bob issueshttps://gitlab.idiap.ch/groups/bob/-/issues2019-01-25T14:54:02Zhttps://gitlab.idiap.ch/bob/bob.devtools/-/issues/9Build command fails, complaining about zstd2019-01-25T14:54:02ZJaden DIEFENBAUGHBuild command fails, complaining about zstdRunning `bdt build --help` gets:
```
read filter "zstd" is not supported
write filter "zstd" is not supported
Usage: bdt build [OPTIONS]
Warning: entry point could not be loaded. Contact its author for help.
Traceback (most recent ...Running `bdt build --help` gets:
```
read filter "zstd" is not supported
write filter "zstd" is not supported
Usage: bdt build [OPTIONS]
Warning: entry point could not be loaded. Contact its author for help.
Traceback (most recent call last):
File "/home/jae/miniconda3/envs/bdt/lib/python3.6/site-packages/click_plugins/core.py", line 37, in decorator
group.add_command(entry_point.load())
File "/home/jae/miniconda3/envs/bdt/lib/python3.6/site-packages/pkg_resources/__init__.py", line 2318, in load
return self.resolve()
File "/home/jae/miniconda3/envs/bdt/lib/python3.6/site-packages/pkg_resources/__init__.py", line 2324, in resolve
module = __import__(self.module_name, fromlist=['__name__'], level=0)
File "/home/jae/miniconda3/envs/bdt/lib/python3.6/site-packages/bob/devtools/scripts/build.py", line 75, in <module>
@bdt.raise_on_error
File "/home/jae/miniconda3/envs/bdt/lib/python3.6/site-packages/click/decorators.py", line 170, in decorator
_param_memo(f, OptionClass(param_decls, **attrs))
File "/home/jae/miniconda3/envs/bdt/lib/python3.6/site-packages/click/core.py", line 1460, in __init__
Parameter.__init__(self, param_decls, type=type, **attrs)
TypeError: __init__() got an unexpected keyword argument 'hidden'
Options:
-h, -?, --help Show this message and exit.
```
I get those first two lines about read/write filters with every `bdt` command I try, even just `bdt --help`.
Here's the conda environment (updated as of Jan. 23rd):
```
# Name Version Build Channel
alabaster 0.7.11 py36_0 defaults
asn1crypto 0.24.0 py36_0 defaults
babel 2.6.0 py36_0 defaults
beautifulsoup4 4.7.1 py36_1 defaults
bob 5.0.0 py36hc3b47e9_0 https://www.idiap.ch/software/bob/conda
bob.devtools 0.1.2 py36h6b6bea4_0 https://www.idiap.ch/software/bob/conda
bzip2 1.0.6 h14c3975_5 defaults
ca-certificates 2018.03.07 0 defaults
certifi 2018.4.16 py36_0 defaults
cffi 1.11.5 py36h9745a5d_0 defaults
chardet 3.0.4 py36h0f667ec_1 defaults
click 6.7 py36h5253387_0 defaults
click-plugins 1.0.3 py36_0 defaults
conda 4.5.12 py36_0 defaults
conda-build 3.17.7 py36_0 defaults
conda-env 2.6.0 1 defaults
cryptography 2.2.2 py36h14c3975_0 defaults
docutils 0.14 py36hb0f60f5_0 defaults
filelock 3.0.10 py36_0 defaults
gitdb2 2.0.5 py36_0 defaults
gitpython 2.1.11 py36_0 defaults
glob2 0.6 py36_1 defaults
icu 58.2 h9c2bf20_1 defaults
idna 2.7 py36_0 defaults
imagesize 1.0.0 py36_0 defaults
jinja2 2.10 py36ha16c418_0 defaults
libarchive 3.3.2 hb43526a_6 defaults
libedit 3.1.20170329 h6b74fdf_2 defaults
libffi 3.2.1 hd88cf55_4 defaults
libgcc-ng 7.2.0 hdf63c60_3 defaults
liblief 0.9.0 h1532aa0_0 defaults
libstdcxx-ng 7.2.0 hdf63c60_3 defaults
libxml2 2.9.8 h26e45fe_1 defaults
libxslt 1.1.32 h1312cb7_0 defaults
lxml 4.2.4 py36hf71bdeb_0 defaults
lz4-c 1.8.1.2 h14c3975_0 defaults
lzo 2.10 h49e0be7_2 defaults
markupsafe 1.0 py36hd9260cd_1 defaults
ncurses 6.1 hf484d3e_0 defaults
openssl 1.0.2o h20670df_0 defaults
patchelf 0.9 hf484d3e_2 defaults
pip 18.1 py36_0 defaults
pkginfo 1.5.0.1 py36_0 defaults
psutil 5.4.6 py36h14c3975_0 defaults
