bob issueshttps://gitlab.idiap.ch/groups/bob/-/issues2016-08-04T09:27:53Zhttps://gitlab.idiap.ch/bob/bob/-/issues/34inconsistent function naming convention2016-08-04T09:27:53ZAndré Anjosinconsistent function naming convention*Created by: siebenkopf*
Currently, there are different ways, python functions are called. One type uses CamelCase, whereas other types use under_score_names. Since the default python ways seems to be under_scores, we should rename the ...*Created by: siebenkopf*
Currently, there are different ways, python functions are called. One type uses CamelCase, whereas other types use under_score_names. Since the default python ways seems to be under_scores, we should rename the camelCase functions in the API.v1.0https://gitlab.idiap.ch/bob/bob/-/issues/33New databases "casia_fasd" and "nuaa" do not work with dbmanage.py2016-08-04T09:27:51ZAndré AnjosNew databases "casia_fasd" and "nuaa" do not work with dbmanage.py*Created by: anjos*
Please integrate those ASAP. Check other implemented drivers like the "iris" database or the "replay" database.*Created by: anjos*
Please integrate those ASAP. Check other implemented drivers like the "iris" database or the "replay" database.v1.0https://gitlab.idiap.ch/bob/bob/-/issues/32Tests failing on Ubuntu 12.04 LTS 32 bits2016-08-04T09:27:49ZAndré AnjosTests failing on Ubuntu 12.04 LTS 32 bits*Created by: laurentes*
The three following tests are failing on Ubuntu 12.04 LTS 32 bits:
- cxx-ip-facenorm
- python-trainer-rprop-main
- python-face_verification-bancasmall-main
In all three cases, this seems to be caused by numer...*Created by: laurentes*
The three following tests are failing on Ubuntu 12.04 LTS 32 bits:
- cxx-ip-facenorm
- python-trainer-rprop-main
- python-face_verification-bancasmall-main
In all three cases, this seems to be caused by numerical issues. In particular, "python-face_verification-bancasmall-main" makes use of facenorm.v1.0https://gitlab.idiap.ch/bob/bob/-/issues/31Add a new command to check if the data of a database are in place2016-08-04T09:27:47ZAndré AnjosAdd a new command to check if the data of a database are in place*Created by: laurentes*
It seems that such a command would be useful to check that a folder containing a database has the expectated file organization.*Created by: laurentes*
It seems that such a command would be useful to check that a folder containing a database has the expectated file organization.v1.0https://gitlab.idiap.ch/bob/bob/-/issues/30Order of the eigenvalues with FisherLDA2016-08-04T09:27:43ZAndré AnjosOrder of the eigenvalues with FisherLDA*Created by: laurentes*
Following the introduction of the doctest, it seems that the behaviour of our FisherLDA implementation is currently platform dependent: the order of the returned eigenvalues and eigenvectors differs from one plat...*Created by: laurentes*
Following the introduction of the doctest, it seems that the behaviour of our FisherLDA implementation is currently platform dependent: the order of the returned eigenvalues and eigenvectors differs from one platform to another one. This might be caused by a different behaviour of the LAPACK implementations. We need to fix the problem ASAP.v1.0https://gitlab.idiap.ch/bob/bob/-/issues/29bob.ip.Rotate needs a review2016-08-04T09:27:41ZAndré Anjosbob.ip.Rotate needs a review*Created by: siebenkopf*
The bob.ip.Rotate class needs a mayor revision. First, it expects the rotation angle in degrees (which is rather uncommon), but this is documented nowhere. The resulting rotated image is assumed to be of type fl...*Created by: siebenkopf*
The bob.ip.Rotate class needs a mayor revision. First, it expects the rotation angle in degrees (which is rather uncommon), but this is documented nowhere. The resulting rotated image is assumed to be of type float, even when the input image is uint8. I am not sure if this behavior is intended.
