generators issueshttps://git.rwth-aachen.de/groups/monticore/EmbeddedMontiArc/generators/-/issues2018-10-25T12:09:27+02:00https://git.rwth-aachen.de/monticore/EmbeddedMontiArc/generators/EMAM2Middleware/-/issues/1Linux: Linked Armadillo does not work2018-10-25T12:09:27+02:00Alexander David HellwigLinux: Linked Armadillo does not workReplace with copied build installReplace with copied build installAlexander David HellwigAlexander David Hellwighttps://git.rwth-aachen.de/monticore/EmbeddedMontiArc/generators/EMAM2Middleware/-/issues/2CMake generated from Cpp generator does not find Armadillo in installed in Ar...2018-10-25T12:09:13+02:00Alexander David HellwigCMake generated from Cpp generator does not find Armadillo in installed in Armadillo_HOMEAlexander David HellwigAlexander David Hellwighttps://git.rwth-aachen.de/monticore/EmbeddedMontiArc/generators/CNNArch2Caffe2/-/issues/2Fix CI2018-09-01T13:40:04+02:00Evgeny KusmenkoFix CIadd maven dependenciesadd maven dependenciesCarlos Alfredo Yeverino RodriguezCarlos Alfredo Yeverino Rodriguezhttps://git.rwth-aachen.de/monticore/EmbeddedMontiArc/generators/EMAM2Cpp/-/issues/25StreamTest matrix ports input value setting2018-08-16T17:39:48+02:00Sascha Niklas SchneidersStreamTest matrix ports input value settingThe "component.${portName} = ..." part was not generated correctly for matrices used in models like detection.ObjectDetector1The "component.${portName} = ..." part was not generated correctly for matrices used in models like detection.ObjectDetector1Sascha Niklas SchneidersSascha Niklas Schneidershttps://git.rwth-aachen.de/monticore/EmbeddedMontiArc/generators/EMAM2Cpp/-/issues/24StreamTest Matrix output generation check not working properly2018-08-16T17:39:40+02:00Sascha Niklas SchneidersStreamTest Matrix output generation check not working properlyThe "REQUIRE(..)" part was not generated correctly for matrices used in models like detection.ObjectDetector1The "REQUIRE(..)" part was not generated correctly for matrices used in models like detection.ObjectDetector1Sascha Niklas SchneidersSascha Niklas Schneidershttps://git.rwth-aachen.de/monticore/EmbeddedMontiArc/generators/EMAM2Cpp/-/issues/19C++ float division2018-08-08T15:51:32+02:00Christoph RichterC++ float divisionFloating point division delivers wrong results in C++.
`Q x = 1 / 2`;
The value of `x` should be 0.5.
The generated C++ code looks like:
`double x = 1 / 2`;
In this case C++ performs an integer division, because `1` and `2` are ha...Floating point division delivers wrong results in C++.
`Q x = 1 / 2`;
The value of `x` should be 0.5.
The generated C++ code looks like:
`double x = 1 / 2`;
In this case C++ performs an integer division, because `1` and `2` are handled as Integer values. The (wrong result is `x == 0`).Christoph RichterChristoph Richterhttps://git.rwth-aachen.de/monticore/EmbeddedMontiArc/generators/EMAM2Cpp/-/issues/17Column Vector is not initialized if declared with 2 dimensions2018-08-08T13:46:26+02:00Christoph RichterColumn Vector is not initialized if declared with 2 dimensionsFor EMAM statement `Q^{n,1} a;` the following C++ code is generated: `colvec a;` This vector can be used in EMAM but is not initialized in C++.For EMAM statement `Q^{n,1} a;` the following C++ code is generated: `colvec a;` This vector can be used in EMAM but is not initialized in C++.Christoph RichterChristoph Richterhttps://git.rwth-aachen.de/monticore/EmbeddedMontiArc/generators/EMADL2CPP/-/issues/2Generator Factory2018-08-02T14:12:40+02:00Evgeny KusmenkoGenerator FactoryWe need a factory to create a train generator and an architecture generator,
see Abstract Factory Pattern
The Factories should be defined in their respective projects, i.e. CNNArch2Caffe2 should contain a Caffe2GeneratorFactory; the CN...We need a factory to create a train generator and an architecture generator,
see Abstract Factory Pattern
The Factories should be defined in their respective projects, i.e. CNNArch2Caffe2 should contain a Caffe2GeneratorFactory; the CNNArch2MXNet Generator project should contain the MXNetFactory.Carlos Alfredo Yeverino RodriguezCarlos Alfredo Yeverino Rodriguezhttps://git.rwth-aachen.de/monticore/EmbeddedMontiArc/generators/EMAM2Cpp/-/issues/18Matrix containing variables is created as constant2018-07-27T10:42:35+02:00Christoph RichterMatrix containing variables is created as constantIf a matrix definition contains a variable it is defined as constant. This creates wrong results if the matrix was modified in any situation. Also ports are always passed as a constant value.