py-lief 0.9.0 py36h1532aa0_0 defaults
pycosat 0.6.3 py36h14c3975_0 defaults
pycparser 2.18 py36hf9f622e_1 defaults
pygments 2.2.0 py36h0d3125c_0 defaults
pyopenssl 18.0.0 py36_0 defaults
pysocks 1.6.8 py36_0 defaults
python 3.6.6 hc3d631a_0 defaults
python-dateutil 2.7.3 py36_0 defaults
python-gitlab 1.7.0 py36_0 https://www.idiap.ch/software/bob/conda
python-libarchive-c 2.8 py36_6 defaults
pytz 2018.5 py36_0 defaults
pyyaml 3.12 py36hafb9ca4_1 defaults
readline 7.0 ha6073c6_4 defaults
requests 2.19.1 py36_0 defaults
requests-toolbelt 0.8.0 py36_0 defaults
ruamel_yaml 0.15.46 py36h14c3975_0 defaults
setuptools 39.2.0 py36_0 defaults
six 1.11.0 py36h372c433_1 defaults
smmap2 2.0.5 py36_0 defaults
snowballstemmer 1.2.1 py36h6febd40_0 defaults
soupsieve 1.7.1 py36_0 defaults
sphinx 1.6.3 py36he5f0bdb_0 defaults
sphinxcontrib 1.0 py36h6d0f590_1 defaults
sphinxcontrib-websupport 1.1.0 py36_1 defaults
sqlite 3.24.0 h84994c4_0 defaults
tk 8.6.7 hc745277_3 defaults
tqdm 4.29.1 py_0 defaults
twine 1.11.0 py36_0 defaults
typing 3.6.4 py36_0 defaults
urllib3 1.23 py36_0 defaults
wheel 0.32.3 py36_0 defaults
xz 5.2.4 h14c3975_4 defaults
yaml 0.1.7 had09818_2 defaults
zlib 1.2.11 ha838bed_2 defaults
zstd 1.3.3 h84994c4_0 defaults
```André AnjosAndré Anjoshttps://gitlab.idiap.ch/bob/bob.devtools/-/issues/8Too many environment variables required2019-01-20T21:57:34ZAndré AnjosToo many environment variables requiredThe current build system depends on too many environment variables which are not explicitly passed to the build functions, but inferred from the current environment.
While this works, global (hidden) environment variables make the under...The current build system depends on too many environment variables which are not explicitly passed to the build functions, but inferred from the current environment.
While this works, global (hidden) environment variables make the understanding of the program functionality harder. We should strive to make explicit all environment variable usage.André AnjosAndré Anjoshttps://gitlab.idiap.ch/bob/bob.devtools/-/issues/7Need global testing for README.rst2019-01-21T16:45:25ZAndré AnjosNeed global testing for README.rstFollowing docs#3, it would be good if a global test would be created to make sure:
1. All packages contain a `README.rst` file
2. This file passes basic linting - something like `rst2html README.rst >& /dev/null` should return 0 and emi...Following docs#3, it would be good if a global test would be created to make sure:
1. All packages contain a `README.rst` file
2. This file passes basic linting - something like `rst2html README.rst >& /dev/null` should return 0 and emit no warning messages.https://gitlab.idiap.ch/bob/bob.devtools/-/issues/6Please organize the development docs from here2019-01-25T14:42:20ZAmir MOHAMMADIPlease organize the development docs from hereAs discussed with @andre.anjos before, the docs in the template folder of bob.admin should go to bob.extension.As discussed with @andre.anjos before, the docs in the template folder of bob.admin should go to bob.extension.Conda-based CIAndré AnjosAndré Anjoshttps://gitlab.idiap.ch/bob/bob.devtools/-/issues/5Out-of-source test-only runs are bound to fail2019-02-20T13:58:52ZAndré AnjosOut-of-source test-only runs are bound to failThis is not an issue per see, but more like a feature request. With the current test strategy implemented in bob.admin#100, conda packages are supposed to contain all material that is required to test them.