Furthermore, its function bob.ip.Rotate.getOutputShape(image,angle) does not use the angle member variable, but requires the angle to be specified again. There are two solutions for this last problem. Either the class is removed and only the two functions bob.ip.rotate(image,image,angle) and bob.ip.getRotatedShape(image,angle) are preserved, or the getRotatedShape(image) function uses the angle member variable. I vote for the first solution since this would mimic the bob.ip.scale(image,image,scale) function.
v1.0https://gitlab.idiap.ch/bob/bob/-/issues/28Temporary files not being erased2016-08-04T09:27:39ZAndré AnjosTemporary files not being erased*Created by: anjos*
After running bob's test suite, lots of files seem not to be erased from the temporary directory. We should verify this as soon as possible.*Created by: anjos*
After running bob's test suite, lots of files seem not to be erased from the temporary directory. We should verify this as soon as possible.v1.0https://gitlab.idiap.ch/bob/bob/-/issues/27Usage of abbreviations in namespaces2016-08-04T09:27:37ZAndré AnjosUsage of abbreviations in namespaces*Created by: siebenkopf*
When I try to read some C++ code of Bob, I often step over an abbreviated namespace (like tp for bob::python). Also in the python code, sometimes packages are abbreviated or, even worse, "from xxx import yyy as ...*Created by: siebenkopf*
When I try to read some C++ code of Bob, I often step over an abbreviated namespace (like tp for bob::python). Also in the python code, sometimes packages are abbreviated or, even worse, "from xxx import yyy as zzz" is used. This makes the code really hard to read since I always have to scroll up to the definition of abbreviation to understand, what is actually called.
I definitely would go for removing all these abbreviations and using the fully qualified names. I know that writing code this way may take a little longer, but I think it is worth the time. And most of the editors provide automatic source completion, that speeds up the typing.
What do you think about this?v1.2https://gitlab.idiap.ch/bob/bob/-/issues/26Bob installation documentation needs to be rewritten2016-08-04T09:27:34ZAndré AnjosBob installation documentation needs to be rewritten*Created by: siebenkopf*
In the "Using bob" section of the "Installation" section of the documentation still the bin/shell.py is called. This should be replaced by the current way to call python, namely bin/python. Also, the mentioned b...*Created by: siebenkopf*
In the "Using bob" section of the "Installation" section of the documentation still the bin/shell.py is called. This should be replaced by the current way to call python, namely bin/python. Also, the mentioned bin/bob_config.py is no longer existent.
Additionally, it might be a good idea to include the BobMaker here (maybe marked to be used only Idiap-internally, if this is the case).v1.0https://gitlab.idiap.ch/bob/bob/-/issues/25cblas, clapack and LAPACK2016-08-04T09:27:32ZAndré Anjoscblas, clapack and LAPACK*Created by: laurentes*
In the early years, we've got rid of the usage of C functions from cblas and clapack, calling the Fortran functions from LAPACK directly. This is not reflected in the external/cblasConfig.cmake, which still requi...*Created by: laurentes*
In the early years, we've got rid of the usage of C functions from cblas and clapack, calling the Fortran functions from LAPACK directly. This is not reflected in the external/cblasConfig.cmake, which still requires cblas.h and clapack.h to exist.v1.0https://gitlab.idiap.ch/bob/bob/-/issues/24euklidean_distance should be renamed2016-08-04T09:27:30ZAndré Anjoseuklidean_distance should be renamed*Created by: laurentes*
I've nothing against the beautiful german language, but it would be better to stick to English for the Bob project. Furthermore, the function euklidean_distance() in python/math/lib/__init__.py should be renamed ...*Created by: laurentes*