Example:
```
ports in Q in1,
...If a matrix definition contains a variable it is defined as constant. This creates wrong results if the matrix was modified in any situation. Also ports are always passed as a constant value.
Example:
```
ports in Q in1,
out Q out1;
implementation Math{
Q^{3} A = [0, in1, 1];
out1 = sum(A);
}
```
creates C++ code:
```
class test_portInMatrixDefinition{
public:
double in1;
double out1;
rowvec CONSTANTCONSTANTVECTOR0;
void init()
{
CONSTANTCONSTANTVECTOR0 = rowvec(3);
CONSTANTCONSTANTVECTOR0(0,0) = 0;
CONSTANTCONSTANTVECTOR0(0,1) = in1;
CONSTANTCONSTANTVECTOR0(0,2) = 1;
}
void execute()
{
colvec A = CONSTANTCONSTANTVECTOR0;
out1 = (accu(A));
}
```
https://git.rwth-aachen.de/monticore/EmbeddedMontiArc/generators/EMAM2Cpp/-/issues/14Vector colon expression generates invalid code!2018-07-18T17:57:22+02:00Christoph RichterVector colon expression generates invalid code!Expressions:
```
Q^{1,10} A = 1:10;
Q^{1,5} B = 1:2:10;
```
generates the following invalid expression:
```
rowvec A = ;
rowvec B = ;
```Expressions:
```
Q^{1,10} A = 1:10;
Q^{1,5} B = 1:2:10;
```
generates the following invalid expression:
```
rowvec A = ;
rowvec B = ;
```Christoph RichterChristoph Richterhttps://git.rwth-aachen.de/monticore/EmbeddedMontiArc/generators/EMAM2Cpp/-/issues/13Element-wise multiplication does not generate correct armadillo code!2018-07-18T17:56:53+02:00Christoph RichterElement-wise multiplication does not generate correct armadillo code!Math statements:
```
Q^{2,2} A = [1,1;1,1];
Q^{2,2} B = [1,2;3,4];
Q^{2,2} C = A .* B;
```
produce wrong armadillo C++ code:
`mat C = A .* B;`
Armadillo uses % for element wise multiplication.Math statements:
```
Q^{2,2} A = [1,1;1,1];
Q^{2,2} B = [1,2;3,4];
Q^{2,2} C = A .* B;
```
produce wrong armadillo C++ code:
`mat C = A .* B;`
Armadillo uses % for element wise multiplication.Christoph RichterChristoph Richterhttps://git.rwth-aachen.de/monticore/EmbeddedMontiArc/generators/visualisation/-/issues/6All Layers except of Layer 3 are kind of defect2018-07-16T23:55:24+02:00vonWencksternAll Layers except of Layer 3 are kind of defectThe third layer looks pretty good: https://embeddedmontiarc.github.io/VisualizationEMAM/models/PacMan/index.html#de.rwth.pacman.pacManWrapper.controller.coneSearch:1
But the other layers look terrible:
1. https://embeddedmontiarc.github...The third layer looks pretty good: https://embeddedmontiarc.github.io/VisualizationEMAM/models/PacMan/index.html#de.rwth.pacman.pacManWrapper.controller.coneSearch:1
But the other layers look terrible:
1. https://embeddedmontiarc.github.io/VisualizationEMAM/models/PacMan/index.html#de.rwth.pacman.pacManWrapper.controller.coneSearch:3
* ![image](/uploads/600c08e30d3a7065bd9c200154adec51/image.png)
* **this is the worst, here are even the connectors between components missing**
2. https://embeddedmontiarc.github.io/VisualizationEMAM/models/PacMan/index.html#de.rwth.pacman.pacManWrapper.controller.coneSearch:2
* ![image](/uploads/fc980e12bc364a69f09d4f4658fb9cfe/image.png)
* here the component width is too small
3. https://embeddedmontiarc.github.io/VisualizationEMAM/models/PacMan/index.html#de.rwth.pacman.pacManWrapper.controller.coneSearch:0
* ![image](/uploads/f49d7a040e55d843e7910f69bc3d828f/image.png)
* here components are much too fat
The first (1.) issue is not only ugly is also incorrect, as the connections are missign.