This is not true for the `do...This is not an issue per see, but more like a feature request. With the current test strategy implemented in bob.admin#100, conda packages are supposed to contain all material that is required to test them.
This is not true for the `doc` directory of source packages, therefore testing any package inside `bob` out of its source present is bound to fail.
My impression is that this could be mitigated if we were not "building" the documentation during "test" time, but rather during build time.
It would be nice to understand why that is not the case, @amohammadi.André AnjosAndré Anjoshttps://gitlab.idiap.ch/bob/bob.devtools/-/issues/4Differentiating between errors in `conda build --test` and build skips2019-01-21T16:49:30ZAndré AnjosDifferentiating between errors in `conda build --test` and build skipsSome of our packages need to skip the build in certain architecture combinations. E.g. bob.ip.mtcnn should be skipped on everything but linux.
In order to detect this, I implemented the following condition:
```sh
$ pkgname=$(conda bui...Some of our packages need to skip the build in certain architecture combinations. E.g. bob.ip.mtcnn should be skipped on everything but linux.
In order to detect this, I implemented the following condition:
```sh
$ pkgname=$(conda build ... --output conda)
$ if [ -z "${pkgname}" ]; then echo "Build is going to be skipped..."; else echo "Runs the build"; fi
```
This has been working fine on the nightlies, where we build all packages and must skip some. Example:
https://gitlab.idiap.ch/bob/bob.nightlies/-/jobs/150738/raw
(look for the pattern `is NOT supposed to be built for`)
The problem is that if there is an error in the `$(conda build ... --output conda)` command, then it still exits with status 0 and returns empty.
In this case an error must be signed instead of just skipping the build. This is a problem, for example, when connectivity is down. See an example build that is supposed to fail (network error) and passes with a build skip:
https://gitlab.idiap.ch/beat/beat.core/-/jobs/150701
I tested locally by adding a `skip: true` line in the build section of a/any package and running `conda build ... --output conda` on it:
``sh
$ #edit conda/meta.yaml, add "skip: true" in the build section
$ conda build ../bob/bob.admin/conda/conda-build.sh --test --output --python=3.6 conda
$ echo $?
0
``
Now I turn off the network on my computer, to provoke an error:
```sh
$ conda build ... --output conda
WARNING:urllib3.connectionpool:Retrying (Retry(total=2, connect=None, read=None, redirect=None, status=None)) after connection broken by 'NewConnectionError('<urllib3.connection.VerifiedHTTPSConnection object at 0x109753e48>: Failed to establish a new connection: [Errno 8] nodename nor servname provided, or not known',)': /pkgs/main/osx-64/repodata.json.bz2
WARNING:urllib3.connectionpool:Retrying (Retry(total=1, connect=None, read=None, redirect=None, status=None)) after connection broken by 'NewConnectionError('<urllib3.connection.VerifiedHTTPSConnection object at 0x10975f080>: Failed to establish a new connection: [Errno 8] nodename nor servname provided, or not known',)': /pkgs/main/osx-64/repodata.json.bz2
WARNING:urllib3.connectionpool:Retrying (Retry(total=0, connect=None, read=None, redirect=None, status=None)) after connection broken by 'NewConnectionError('<urllib3.connection.VerifiedHTTPSConnection object at 0x10975f128>: Failed to establish a new connection: [Errno 8] nodename nor servname provided, or not known',)': /pkgs/main/osx-64/repodata.json.bz2
Returning non-final recipe for a.b.c; one or more dependencies was unsatisfiable:
Build: None
Host: python, setuptools
WARNING:conda_build.render:Returning non-final recipe for a.b.c; one or more dependencies was unsatisfiable:
Build: None
Host: python, setuptools
/Users/user/conda/conda-bld/osx-64/a.b.c-py36h51fb0fc_0.tar.bz2
$ echo $?