I've nothing against the beautiful german language, but it would be better to stick to English for the Bob project. Furthermore, the function euklidean_distance() in python/math/lib/__init__.py should be renamed to something like euclidean_distance().v1.0https://gitlab.idiap.ch/bob/bob/-/issues/23ffmpeg 0.8 on Ubuntu 12.04 deprecates some symbols2016-08-04T09:27:28ZAndré Anjosffmpeg 0.8 on Ubuntu 12.04 deprecates some symbols*Created by: anjos*
Our bindings to ffmpeg need a revamp. The VideoReader compilation reports:
```
/home/andre/work/bob/cxx/io/src/VideoReader.cc: In member function ‘void bob::io::VideoReader::open()’:
/home/andre/work/bob/cxx/io/...*Created by: anjos*
Our bindings to ffmpeg need a revamp. The VideoReader compilation reports:
```
/home/andre/work/bob/cxx/io/src/VideoReader.cc: In member function ‘void bob::io::VideoReader::open()’:
/home/andre/work/bob/cxx/io/src/VideoReader.cc:128:7: warning: ‘int av_find_stream_info(AVFormatContext*)’ is deprecated (declared at /usr/include/libavformat/avformat.h:1412) [-Wdeprecated-declarations]
/home/andre/work/bob/cxx/io/src/VideoReader.cc:128:38: warning: ‘int av_find_stream_info(AVFormatContext*)’ is deprecated (declared at /usr/include/libavformat/avformat.h:1412) [-Wdeprecated-declarations]
/home/andre/work/bob/cxx/io/src/VideoReader.cc:129:5: warning: ‘void av_close_input_file(AVFormatContext*)’ is deprecated (declared at /usr/include/libavformat/avformat.h:1580) [-Wdeprecated-declarations]
/home/andre/work/bob/cxx/io/src/VideoReader.cc:129:36: warning: ‘void av_close_input_file(AVFormatContext*)’ is deprecated (declared at /usr/include/libavformat/avformat.h:1580) [-Wdeprecated-declarations]
/home/andre/work/bob/cxx/io/src/VideoReader.cc:147:5: warning: ‘void av_close_input_file(AVFormatContext*)’ is deprecated (declared at /usr/include/libavformat/avformat.h:1580) [-Wdeprecated-declarations]
/home/andre/work/bob/cxx/io/src/VideoReader.cc:147:36: warning: ‘void av_close_input_file(AVFormatContext*)’ is deprecated (declared at /usr/include/libavformat/avformat.h:1580) [-Wdeprecated-declarations]
/home/andre/work/bob/cxx/io/src/VideoReader.cc:163:5: warning: ‘void av_close_input_file(AVFormatContext*)’ is deprecated (declared at /usr/include/libavformat/avformat.h:1580) [-Wdeprecated-declarations]
/home/andre/work/bob/cxx/io/src/VideoReader.cc:163:36: warning: ‘void av_close_input_file(AVFormatContext*)’ is deprecated (declared at /usr/include/libavformat/avformat.h:1580) [-Wdeprecated-declarations]
/home/andre/work/bob/cxx/io/src/VideoReader.cc:168:7: warning: ‘int avcodec_open(AVCodecContext*, AVCodec*)’ is deprecated (declared at /usr/include/libavcodec/avcodec.h:4034) [-Wdeprecated-declarations]
/home/andre/work/bob/cxx/io/src/VideoReader.cc:168:37: warning: ‘int avcodec_open(AVCodecContext*, AVCodec*)’ is deprecated (declared at /usr/include/libavcodec/avcodec.h:4034) [-Wdeprecated-declarations]
/home/andre/work/bob/cxx/io/src/VideoReader.cc:169:5: warning: ‘void av_close_input_file(AVFormatContext*)’ is deprecated (declared at /usr/include/libavformat/avformat.h:1580) [-Wdeprecated-declarations]
/home/andre/work/bob/cxx/io/src/VideoReader.cc:169:36: warning: ‘void av_close_input_file(AVFormatContext*)’ is deprecated (declared at /usr/include/libavformat/avformat.h:1580) [-Wdeprecated-declarations]
/home/andre/work/bob/cxx/io/src/VideoReader.cc:228:3: warning: ‘void av_close_input_file(AVFormatContext*)’ is deprecated (declared at /usr/include/libavformat/avformat.h:1580) [-Wdeprecated-declarations]
/home/andre/work/bob/cxx/io/src/VideoReader.cc:228:34: warning: ‘void av_close_input_file(AVFormatContext*)’ is deprecated (declared at /usr/include/libavformat/avformat.h:1580) [-Wdeprecated-declarations]
