRecookInterval
: int
Property
When auto recooking, the interval in milliseconds between recooks.
Defaults to 100.
blocking
: bool
Property
Indicates whether or not the cook should block the caller, or cook in the background.
Defaults to False