This file is a condensed form of the PreonLab changelog. It contains all backwards incompatible changes and deprecations regarding the preonpy API interface and the command line arguments of PreonLab, PreonCLI and PreonNode. It also lists changes of supported environments, like Python versions or operating systems. In some cases it also lists new features that are useful for IT related tasks. The target audience for it are people responsible for deploying our products on workstations or clusters or developing automatization workflows.
This document does not list changes in the feature set of PreonLab in terms of its simulation or pre- or postprocessing capabilities. For this, please consider the changelog document.
--gpuSelection parameter to select which GPUs
should be used for the calculations.GPU Single Node Unlimited license type in the respective
CLI, PreonPy APIs, and PreonLab UIs.global fluid temperature to
global reference temperature.orientation constraint property
that replaces and extends the previous fixed view axis
property. It also now allows fixing the viewing direction of the camera,
which is enabled by default for orthographic cameras.Thermal Sensor: Solid Temperature to
Thermal Sensor: Temperature.Boundary Condition: to the
boundary condition buffer names (ex: renamed Temperature to
Boundary Condition: Temperature).get_image method to objects of type
FloatFieldObject that allows to get the texture data.get_texture3d method to rigids that allows to
get the coloring texture data, e.g., when a force sensor is
connected.do_not_load_resources parameter when opening
a scene. This allows opening a scene without resources that might take a
long time to load, e.g., geometries, pathlines, or air flow data. To be
able to load this data later, a load_initial_resources
method was added to the scene.preonpy.plugin_args(…) interface as a way to pass
configuration to plugins. This can be used to select which GPUs should
be used for the calculations.names_of_measured_values() method for sensors now
excludes quantities for which it is known that they are currently not
available.Scene.create_object method now accepts the
typename parameter in the new common naming scheme, where
each word starts with a capitalized letter. Old names will continue to
be accepted for some time to ensure a smoother transition.Object.get_connected_objects(…) or
Object.get_connection_slots(…) were called without setting
include_target_slot_name to True. We don’t
plan to change the default value in future versions.preonpy module. This mainly prevents the use of misspelled
attribute names.as_scalars of the PreonPy
method Object.statistic_buffers(…) to be
False. Therefore, scripts that do not specify that
parameter will behave the same as in version 7.0 when it is
removed.export_mesh() method for pathlines now allows
specifying an export unit. The default export unit is meters.load_only_object_creations=True, now
preonpy.DataNotAvailable is raised instead of a
TypeError.Full license was renamed to the Base
license in version 6.0. Starting with version 7.0, the PreonPy API and
the PreonCLI and PreonNode command-line interfaces will remove support
for the old naming scheme.as_scalars of the PreonPy method
Object.statistic_buffers(…) is deprecated and will be
removed in version 7.0.load_only_object_creations of
preonpy.Scene(…) is deprecated and will be removed in
version 7.0. Please consider using do_not_load_resources
instead.scene_path as a
positional argument to preonpy.Scene(…) is deprecated.
Starting with version 7.0, such a parameter will only be accepted as a
keyword argument.preonpy.current_scene has been
deprecated since version 5.0. It will be removed in version 7.0. Please
use get_scenes() and preonlab.current_scene
instead.preonpy.show_progressbar has been
deprecated since version 5.2. It will be removed in version 7.0. Please
use preonpy.show_progress instead.--continue parameter that allows a
previously canceled or otherwise stopped simulation to be continued
based on the last written scene statistics.--featureOnGPU parameter
for activating the GPU support.GPU Memory statistic to the “Scene” object,
which tracks the amount of memory spent while simulating on the GPU.
Note that this does not account for memory spent on visualization (e.g.,
rendering).--lBoost=sim_threads_max and not enough licenses are
available to cover all threads, it will cancel with an error message. In
earlier releases, only a warning was printed, and the simulation started
with fewer simulation threads than the system would provide. The user
should specify the exact number of additional simulation thread licenses
to cover the whole machine (e.g., --lBoost=10).OMP_NUM_THREADS is set to a value that is higher than the
number of available cores. As before, the number of actually used
threads will be reduced to the number of available cores. But now a
warning is logged, and the number of simulation thread licenses
requested will align with that number.Full license is now called the Base
license to better express its extendability with a GPU
license or licenses that extend the number of CPU simulation threads in
MPI contexts. The renaming is visible in the GUI of PreonLab, but also
in the preonpy.checkout_license function of the PreonPy API
and in the -licenseMain argument of
PreonCLI/PreonNode.preonpy.LicenseTypes.GPU to be used in
preonpy.checkout_license(...) calls and to be returned in
preonpy.get_license() calls.preonpy.set_features to enable and disable
features. This is currently mainly used to enable GPU usage when PreonPy
is used as a package within a user-provided Python environment,
e.g. without using PreonCLI.statistic_names() will now return the
names of all statistics, even if they have no data available.