/home/andre/work/bob/cxx/io/src/VideoReader.cc: In member function ‘void bob::io::VideoReader::const_iterator::init()’:
/home/andre/work/bob/cxx/io/src/VideoReader.cc:346:7: warning: ‘int av_find_stream_info(AVFormatContext*)’ is deprecated (declared at /usr/include/libavformat/avformat.h:1412) [-Wdeprecated-declarations]
/home/andre/work/bob/cxx/io/src/VideoReader.cc:346:40: warning: ‘int av_find_stream_info(AVFormatContext*)’ is deprecated (declared at /usr/include/libavformat/avformat.h:1412) [-Wdeprecated-declarations]
/home/andre/work/bob/cxx/io/src/VideoReader.cc:377:7: warning: ‘int avcodec_open(AVCodecContext*, AVCodec*)’ is deprecated (declared at /usr/include/libavcodec/avcodec.h:4034) [-Wdeprecated-declarations]
/home/andre/work/bob/cxx/io/src/VideoReader.cc:377:41: warning: ‘int avcodec_open(AVCodecContext*, AVCodec*)’ is deprecated (declared at /usr/include/libavcodec/avcodec.h:4034) [-Wdeprecated-declarations]
/home/andre/work/bob/cxx/io/src/VideoReader.cc: In member function ‘void bob::io::VideoReader::const_iterator::reset()’:
/home/andre/work/bob/cxx/io/src/VideoReader.cc:463:5: warning: ‘void av_close_input_file(AVFormatContext*)’ is deprecated (declared at /usr/include/libavformat/avformat.h:1580) [-Wdeprecated-declarations]
/home/andre/work/bob/cxx/io/src/VideoReader.cc:463:38: warning: ‘void av_close_input_file(AVFormatContext*)’ is deprecated (declared at /usr/include/libavformat/avformat.h:1580) [-Wdeprecated-declarations]
```
Where as the VideoWriter.cc reports:
```
/home/andre/work/bob/cxx/io/src/VideoWriter.cc: In member function ‘AVStream* bob::io::VideoWriter::add_video_stream()’:
/home/andre/work/bob/cxx/io/src/VideoWriter.cc:229:8: warning: ‘AVStream* av_new_stream(AVFormatContext*, int)’ is deprecated (declared at /usr/include/libavformat/avformat.h:1604) [-Wdeprecated-declarations]
/home/andre/work/bob/cxx/io/src/VideoWriter.cc:229:38: warning: ‘AVStream* av_new_stream(AVFormatContext*, int)’ is deprecated (declared at /usr/include/libavformat/avformat.h:1604) [-Wdeprecated-declarations]
/home/andre/work/bob/cxx/io/src/VideoWriter.cc: In member function ‘void bob::io::VideoWriter::open_video()’:
/home/andre/work/bob/cxx/io/src/VideoWriter.cc:300:7: warning: ‘int avcodec_open(AVCodecContext*, AVCodec*)’ is deprecated (declared at /usr/include/libavcodec/avcodec.h:4034) [-Wdeprecated-declarations]
/home/andre/work/bob/cxx/io/src/VideoWriter.cc:300:28: warning: ‘int avcodec_open(AVCodecContext*, AVCodec*)’ is deprecated (declared at /usr/include/libavcodec/avcodec.h:4034) [-Wdeprecated-declarations]
```v1.0https://gitlab.idiap.ch/bob/bob/-/issues/22debugging problem2016-08-04T09:27:26ZAndré Anjosdebugging problem*Created by: ivana7c*
Debugging using gdb requires that the PYTHONPATH is set before every call of gdb. Script that will do this automatically should be provided.*Created by: ivana7c*
Debugging using gdb requires that the PYTHONPATH is set before every call of gdb. Script that will do this automatically should be provided.https://gitlab.idiap.ch/bob/bob/-/issues/21Can't read .ppm image with stock Ubuntu 10.10 and 11.042016-08-04T09:27:23ZAndré AnjosCan't read .ppm image with stock Ubuntu 10.10 and 11.04*Created by: laurentes*
With the stock installation of Ubuntu 10.10 and 11.04, loading ppm images inside bob is failing. This is caused by a change in the behaviour of the ImageMagick library, which interprets ppm images as color images...*Created by: laurentes*
With the stock installation of Ubuntu 10.10 and 11.04, loading ppm images inside bob is failing. This is caused by a change in the behaviour of the ImageMagick library, which interprets ppm images as color images, instead of grayscale. This might be one of the possible explanation of why canonical did a downgrade to ImageMagick 6.6.0.4-3 in Ubuntu 11.10.