@toezen1 can you please fix it.
@toezen1 and why are there not tests which fail? I mean there is no single connector drawn.
@jean\-marc.ronck can you please write @toezen1 what version of visualization you used; and when @toezen1 fixed this bug; can please update the version and the models at https://embeddedmontiarc.github.io/VisualizationEMAM.
@toezen1 as we published this models for our paper (http://www.se-rwth.de/materials/embeddedmontiarc/), and the reviewing process starts at Wednesday (July 18th), can you please fix the issues 1. and 2. immediately. And when you fixed it. please write an E-Mail to @jean\-marc.ronck so that he updates the models at github-pages; I will then update the models for our paper homepage.
(You do not need to fix 3. as this is only very ugly but not wrong)toezen1toezen1https://git.rwth-aachen.de/monticore/EmbeddedMontiArc/generators/EMAM2Cpp/-/issues/15Sum command for armadillo not supported2018-07-13T12:27:03+02:00Christoph RichterSum command for armadillo not supportedUntil now EMAM only supports `sum(X)` which creates `accu(X)` in armadillo.
Nevertheless armadillo and MATLAB also support `sum(X,1)` or `sum(X,2)` to only sum rows or columns. Would be nice if EMAM also supports this.Until now EMAM only supports `sum(X)` which creates `accu(X)` in armadillo.
Nevertheless armadillo and MATLAB also support `sum(X,1)` or `sum(X,2)` to only sum rows or columns. Would be nice if EMAM also supports this.Christoph RichterChristoph Richterhttps://git.rwth-aachen.de/monticore/EmbeddedMontiArc/generators/EMAM2RosCpp/-/issues/1Add error msg for non recoverable RosConnection error2018-07-06T15:03:02+02:00Alexander David HellwigAdd error msg for non recoverable RosConnection errorMissing Topic name without connector leads to uncaught exceptionMissing Topic name without connector leads to uncaught exceptionAlexander David HellwigAlexander David Hellwig2018-07-07https://git.rwth-aachen.de/monticore/EmbeddedMontiArc/generators/MathPrettyPrinter/-/issues/8Hadamard-Produkt2018-07-04T19:36:45+02:00Evgeny KusmenkoHadamard-ProduktAnalog zu
https://embeddedmontiarc.github.io/MathPrettyPrinter/montimath/vanilla/PowerWise/index.html
sollte auch das hadamard-Produkt aussehen. Aus .* sollte also ein Kreis entstehen, zB hier:
https://embeddedmontiarc.github.io/MathPre...Analog zu
https://embeddedmontiarc.github.io/MathPrettyPrinter/montimath/vanilla/PowerWise/index.html
sollte auch das hadamard-Produkt aussehen. Aus .* sollte also ein Kreis entstehen, zB hier:
https://embeddedmontiarc.github.io/MathPrettyPrinter/montimath/vanilla/TimeWise/index.html
Siehe:
https://en.wikipedia.org/wiki/Hadamard_product_(matrices)
Bei .+ bzw .- soll zudem der Punkt entfernt werden, Addition / Subtraktion immer elementweise sind.
Bei ./ verhält es sich folgendermaßen: A./B = A°B^{°(-1)} , wobei ^{} "hochgestellt" symbolisieren soll. Siehe:
https://math.stackexchange.com/questions/1321879/is-there-a-formula-for-the-inverse-of-hadamard-productArmin MokhtarianArmin Mokhtarianhttps://git.rwth-aachen.de/monticore/EmbeddedMontiArc/generators/MathPrettyPrinter/-/issues/4Unnötige Klammern bei Zeilen2018-07-02T11:49:34+02:00vonWencksternUnnötige Klammern bei ZeilenIn https://git.rwth-aachen.de/monticore/EmbeddedMontiArc/generators/MathPrettyPrinter/blob/master/src/test/resources/models/montimath/vanilla/SolEqu.m#L5
sieht die Ausgabe wie folgt aus: (Bug #2 behoben)
```
Q^{3,1} A = [3 6 2; 1 2 8; 7...In https://git.rwth-aachen.de/monticore/EmbeddedMontiArc/generators/MathPrettyPrinter/blob/master/src/test/resources/models/montimath/vanilla/SolEqu.m#L5
sieht die Ausgabe wie folgt aus: (Bug #2 behoben)
```
Q^{3,1} A = [3 6 2; 1 2 8; 7 9 4] \ [2;3;4];
```
Aber die Ausgabe sieht wie folgt aus:
![image](/uploads/6705613da019557155aba8373052f145/image.png)
Gerade bei Vektoren wie `[(2); (3); (4)]`, sieht die unnötige Klammerung pro Zeile hässlich aus; viel besser ist ` [2; 3; 4]`.
