It is time for a new release

Hi people,

I'm just finishing to publish some packages from bob. So, It is time for new release of bob.bio.* (bob.io.base, bob.bio.face, bob.bio.video, bob.bio.vein, bob.bio.spear, bob.db.bio_filelist and bob.bio.csu). I've seem some people working on master branches and this is not healthy at all.

Can I do it tomorrow?? We need to end this cycle of development.

Cheers

@bob

To be more objective, what is blocking us to have a new version??

A big green light from my side.

I think, bob.bio.csu is not green at the moment. There are still three things missing:

1. To be able to run the complete test suite, you need to incorporate the steps from the old travis.yml: https://github.com/bioidiap/bob.bio.csu/blob/master/.travis.yml#L24 into the gitlab-ci.yml
2. You need to update the buildout.cfg to incorporate the [csu-dir] from the develop.cfg: https://gitlab.idiap.ch/bob/bob.bio.csu/blob/master/develop.cfg#L40 and L43
3. We have to implement the data loading in https://github.com/bioidiap/bob.bio.csu/blob/master/bob/bio/csu/preprocessor/LDAIR.py#L144 and https://github.com/bioidiap/bob.bio.csu/blob/master/bob/bio/csu/preprocessor/LRPCA.py#L133 to use the new version, i.e., passing the read_original_data to the base class' constructor.

Steps 2 and 3, I can do, but not step 1.

Cheers Manuel

I'm having a look

Reassigned to @tiago.pereira

mentioned in issue bob.bio.csu#1 (closed)

I did steps 2 and 3. They are working fine in a branch that I created, but step one is difficult to do.

Please read bob.bio.csu#1 (closed) for more info.

Can we move forward?

I will work on bob.bio.csu today, i.e., in order to make it work with the latest version of bob.bio.base and bob.bio.face.

One question: Did you actually modify the bob.bio.face, bob.bio.video and bob.bio.spear packages to work with the new way of reading original data in bob.bio.base? I somehow lost track of that...

Cheers Manuel

Hey @mguenther, I did that alredy

Check the branch https://gitlab.idiap.ch/bob/bob.bio.csu/tree/standarization

I am not sure that this works. The read_original_data function has changed it's interface (it is now getting BioFile instances instead of filenames). I will update that.

But thanks for letting me know about the branch :-)

I should squash some commits. I introduced a lot of noise when I was trying to fix the CI

I also don't like the implementation in bob.bio.video. The preprocessor.Wrapper gets a frame_selector as parameter: https://gitlab.idiap.ch/bob/bob.bio.video/blob/master/bob/bio/video/preprocessor/Wrapper.py#L53 but this is used nowhere. Instead, you have to write a new read_original_data function, when you want to use a different frame_selector: https://gitlab.idiap.ch/bob/bob.bio.video/blob/master/bob/bio/video/test/test_preprocessor.py#L47

I will check if I cannot find a better solution...

Btw, since we are releasing bob.bio.* packages, I would also release bob.pad.base and bob.pad.voice, since they depend and extend bob.bio.* family. Unfortunately, we don't have any bob.pad.face yet, but I hope it will materialize at some point soon.

See bob.bio.video!4 (merged) for a better implementation of the IO for bob.bio.video

Oh, yet another point that needs to be fixed before we can make a new release. We have to fix #36 (closed). With improper/outdated documentation, code is not worth much.

In fact, current build are also failing because of documentation issues.

I will take care of issue #36 (closed) and the builds. @andre.anjos made all the warnings as errors in the sphinx-build. We'll see more of this. I will take care of this too.

wow, bob.bio.base has 143 WARNINGS in the documentation.

enjoy

I hope this flag does improve our doc quality.

Hi, follow bellow the packages that are ready to be published:

• bob.bio.gmm
• bob.bio.face
• bob.bio.spear
• bob.db.bio_filelist
• bob.bio.video

We still have some bits in thebob.bio.base. We have three issues open.