0
```
So, is there a way to check for this correctly? @amohammadi: any suggestions?André AnjosAndré Anjoshttps://gitlab.idiap.ch/bob/bob.devtools/-/issues/3"New" project templates2019-01-25T14:42:19ZAndré Anjos"New" project templatesA missing feature of this package is the ability to "start" new projects with the templates we used to have in bob.admin.
These have not yet been migrated, but should before the 1.0.0 release.
This issue tracks development on this.A missing feature of this package is the ability to "start" new projects with the templates we used to have in bob.admin.
These have not yet been migrated, but should before the 1.0.0 release.
This issue tracks development on this.André AnjosAndré Anjoshttps://gitlab.idiap.ch/bob/bob.devtools/-/issues/2Garbled output2019-04-19T19:10:12ZAndré AnjosGarbled outputWith the new build system, the output of commands and the CI infrastructure can sometimes get garbled:
https://gitlab.idiap.ch/bob/bob.devtools/-/jobs/154891
It would be nice to fix this.With the new build system, the output of commands and the CI infrastructure can sometimes get garbled:
https://gitlab.idiap.ch/bob/bob.devtools/-/jobs/154891
It would be nice to fix this.André AnjosAndré Anjoshttps://gitlab.idiap.ch/bob/bob/-/issues/257New Release - Bob - 6.0.02019-07-01T13:25:38ZTiago de Freitas PereiraNew Release - Bob - 6.0.0Nightlies are green;
https://gitlab.idiap.ch/bob/bob.nightlies/pipelines
Shall we consider a new release?Nightlies are green;
https://gitlab.idiap.ch/bob/bob.nightlies/pipelines
Shall we consider a new release?https://gitlab.idiap.ch/bob/bob.devtools/-/issues/1This packages ships its own conda settings2019-01-15T15:53:10ZAmir MOHAMMADIThis packages ships its own conda settingsEspecially I am talking about the conda_build_config.yaml file which is shipped with this package and it's not in sync with the one in bob.adminEspecially I am talking about the conda_build_config.yaml file which is shipped with this package and it's not in sync with the one in bob.adminhttps://gitlab.idiap.ch/bob/bob.db.casiasurf/-/issues/1Samples are images, not video2019-01-16T13:59:56ZGuillaume HEUSCHSamples are images, not videoSo a good idea would be to have an ImagePadFile class ... Right now everything is VideoPadFile.
Well, we can still consider an image as a video with one frameSo a good idea would be to have an ImagePadFile class ... Right now everything is VideoPadFile.
Well, we can still consider an image as a video with one frameGuillaume HEUSCHGuillaume HEUSCHhttps://gitlab.idiap.ch/bob/nightlies/-/issues/44Nightlies failure - bob.bio.vein2019-01-15T09:45:47ZTiago de Freitas PereiraNightlies failure - bob.bio.veinHi guys, I was looking at the log from the last nightly build and I saw that it fails on `bob.bio.vein`.
Looking at the both logs it fails on `bob.bio.vein` in the after build script.
https://gitlab.idiap.ch/bob/bob.nightlies/-/jobs/153...Hi guys, I was looking at the log from the last nightly build and I saw that it fails on `bob.bio.vein`.
Looking at the both logs it fails on `bob.bio.vein` in the after build script.
https://gitlab.idiap.ch/bob/bob.nightlies/-/jobs/153819/raw
https://gitlab.idiap.ch/bob/bob.nightlies/-/jobs/153820/raw
The last message is:
```
section_end:1546459397:after_script
�[0Ksection_start:1546459397:upload_artifacts_on_failure
�[0Ksection_end:1546459399:upload_artifacts_on_failure
�[0K�[31;1mERROR: Job failed: exit code 1
�[0;m
```
I suspect that something is wrong (maybe the size of the artifact is to big) with our gitlab server.