Furthermore, the function will only return one statistic name per
vector-valued statistic. Example: Only Linear Velocity will
be returned instead of Linear Velocity Norm,
Linear Velocity X, Linear Velocity Y and
Linear Velocity Z.adaptive spacing to
continuous particle size, min. spacing to
min. particle size, spacing to
particle size/max. particle size,
coarse particles to coarsen particles and
source spacing to particle size. These
renamings were introduced in version 5.2.get_statistic(…), get_statistic_avg(…),
get_statistic_min(…) and get_statistic_max(…))
will be removed in version 6.1. The new Statistic API, which can be
reached via Object.statistic(…), is now recommended to be
used instead.--licenseModelBoost
command-line parameter and for PreonPy the
license_model_boost parameter for the
checkout_license function.preonpy.get_license function now returns a fourth
tuple element containing the boost license model.Total Wait Time now accumulates the time
of the node that had to wait the longest each simulation step.statistic_names() will change from 6.0.0
on. Then, it will return names of all statistics, even if they have no
data available. Furthermore, the function will only return one statistic
name per vector-valued statistic. Example: Only ‘Linear Velocity’ will
be returned instead of ‘Linear Velocity Norm’, ‘Linear Velocity X’,
‘Linear Velocity Y’ and ‘Linear Velocity Z’.PathLike objects such as
pathlib.Path in all places where previously only
str was accepted as a path.convergence criterion to
thermal timestep limiter.FlowRate to
VolumeFlowRate.heat capacity to
specific heat capacity and flow rate to
volume flow rate.Licensed Threads to
Tot. Licensed Threads.Scene.get_files for retrieving all existing files
that belong to the scene.preonpy.show_progress attribute which is
enabled per default. When enabled, it prints a status log per frame for
long-running tasks such as simulate,
post_process or exporting a CSV file. The behaviour
substitutes the previously shown progress bar for the long-running
tasks, which required the progressbar2 package. This
requirement is not needed anymore. Previously, the non-existence of the
progressbar2 package disabled logging, which is not true
anymore. The preonpy.show_progress attribute should be used
instead of preonpy.show_progressbar attribute, which is
deprecated and will be removed in future versions.particleBasedLoadBalancing which is
now the default. Furthermore, the guiding particle-count-based metric
was optimized to predict the computational effort more accurately.timingBasedLoadBalancing which enables
the timings-based load balancer in previous PreonLab versions. This is
now only recommended when using MPI nodes that vary in computation
speed.mode and version field in the status
file (--statusFile). The mode field can be of
value postprocess or simulate, depending on
which type of run was invoked. version will have the value
1.1 and will update with future changes to this file.--licenseList.Decimeter and
Inch, as well as for the temperature unit
Fahrenheit.mm).
Previously, some properties and methods expected the unit to be
represented by its full name (e.g., Millimeter).particle_buffers was called on a solid with
dynamic sampling turned on, a RuntimeError was
raised. Now a preonpy.SubsamplingError is raised
instead.Fluids to
Solvers.thermodynamics. It replaces the old
system type property. If the removed system type
isolated is loaded from an old scene, thermodynamics is set
to Off in this case.system type to
boundary type.temperature to
constant temperature.ClosedFixed to
temperature. The property constant temperature
can only be set as a boundary condition if this particular system type
is set.Heat flux sensor to
Thermal sensor.physical value from the force
sensor. The decision on which physical value to display is now defined
by the coloring mode alone.car body mass to
sprung mass.reference point from
axleLoad and uncompressedSpring to
compressed state and uncompressed state,
respectively.mapped unit to
mapped.mapped from
NewtonPerMeter and Newton to
Coefficient and Force, respectively.Front and
Rear, respectively. The PreonPy names of the properties in
these two groups include the group name (e.g. “Front::”).parameter input. The per-wheel
suspension parameters have to be provided for the wheels at the front
axle and the wheels at the rear axle separately.close_existing argument to the
preonpy.Scene constructor to specify what should happen
with currently opened scenes. Also preonpy.multiple_scenes
can be used to set the default behavior if close_existing
was not passed.preonpy.set_log_dir and
preonpy.get_log_dir for specifying where log files should
be written.set_default_scene_dir and
get_default_scene_dir for specifying where new scenes
should be created.current_scene property, the
function get_scenes() (and for the PreonLab Console the
property preonlab.current_scene) was added.ClosedFixed to
constant temperature for solids.thermodynamics. It replaces the old
system type property.preonpy.multiple_scenes = False or pass
close_existing=True to the preonpy.Scene
constructor to get the old behavior again.import preonpy anymore.current_scene will be removed in
one of the upcoming releases. Please use get_scenes() and
preonlab.current_scene.live csv export to
live CSV export.csv file to
CSV file. This read-only property is only available in
PreonPy.noninear damper coefficient (rear) to
nonlinear damper coefficient (rear).BoundaryDomain slot.smoothShaded to
visible.TemperatureBased is now called
temperature. You can still use the old names in
PreonPy.Mesh particle sampling method for
solids.record statistics property.specify thread affinity option from the
scene settings.temperature has been removed from the
solver. If you want to set a temperature for all currently existing
particles of a solver, use the new right-click action
Set temperature.compute initial deflection.timestep property of the scene.--noPerformanceChecks option
is given). Before, this error was shown only with three or more
instances.preonpy.BufferSet got renamed to
preonpy.BufferSetRef.precipitation height ("mm/min") got renamed to
precipitation height.intial Young Modulus
to Young Modulus.load_only_object_creations
flag, the scene statistics are not loaded by default. Please use the new
reload_statistics method.