Native implementation of the wrappers and translators by the SimPhoNy team at Fraunhofer IWM.
The GMSHWrapper is a SimPhoNy-package for ontologized modelling of finite-element geometries (.stl-format), volume calculation and mesh calculations of mold-models as inputs for CFD-simulations, using GMSH.
Contact: Matthias Büschelberger from the Material Informatics team, Fraunhofer IWM
The GMSHWrapper is currently able to automatically generate meshes for rectangles and cylinders by providing the numerical values of their lengths and resolution. The metric SI-units of these quantities (such as mm, cm or dm) will be defined via String-input. If the want to model the filling of a mold which is represented by a mesh, you can also provide a numerical value for the filling_fraction between 0-1. This value is associated to the relative z-value of the domain.
Apart from that, the wrapper provides the functionality to read an already available .stl-file with any arbitrary geometry-abstraction and to calculate its bulk 3D-volume as well as the volume of the 3D-mold-filling. This is achieved by calculating the sum of all determinates for each triangular facet after Zhang & Chen (2001).
In addition to the volume-calculations, the maximum extents in xyz-directions of these forms are provided in the resulting CUDS-objects after the execution of the wrapper.
We recommend to use the GMSHWrapper in combination with the CFDWrapper for SimPhoNy, since the ontologized mesh-data in the form of CUDS can be used in order to automatically create domains for CFD-simulations with e.g. OpenFOAM®.
Please use a git-version >= 2.20 and python >= 3.6.12 for the installation.
We tested the wrapper for the GMSH-backend mainly on Ubuntu>=18.04 and Debian GNU/Linux 10. If you consider to use the same os, you will need to install the following additional packages via apt-get:
apt-get update && apt-get install -y ffmpeg libsm6 libxext6 libglu1-mesa-dev libgl1-mesa-glx
Now clone the wrapper-repository.
git clone https://github.com/simphony/GMSHWrapper.git
cd GMSHWrapper
pip install -r requirements.txt
python setup.py install
The installation of the namespaces of the EMMO CFD into osp-core will be automatically executed via pico install during this step.
The GMSHWrapper is expecting an osp-core-version >= 3.5.2. A compatible version will be installed via pip during the setup of the CUDSTranslator.
See the dependency matrix:
| Python Package | Minimum Version |
|---|---|
osp-core |
3.5.2-beta |
CUDSTranslator |
v0.1 |
EMMO-CFD |
v0.1 |
The package is providing scripts for unittesting in the tests-directories. Run via:
python -m unittest
For the following examples, we are considering the following python-imports:
from osp.wrappers.gmsh_wrapper.gmsh_session import GMSHSession
from osp.wrappers.gmsh_wrapper.gmsh_cuds_translator import Rectangle, Cylinder, Complex
from osp.core.namespaces import emmo, cuba
In the first example, we like to generate a .stl-file for a rectangle through semantic data structures of RDF and CUDS within GMSHSession.
For the instantiation of CUDS-objects, we use the Rectangle-class which is a child from the CUDSTranslator.
# start session
with GMSHSession() as session:
# initialize wrapper
wrapper = cuba.Wrapper(session=session)
# create rectangle
rec = Rectangle(
'path/to/my/output/directory',
values={
'x': 20,
'y': 10,
'z': 150,
'filling_fraction': 0.5,
'resolution': 1
},
units={
'lengths': "mm",
'resolution': "mm"
},
session=session
)
# get the CUDS if emmo.MeshGenetation
my_model = rec.get_model()
# add the CUDS to the wrapper
wrapper.add(my_model, rel=emmo.hasPart)
# run the computation
session.run()
# get the CUDS of type emmo.MeshGeneration again
my_model_with_outputs = wrapper.get(oclass=emmo.MeshGeneration)[0]
The python-variable of my_model is the CUDS-objects of type MeshGeneration, which is linked to the ontologized parameters such as lengths, os-paths, resolution etc.
The variable of my_model_with_outputs is the exacly same CUDS of type MeshGeneration, but with the derived meta-data from the computation. The .stl of interest has been generated under the desired output-directory.