Stock imagemagick versions of several Ubuntu releases:
- Ubuntu 10.04LTS -> ImageMagick 6.5.7.8-1
- Ubuntu 10.10 -> ImageMagick 6.6.2.6-1
- Ubuntu 11.04 -> ImageMagick 6.6.2.6-1
- Ubuntu 11.10 -> ImageMagick 6.6.0.4-3
This currently makes the following tests failing, as they use .ppm image files:
- cxx-io-image_codec
- cxx-ip-rotate
- python-ip-histo-mainhttps://gitlab.idiap.ch/bob/bob/-/issues/20DCT 1D c++ tests failing in release mode on Ubuntu 10.10+2016-08-04T09:27:21ZAndré AnjosDCT 1D c++ tests failing in release mode on Ubuntu 10.10+*Created by: laurentes*
The current DCT1D C++ test is failing in release mode on Ubuntu 10.10, as well as on more recent Ubuntu versions. In contrast, everything is working fine in debug mode. The DCT1DNaive implementation seems to be t...*Created by: laurentes*
The current DCT1D C++ test is failing in release mode on Ubuntu 10.10, as well as on more recent Ubuntu versions. In contrast, everything is working fine in debug mode. The DCT1DNaive implementation seems to be the cause of the problem.
https://gitlab.idiap.ch/bob/bob/-/issues/18ZT Norm returns NaN when the standard deviation of the scores is equal to 02016-08-04T09:27:19ZAndré AnjosZT Norm returns NaN when the standard deviation of the scores is equal to 0*Created by: laurentes*
The current ZT-Norm implementation doesn't check whether the standard deviation of the input scores (B, C, and D matrices of scores) is 0 or not. This leads to NaN normalized scores. We need to introduce some fur...*Created by: laurentes*
The current ZT-Norm implementation doesn't check whether the standard deviation of the input scores (B, C, and D matrices of scores) is 0 or not. This leads to NaN normalized scores. We need to introduce some further checks to deal with the problemv1.0https://gitlab.idiap.ch/bob/bob/-/issues/17Database documentation should include expected filesystem mapping2016-08-04T09:27:17ZAndré AnjosDatabase documentation should include expected filesystem mapping*Created by: anjos*
In order go make our current DB API and practices useful, we should precisely document what is the expected layout expected by each database in Bob.*Created by: anjos*
In order go make our current DB API and practices useful, we should precisely document what is the expected layout expected by each database in Bob.v1.0https://gitlab.idiap.ch/bob/bob/-/issues/16Loading an hdf5 Arrayset requires write access2012-02-29T14:05:48ZAndré AnjosLoading an hdf5 Arrayset requires write access*Created by: laurentes*
Loading an .hdf5 file as an Arrayset with read permission but without write permission generates the following runtime error:
<pre><code>
bob.io.Arrayset('/idiap/temp/pmotlic/TEMP.2/MOBIO/DATA/f007/f007_04_r06_...*Created by: laurentes*
Loading an .hdf5 file as an Arrayset with read permission but without write permission generates the following runtime error:
<pre><code>
bob.io.Arrayset('/idiap/temp/pmotlic/TEMP.2/MOBIO/DATA/f007/f007_04_r06_i0_0.hdf5')
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
RuntimeError: Call to HDF5 C-function 'H5Fopen' returned '-1'. HDF5 error statck follows:
</code></pre>
whereas it works fine for Array or using bob.io.load:
<pre><code>
bob.io.Array('/idiap/temp/pmotlic/TEMP.2/MOBIO/DATA/f007/f007_04_r06_i0_0.hdf5')
Array float64@(351, 60)
bob.io.load('/idiap/temp/pmotlic/TEMP.2/MOBIO/DATA/f007/f007_04_r06_i0_0.hdf5')
array([[ -1.23146095e+01, -7.83866215e+00, 1.74252892e+01, ...,
-2.80604903e-02, -3.33320588e-01, 1.99772105e-01],