• The issue #36 (closed) I updated the installation instructions and the database implementation details. I think it is fine, if someone wants to include something fell free to do it
• The issue #43 (closed) @andre.anjos will solve it
• The issue #41 (closed) is on the way to be merged

@pkorshunov , in your side, with the bob.pad framework, is everything ok?

You are waiting for us, right?

Thanks

@tiago.pereira Well, I kinda am waiting, yes, but since bob.pad packages depend on bob.bio.base, we first need to ensure that everything passes with the latest bob.bio.base.

Things are good now IMHO. We are ready for a release after we merge/not merge !44 (merged)

@mguenther we would really appreciate it if you would not directly push to master.

And your /scratch/mgunther is not visible in our machines, only yours.

Sorry, there must have been a copy-paste error while writing the commit message. I actually wanted to point to #40 (closed).

I agree, pushing to master is not something that one (I) should do regularly. However, this fix was soo small that creating a PR would be soo much of an overhead.

Anyways, if you insist, I'll try to avoid pushing to master next time.

Yes I do insist. I would like to know what is going on and when you push directly to master, I don't notice it. Thank you.

Hey @mguenther , let's try to avoid push code direct to master, it is good for everybody. However, for minor things, and for an eminent release (where everyone is aware on what is going on) I don't see problem.

It seems that the number of issues are increasing. However, the issues are not blocking so I vote to release bob.bio (bob.bio.base, bob.bio.gmm, bob.bio.face, bob.bio.spear, bob.db.bio_filelist, bob.bio.video ) today, otherwise we will never do it. Other issues can be solved with patch versions.

Furthermore, we have to publish ALL the database packages in order to be compliant with the new bob.bio. This will take some time considering that we need to fix the sphinx WARNINGS one by one.

I agree to release the bob.bio family of packages. It will push us to update everything depending on this, including databases and bob.pad family.

I fixed the sphinx warnings for most of database packages. I will start to publish them.

I will bump the minor versions

@pkorshunov is it ok to publish bob.db.asvspoof and bob.db.avspoof?

@tiago.pereira Yes, these packages can be publish but I can publish them, no worries.

THanks @pkorshunov , I will do it. I'm already with the my finger on the trigger :-P

The databases are published.

Shall we move to bob.bio and bob.pad??

@amohammadi, is everything ok in your side?

Can I do it @amohammadi ?

Please go ahead, i don't think we are going to break API anymore. We'll have to fix a couple of things with patch versions.

Ok

Hi people,

"That's one small step for man, one giant leap for mankind"

We have our first release of bob.bio framework after the refactoring 2016. Follow bellow the released softwares:

• bob.bio.base in https://pypi.python.org/pypi/bob.bio.base/3.0.0 (103 commits - https://gitlab.idiap.ch/bob/bob.bio.base/tags/v3.0.0#v300)
• bob.bio.face in https://pypi.python.org/pypi/bob.bio.face/3.0.0 (42 commits - https://gitlab.idiap.ch/bob/bob.bio.face/tags/v3.0.0#v300)
• bob.bio.spear in https://pypi.python.org/pypi/bob.bio.spear/3.0.0 (36 commits - https://gitlab.idiap.ch/bob/bob.bio.spear/tags/v3.0.0#v300)
• bob.bio.video in https://pypi.python.org/pypi/bob.bio.video/3.0.0 (28 commits - https://gitlab.idiap.ch/bob/bob.bio.video/tags/v3.0.0#v300)
• bob.bio.gmm in https://pypi.python.org/pypi/bob.bio.gmm/3.0.0 (13 commits - https://gitlab.idiap.ch/bob/bob.bio.gmm/tags/v3.0.0#v300)

We also refactored around 30 database package which makes the list of commits even bigger.

Thanks everybody for the work, specially @amohammadi that organized everything. In case of problems, don't hesitate to open an issue and assign to him :-P

Status changed to closed