Same issue with `bob.bio.vein` builds
https://gitlab.idiap.ch/bob/bob.bio.vein/pipelines/25654https://gitlab.idiap.ch/bob/nightlies/-/issues/43Why are the nightly builds turned off?2018-12-19T08:22:01ZTiago de Freitas PereiraWhy are the nightly builds turned off?https://gitlab.idiap.ch/bob/bob.io.video/-/issues/14The video reader leaves the video files openned2018-11-29T08:26:31ZAmir MOHAMMADIThe video reader leaves the video files opennedLooks like the video reader `bob.io.video.reader` does not close the video file when iterating over frames AND the iteration does not finish fully.
For example:
```python
video = bob.io.base.load('testvideo.avi')
# closes the video file...Looks like the video reader `bob.io.video.reader` does not close the video file when iterating over frames AND the iteration does not finish fully.
For example:
```python
video = bob.io.base.load('testvideo.avi')
# closes the video file properly
for frame in reader('testvideo.avi'):
pass
# closes the video file properly
for frame in reader('testvideo.avi'):
break
# the video file is left open!
import gc
gc.collect()
# the video file is still open!
```Amir MOHAMMADIAmir MOHAMMADIhttps://gitlab.idiap.ch/bob/bob.pad.face/-/issues/26Frame-level scoring is the default for (some) baselines?2022-03-03T17:26:17ZAmir MOHAMMADIFrame-level scoring is the default for (some) baselines?Recently I noticed that the SVM PAD algorithm dumps scores as a frame-level basis and it is used by default in some baselines (e.g. `qm-svm`). However the evaluation using this frame-level analysis is much worse compared to video-level s...Recently I noticed that the SVM PAD algorithm dumps scores as a frame-level basis and it is used by default in some baselines (e.g. `qm-svm`). However the evaluation using this frame-level analysis is much worse compared to video-level scoring. For example, I saw the EER decrease from 28% to 15% when switching to video-level scoring. Shouldn't we change the defaults in our baselines?Amir MOHAMMADIAmir MOHAMMADIhttps://gitlab.idiap.ch/bob/nightlies/-/issues/42Nightlies are broken...... but for another reason2018-11-23T13:31:57ZTiago de Freitas PereiraNightlies are broken...... but for another reasonhttps://gitlab.idiap.ch/bob/bob.nightlies/-/jobs/152189
```
conda.exceptions.UnsatisfiableError: The following specifications were found to be in conflict:
- bob-devel=2018.10.16 -> cudatoolkit==9.0=h13b8566_0
- bob-devel=2018.10.16...https://gitlab.idiap.ch/bob/bob.nightlies/-/jobs/152189
```
conda.exceptions.UnsatisfiableError: The following specifications were found to be in conflict:
- bob-devel=2018.10.16 -> cudatoolkit==9.0=h13b8566_0
- bob-devel=2018.10.16 -> scikit-learn==0.19.2=py36h4989274_0
- numpy
Use "conda info <package>" to see the dependencies for each package.