[ -1.34626703e+01, -8.39167404e+00, 1.67254601e+01, ...,
1.30144298e-01, -1.32815793e-01, 2.55901814e-01],
[ -7.23934698e+00, -3.61744595e+00, 1.45476503e+01, ...,
9.95570198e-02, 2.48753905e-01, 1.26341403e-01],
...,
[ -1.02341805e+01, -3.34847593e+00, 1.06275997e+01, ...,
1.29639900e+00, 3.63285094e-01, 1.51863798e-01],
[ -6.78214598e+00, 2.78683591e+00, 1.41695900e+01, ...,
7.74318576e-01, 8.26705305e-04, 6.93367869e-02],
[ -7.07084513e+00, 1.43504307e-01, 1.79709702e+01, ...,
-2.10530292e-02, -2.02032804e-01, -1.68551896e-02]])
</code></pre>
If I copy the file to get write permission, then the loading can be done correctly:
<pre><code>
bob.io.Arrayset('/idiap/home/lelshafey/f007_04_r06_i0_0.hdf5')
Arrayset[351] float64@(60,)
</code></pre>
I guess there is a write flag which is set incorrectly in the hdf5 support part.v1.0https://gitlab.idiap.ch/bob/bob/-/issues/15Blitz limitation with the debugging mode2016-08-04T09:27:13ZAndré AnjosBlitz limitation with the debugging mode*Created by: laurentes*
Blitz has a current limitation when combining reduction and complex expressions, AND enabling the debugging mode using the -DBZ_DEBUG flag. This problem is described in this blitz <a href="http://sourceforge.net/...*Created by: laurentes*
Blitz has a current limitation when combining reduction and complex expressions, AND enabling the debugging mode using the -DBZ_DEBUG flag. This problem is described in this blitz <a href="http://sourceforge.net/tracker/?func=detail&aid=2058441&group_id=63961&atid=505791">ticket</a>.
In particular, this means that such a piece code will throw an assert when the DBZ_DEBUG flag is set, but will work fine otherwise:
<pre><code>
blitz::firstIndex bi;
blitz::secondIndex bj;
blitz::thirdIndex bk;
blitz::Array<double,3> A(2,2,2);
A = 2.;
blitz::Array<double,2> B(2,2);
B = 7.;
blitz::sum(A(bi,bj,bk)*B(bi,bj),bk); // throws an assert
// linux-x86_64/include/blitz/array/expr.h:492 Two array operands have different orders: for rank 0, the orders are 2 and 1
// linux-x86_64/include/blitz/array/expr.h:493: static int blitz::bounds::compute_ordering(int, int, int): Assertion `0' failed.
</code></pre>
More precisely, this happens when the expression inside the reduction is built on arrays of different dimensions. The current workaround to make everything work with the DBZ_DEBUG flag enabled is to declare an intermediate array:
<pre><code>
blitz::Array<double,3> C(A(bi,bj,bk)*B(bi,bj));
blitz::sum(C,bk); // throw an assert
</code></pre>
Please be aware that this might generate some overhead.https://gitlab.idiap.ch/bob/bob/-/issues/14Clicking "Issues" gives 404 error if not logged in to github2016-08-04T09:27:11ZAndré AnjosClicking "Issues" gives 404 error if not logged in to github*Created by: Waldo000000*
From the Bob homepage (http://idiap.github.com/bob/), clicking on "Issues" (http://github.com/idiap/bob/issues) or "GitHub" (http://github.com/idiap/bob) both result in a 404 error. The user has to first go to ...*Created by: Waldo000000*
From the Bob homepage (http://idiap.github.com/bob/), clicking on "Issues" (http://github.com/idiap/bob/issues) or "GitHub" (http://github.com/idiap/bob) both result in a 404 error. The user has to first go to github.com and log in, then return to the Bob homepage.
Suggested fixes:
1) The user should be able to log in directly from the Bob homepage, or
2) Rather than 404, the user should be issued with a more informative error & instructionsv1.0