In the second example, we would like to generate a .stl-file for a cylinder:
with GMSHSession() as session:
# initialize wrapper
wrapper = cuba.Wrapper(session=session)
# create cylinder
cyl = Cylinder(
'path/to/my/output/directory',
values={
'length': 20,
'radius': 5,
'filling_fraction': 0.5,
'resolution': 0.14
},
units={
'lengths': "cm",
'resolution': "cm"
},
session=session
)
# get the CUDS of type emmo.MeshGenetation
my_model = cyl.get_model()
# add the model to the wrapper
wrapper.add(my_model, rel=emmo.hasPart)
# run the computation
session.run()
# get the CUDS of type emmo.MeshGeneration again
my_model_with_outputs = wrapper.get(oclass=emmo.MeshGeneration)[0]
Now, we may print the results CUDS with the results of the computation:
from osp.core.utils import pretty_print
pretty_print(my_model_with_outputs)
The results of this pretty_print may look like that:
- Cuds object:
uuid: 85dd67da-b904-4ee7-9980-7089e709fdec
type: emmo.MeshGeneration
superclasses: emmo.Calculation, emmo.Computation, emmo.EMMO, emmo.Holistic, emmo.Item, emmo.MeshGeneration, emmo.Perspective, emmo.Physical, emmo.Process
description:
To Be Determined
|_Relationship emmo.hasInput:
| - emmo.GeometryData cuds object:
| uuid: fe058209-c4cd-4322-8246-aa80902eadee
| |_Relationship emmo.hasQuantitativeProperty:
| | - emmo.Volume cuds object:
| | uuid: cf3bf314-1dc7-40fb-8b8f-dd427949eca6
| | |_Relationship emmo.hasQuantityValue:
| | | - emmo.Real cuds object:
| | | uuid: f96b30e3-b447-4268-a974-ebf3b9a4a146
| | | hasNumericalData: 3e-05
| | |_Relationship emmo.hasReferenceUnit:
| | - emmo.CubicMetre cuds object:
| | uuid: d755b63f-da2d-4008-96aa-e9fcddfa6043
| |_Relationship emmo.hasXLength:
| | - emmo.Length cuds object:
| | uuid: 73cb4f47-9aed-4583-9c75-6930c1769fee
| | |_Relationship emmo.hasQuantityValue:
| | | - emmo.Real cuds object:
| | | uuid: 7c3715f8-02b1-426f-a140-f84372fcc6d0
| | | hasNumericalData: 20
| | |_Relationship emmo.hasReferenceUnit:
| | - emmo.Metre cuds object:
| | . uuid: 20cb1885-0a68-4ff0-9dce-8b687264465e
| | . hasSymbolData: m
| | . |_Relationship emmo.hasPhysicalDimension:
| | . - emmo.LengthDimension cuds object:
| | . uuid: 5a2fcc3e-bd8d-4114-a4a3-1e279c02f34b
| | - emmo.Milli cuds object:
| | uuid: df556889-5dfc-443a-9494-01defb51dc0e
| | hasSymbolData: m
| |_Relationship emmo.hasYLength:
| | - emmo.Length cuds object:
| | uuid: 96c8ec21-ba56-403e-ba9b-dfbcbd123d09
| | |_Relationship emmo.hasQuantityValue:
| | | - emmo.Real cuds object:
| | | uuid: 52544bcc-f8ad-4f22-b13e-1c1d47875303
| | | hasNumericalData: 10
| | |_Relationship emmo.hasReferenceUnit:
| | - emmo.Metre cuds object:
| | . uuid: 8de6607f-30b1-4c23-a4a9-b20964bcc194
| | . hasSymbolData: m
| | . |_Relationship emmo.hasPhysicalDimension:
| | . - emmo.LengthDimension cuds object:
| | . uuid: 13389cdb-9feb-45dd-8b6f-903344cbe190
| | - emmo.Milli cuds object:
| | uuid: 3b838849-091f-4530-8ca5-2c6fcdc31ea6
| | hasSymbolData: m
| |_Relationship emmo.hasZLength:
| | - emmo.Length cuds object:
| | uuid: d6060369-1c9e-4324-81dd-9acb539f120f
| | |_Relationship emmo.hasQuantityValue:
| | | - emmo.Real cuds object:
| | | uuid: 3c355bcd-37be-4410-8be1-22215b31f397
| | | hasNumericalData: 150
| | |_Relationship emmo.hasReferenceUnit:
| | - emmo.Metre cuds object:
| | . uuid: a0ccb3d1-bc72-401c-80ca-c39304844cbe
| | . hasSymbolData: m
| | . |_Relationship emmo.hasPhysicalDimension:
| | . - emmo.LengthDimension cuds object:
| | . uuid: b7a8836e-765d-4653-9383-f10fda339495
| | - emmo.Milli cuds object:
| | uuid: ea57e239-88f9-4bd1-8c5e-0b92027fa9ab
| | hasSymbolData: m
| |_Relationship emmo.standsFor:
| - emmo.Rectangle cuds object:
| uuid: c1f05768-bd1c-4533-b16f-39b3029ab653
|_Relationship emmo.hasOutput:
| - emmo.MeshData cuds object:
| uuid: d93dc3f2-9157-4ce6-b52b-f47c2f62a669
| |_Relationship emmo.hasSign:
| | - emmo.File cuds object:
| | uuid: 28f7a919-1572-4c77-a04d-9e70a0484e32
| | |_Relationship emmo.hasSign:
| | - emmo.DirectorySequence cuds object:
| | . uuid: 69ef1fcb-cc15-422d-99da-363cc6d4a2e2
| | . |_Relationship emmo.hasSpatialDirectPart:
| | . | - emmo.Directory cuds object:
| | . | . uuid: 25fbcc95-c3e7-41d6-b4a2-3b1a94dc9749
| | . | . |_Relationship emmo.hasSign:
| | . | . - emmo.String cuds object:
| | . | . uuid: cb949958-9821-4ce5-bc58-e8c86ef5b8d5
| | . | . hasSymbolData: tmpy1cnsxds
| | . |_Relationship emmo.hasSpatialFirst:
| | . - emmo.Directory cuds object:
| | . uuid: 4a329c23-cc43-49d6-98fe-5d8b8f92dd1e
| | . |_Relationship emmo.hasSign:
| | . | - emmo.String cuds object:
| | . | uuid: 31836603-ff5b-44da-aa95-cb3506b1998b
| | . | hasSymbolData: /tmp
| | . |_Relationship emmo.hasSpatialNext:
| | . - emmo.Directory cuds object:
| | . uuid: 25fbcc95-c3e7-41d6-b4a2-3b1a94dc9749
| | . (already printed)
| | - emmo.STL cuds object:
| | . uuid: a422a69a-53bd-4bf4-9479-6e7390fbda64
| | - emmo.String cuds object:
| | uuid: edc8c517-bd9a-46ea-b7e0-a9e50ae48bfc
| | hasSymbolData: new_surface
| |_Relationship emmo.hasSpatialDirectPart:
| - emmo.Resolution cuds object:
| uuid: d1b7039e-648b-4d30-94ab-820dc1c61971
| |_Relationship emmo.hasQuantityValue:
| | - emmo.Real cuds object:
| | uuid: cdca569a-9e40-43ff-9b82-f80bba14f31e
| | hasNumericalData: 1
| |_Relationship emmo.hasReferenceUnit:
| - emmo.MeterPerUnitOne cuds object:
| . uuid: 94d610c7-3c76-4193-a691-1ee1df94b2e6
| . hasSymbolData: "m/cell"
| . |_Relationship emmo.hasPhysicalDimension:
| . - emmo.ResolutionDimension cuds object:
| . uuid: fa706511-42d9-415c-b10d-6715b96f4cf4
| - emmo.Milli cuds object:
| uuid: 3e096513-bfbe-4235-b634-6cbf87ba61d2
| hasSymbolData: m
|_Relationship emmo.hasProperParticipant:
- emmo.FillingData cuds object:
. uuid: 6bdfa23e-77b1-4876-8409-7c6bfb19b845
. |_Relationship emmo.hasSpatialDirectPart:
. - emmo.FillingFraction cuds object:
. . uuid: a65b0245-6578-4bb2-9e6e-70f015ea2cc3
. . |_Relationship emmo.hasQuantityValue:
. . | - emmo.Real cuds object:
. . | uuid: 6562d18c-a981-458d-83c8-c53db9d56b4e
. . | hasNumericalData: 0.5
. . |_Relationship emmo.hasReferenceUnit:
. . - emmo.VolumeFractionUnit cuds object:
. . uuid: 03366452-4cfe-4bae-8ae9-3b9f9df9e0c9
. . |_Relationship emmo.hasPhysicalDimension:
. . - emmo.DimensionOne cuds object:
. . uuid: a1a778c6-6ecd-4358-85ac-e3d156873859
- emmo.GMSH cuds object:
. uuid: 32b4a27a-5daf-4611-8cd6-81a49daf8261
Further details are provided in the examples-directory of this repository.
For the further use of the CUDS-objects with respect to osp-wrappers for the semantic interoperability to third-party tools, please visit the SimPhoNy-Organisation on GitHub or the Fraunhofer-GitLab.
If you want to run the GMSHWrapper as a flask-app, we can provide you also runner-script with this repository. For the following steps, we assume that you already setup the wrapper and its dependencies as explained under the Installation-section.
If so, then simply run:
cd <gmshwrapper-repo>/flask
python RUN.py
The service will then listen under 0.0.0.0:7000.
From now on, you may run the wrapper remotely and instantiate the CUDS locally. In order to send the CUDS to the server with the wrapper listening via flask, you may simply use the TransportSessionClient from osp.core and the WrapperSession as its base, since the GMSHSession is also based on the WrapperSession.
First of all, import the following lines:
from osp.core.session.transport.transport_session_client import TransportSessionClient
from osp.core.session import WrapperSession
The examples shown in the Useage and examples-section also work with the flask-app, you simply have to exchange the following lines:
with TransportSessionClient(WrapperSession, '0.0.0.0', 7000) as session:
wrapper = cuba.Wrapper(session=session)
...
...
...
With 0.0.0.0 and 7000, we define the host and port, the flask-server the wrapper is listening to. The rest of the example(s) stay(s) exactly the same.
We provide a Docker image for the simplified and automated setup of this wrapper. You should have Docker>=20.10.5 installed on your os.
For building the image, simply run the following command under the top level directory of the wrapper-repository:
docker build -t <wrapper-image-name>:<tag> .
If you want to run the built image, simply run:
docker run -dit <wrapper-image-name>:<tag>
The container is then providing the mentioned flask-app from above as ENTRYPOINT under its port 7000.
- FORCE; Grant agreement number: 721027
The package is licensed under GNU General Public License v3.