During handling of the above exception, another exception occurred:
Traceback (most recent call last):
File "/local/builds/bob/bob.nightlies/miniconda/bin/conda-build", line 11, in <module>
sys.exit(main())
File "/local/builds/bob/bob.nightlies/miniconda/lib/python3.7/site-packages/conda_build/cli/main_build.py", line 439, in main
execute(sys.argv[1:])
File "/local/builds/bob/bob.nightlies/miniconda/lib/python3.7/site-packages/conda_build/cli/main_build.py", line 430, in execute
verify=args.verify, variants=args.variants)
File "/local/builds/bob/bob.nightlies/miniconda/lib/python3.7/site-packages/conda_build/api.py", line 201, in build
notest=notest, need_source_download=need_source_download, variants=variants)
File "/local/builds/bob/bob.nightlies/miniconda/lib/python3.7/site-packages/conda_build/build.py", line 2275, in build_tree
notest=notest,
File "/local/builds/bob/bob.nightlies/miniconda/lib/python3.7/site-packages/conda_build/build.py", line 1398, in build
raise e
File "/local/builds/bob/bob.nightlies/miniconda/lib/python3.7/site-packages/conda_build/build.py", line 1389, in build
channel_urls=tuple(m.config.channel_urls))
File "/local/builds/bob/bob.nightlies/miniconda/lib/python3.7/site-packages/conda_build/environ.py", line 751, in get_install_actions
raise DependencyNeedsBuildingError(exc, subdir=subdir)
conda_build.exceptions.DependencyNeedsBuildingError: Unsatisfiable dependencies for platform linux-64: {'cudatoolkit==9.0=h13b8566_0', 'numpy', 'scikit-learn==0.19.2=py36h4989274_0'}
```
As a consequence, ALL builds are brokenhttps://gitlab.idiap.ch/bob/bob.measure/-/issues/58ROC --no-semilogx plots are broken2019-08-19T13:56:42ZAmir MOHAMMADIROC --no-semilogx plots are brokenIn ROC plots, when --no-semilogx is provided the y axis values change from `1-FRR` to `FRR` and the plot looks very similar to a DET curve which is the expected behavior.
However, the dots shown on the plot (activated using --lines-at 1...In ROC plots, when --no-semilogx is provided the y axis values change from `1-FRR` to `FRR` and the plot looks very similar to a DET curve which is the expected behavior.
However, the dots shown on the plot (activated using --lines-at 1e-3,1e-4,...) are still drawn using the `1-FRR` values. Hence the dots do not fall on the ROC plots.
![roc-1](/uploads/9b4f42423d0d6b5b60abc4034f299411/roc-1.png)
![roc-2](/uploads/47172700f57876cae4185373578d6969/roc-2.png)
I have observed this behavior in `bob bio roc` but since it's implemented here, I suspect the bug is in here.https://gitlab.idiap.ch/bob/bob.bio.video/-/issues/14Frame Containers lose their order when they are saved into an hdf5 file2018-11-27T13:00:03ZAmir MOHAMMADIFrame Containers lose their order when they are saved into an hdf5 fileFor example I have a frame container going from 1 to 150 frames but when loaded back the frames are string sorted:
```python
In [3]: fc = bob.bio.video.FrameContainer(bob.io.base.HDF5File(path))
...For example I have a frame container going from 1 to 150 frames but when loaded back the frames are string sorted:
```python
In [3]: fc = bob.bio.video.FrameContainer(bob.io.base.HDF5File(path))
In [5]: for i, v, q in fc:
...: print(i)
0
1
10
100
101
102
103
104
105
106
```Amir MOHAMMADIAmir MOHAMMADIhttps://gitlab.idiap.ch/bob/bob.pad.base/-/issues/29convert_frame_cont_to_array does not work all the time2018-11-06T09:05:21ZAmir MOHAMMADIconvert_frame_cont_to_array does not work all the timeSometimes, the keys in frame containers are not actually integers and the code fails in https://gitlab.idiap.ch/bob/bob.pad.base/blob/172d46e0ca734ce682323c502f23abc53da15f4b/bob/pad/base/utils/helper_functions.py#L35 when keys are conve...Sometimes, the keys in frame containers are not actually integers and the code fails in https://gitlab.idiap.ch/bob/bob.pad.base/blob/172d46e0ca734ce682323c502f23abc53da15f4b/bob/pad/base/utils/helper_functions.py#L35 when keys are converted from strings to integers.Amir MOHAMMADIAmir MOHAMMADIhttps://gitlab.idiap.ch/bob/bob.db.cuhk_cufsf/-/issues/2Some annotations are inverted2018-11-19T18:26:19ZTiago de Freitas PereiraSome annotations are invertedI would need to redo some experimentsI would need to redo some experiments