@jean\-marc.ronck, @armin.mokhtarian, @bram.kohlen: Bitte beheben.Jean-Marc RonckJean-Marc Ronckhttps://git.rwth-aachen.de/monticore/EmbeddedMontiArc/generators/MathPrettyPrinter/-/issues/5bei mathematischen Brüchen Platz lassen zwischen Zeilen2018-07-02T11:45:22+02:00vonWencksternbei mathematischen Brüchen Platz lassen zwischen ZeilenIch denke folgendes lässt sich besser lesen:
![image](/uploads/547aea10161b01aef183001fb74b34c0/image.png)
Besonders, wenn man folgendes sieht:
![image](/uploads/a478586aa6bc68754c710e4f14c1bb2b/image.png)
@kusmenko: Was meinst du? Sol...Ich denke folgendes lässt sich besser lesen:
![image](/uploads/547aea10161b01aef183001fb74b34c0/image.png)
Besonders, wenn man folgendes sieht:
![image](/uploads/a478586aa6bc68754c710e4f14c1bb2b/image.png)
@kusmenko: Was meinst du? Sollte man bei Brüchen eine leer Zeile mit kleinerer Schriftgröße extra einfügen?Jean-Marc RonckJean-Marc Ronckhttps://git.rwth-aachen.de/monticore/EmbeddedMontiArc/generators/MathPrettyPrinter/-/issues/7Test mit transponierten Operator auf gh-pages hochladen2018-07-02T11:44:13+02:00vonWencksternTest mit transponierten Operator auf gh-pages hochladenBitte einen Test für `X = A'` hochladen.
Und `A'` im LaTeX-Modus bitte wiefolgt darstellen:
![image](/uploads/3fb44603932abe7e296bcf6513e90161/image.png)Bitte einen Test für `X = A'` hochladen.
Und `A'` im LaTeX-Modus bitte wiefolgt darstellen:
![image](/uploads/3fb44603932abe7e296bcf6513e90161/image.png)Jean-Marc RonckJean-Marc Ronckhttps://git.rwth-aachen.de/monticore/EmbeddedMontiArc/generators/MathPrettyPrinter/-/issues/2[Bug in Grammatik] Gleichungssystem lösen; ist nur einfacher Backslash und ni...2018-06-28T21:46:31+02:00vonWenckstern[Bug in Grammatik] Gleichungssystem lösen; ist nur einfacher Backslash und nicht doppelterhttps://git.rwth-aachen.de/monticore/EmbeddedMontiArc/generators/MathPrettyPrinter/blob/tex/src/test/resources/models/montimath/vanilla/SolEqu.m#L5
Da ist ein Fehler drin.
Die Zeile
```
Q^{3,1} A = [3 6 2; 1 2 8; 7 9 4] \\ [2;3;4];
```
...https://git.rwth-aachen.de/monticore/EmbeddedMontiArc/generators/MathPrettyPrinter/blob/tex/src/test/resources/models/montimath/vanilla/SolEqu.m#L5
Da ist ein Fehler drin.
Die Zeile
```
Q^{3,1} A = [3 6 2; 1 2 8; 7 9 4] \\ [2;3;4];
```
müsste
```
Q^{3,1} A = [3 6 2; 1 2 8; 7 9 4] \ [2;3;4];
```
lauten; siehe
https://de.mathworks.com/help/matlab/ref/mldivide.html
@jean\-marc.ronck, @bram.kohlen, @armin.mokhtarian: Könnt ihr die Grammatik bitte so anpassen, dass sich das Beispiel Regelkonform parsen lassen?Jean-Marc RonckJean-Marc Ronckhttps://git.rwth-aachen.de/monticore/EmbeddedMontiArc/generators/CNNArch2Caffe2/-/issues/1CI einrichten2018-06-13T21:44:32+02:00Evgeny KusmenkoCI einrichtenCarlos Alfredo Yeverino RodriguezCarlos Alfredo Yeverino Rodriguez2018-06-13