add dm-acme 0.2.0 (and dm-env 1.3)

[bollwyvl]

Checklist

• Title of this PR is meaningful: e.g. "Adding my_nifty_package", not "updated meta.yaml"
• License file is packaged (see here for an example)
  • all on the up-and-up now!
  • brought in from repo, upstream PRs created
    • add license to source distributions google-deepmind/acme#3
    • add license to source distributions google-deepmind/dm_env#3
• Source is from official source
• Package does not vendor other packages. (If a package uses the source of another package, they should be separate packages or the licenses of all packages need to be packaged)
• If static libraries are linked in, the license of the static library is packaged.
• Build number is 0
• A tarball (url) rather than a repo (e.g. git_url) is used in your recipe (see here for more details)
• GitHub users listed in the maintainer section have posted a comment confirming they are willing to be listed there

Some notes:

• tree and env have tests
• can't build tree (and therefore anything else) because of bazel on windows
  • Windows support bazel-feedstock#36
• forced env to be noarch: python by ignoring EOL pythons
• seeing some linkage warnings on tree, could be my local setup

[conda-forge-linter]

Hi! This is the friendly automated conda-forge-linting service.

I wanted to let you know that I linted all conda-recipes in your PR (recipes/dm-acme, recipes/dm-env, recipes/dm-tree) and found some lint.

Here's what I've got...

For recipes/dm-acme:

• noarch packages can't have selectors. If the selectors are necessary, please remove noarch: python.

For recipes/dm-env:

• noarch packages can't have selectors. If the selectors are necessary, please remove noarch: python.

[conda-forge-linter]

Hi! This is the friendly automated conda-forge-linting service.

I wanted to let you know that I linted all conda-recipes in your PR (recipes/dm-acme, recipes/dm-env, recipes/dm-tree) and found some lint.

Here's what I've got...

For recipes/dm-env:

• Non noarch packages should have python requirement without any version constraints.
• Non noarch packages should have python requirement without any version constraints.

[conda-forge-linter]

Hi! This is the friendly automated conda-forge-linting service.

I just wanted to let you know that I linted all conda-recipes in your PR (recipes/dm-acme, recipes/dm-env, recipes/dm-tree) and found it was in an excellent condition.

[bollwyvl]

acme has a ton of optional dependencies, but many of them are to nightly builds, git remotes, etc. Will see how it develops.

[bollwyvl]

needs investigating:

  ERROR (dm-tree,lib/python3.7/site-packages/tree/_tree.cpython-37m-x86_64-linux-gnu.so): runpaths ['$ORIGIN/../../..'] found in /home/conda/staged-recipes/build_artifacts/dm-tree_1591122096842/_h_env_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_pla/lib/python3.7/site-packages/tree/_tree.cpython-37m-x86_64-linux-gnu.so

Any hot takes, @conda-forge/help-python-c ?

[scopatz]

This seems like it can be noarch. It doesn't compile in any deps

[scopatz]

Does it compile against numpy?

[bollwyvl]

Hm... I tried env/acme with noarch and skip#win (as there would be missing dependency on tree) and the linter was not amused. I can try some other combinations. As to numpy: in tree, it is used in its test suite only for benchmarking.

…

On Tue, Jun 2, 2020, 15:41 Anthony Scopatz ***@***.***> wrote: ***@***.**** commented on this pull request. ------------------------------ In recipes/dm-tree/meta.yaml <#11804 (comment)> : > + run: + - python + - six >=1.12.0 + - wrapt + +test: + imports: + - tree + commands: + - pip check + - python -m unittest tree.tree_test tree.tree_benchmark + requires: + - pip + - attrs + - absl-py + - numpy Does it compile against numpy? — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <#11804 (review)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAALCRBPCIYOYDBDLM4EG7TRUVIWTANCNFSM4NQ6WRTA> .

[scopatz]

RE numpy: ahh ok

[bollwyvl]

Perhaps I have to do some explicit binary patching: https://github.com/pytorch/builder/blob/master/conda/pytorch-nightly/build.sh

…

On Tue, Jun 2, 2020, 15:51 Anthony Scopatz ***@***.***> wrote: RE numpy: ahh ok — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <#11804 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAALCRHIHE3D54TWZQUCETDRUVJ45ANCNFSM4NQ6WRTA> .

[bollwyvl]

Had some more looking around the source... it does some strange things with [headers[(
https://github.com/deepmind/tree/tree/master/external). I am kinda wondering if we could get away with patching out the bazel build chain, as the c++ doesn't look like it uses much off the beaten path... will put some more effort towards it in the coming days, but any more pointers would be greatly appreciated!

[scopatz]

I have never really used bazel - I wish I could give more help here

[bollwyvl]

@scopatz no worries! It seems like quite an interesting beast.

I could imagine a future where bazel+conda made a lot of sense for maintaining an in-house family of channels... and could even someday be useful for conda-forge level work: how slick would a distributed build cache be? I think we'd need a lot more reproducibility in the core conda/conda-build toolchain for it to work for real.

As I mentioned, it seems overkill for this package, and maybe I can figure out patching it to work with stock compiler("c"). I'm not super under the gun to get this out: my goal here is part of a year-long effort to build a reinforcement learning "constructoricculum" which includes state-of-the-art packages like this one, benchmark data sets, (searhable, citeable) research papers, and a jupyterlab to drive it. So I'll likely be stuck having to deal with lots of crazy builds, and will have to up my game in a couple dimensions. Thanks for the support, though!

[stale]

Hi friend!

We really, really, really appreciate that you have taken the time to make a PR on conda-forge/staged-recipes! conda-forge only exists because people like you donate their time to build and maintain conda recipes for use by the community.

In an effort to maintain this repository and increase the signal-to-noise for open PRs, the maintainers of staged-recipes close excessively old PRs after six months. This PR will remain open for another month, and then will be closed.

If you'd like to keep it open, please comment/push and we will be happy to oblige! Note that very old PRs will likely need to be rebased on master so that they can be rebuilt with the most recent CI scripts. If you have any trouble, or we missed reviewing this PR in the first place (sorry!), feel free to ping the team using a special command in a comment on the PR to get the attention of the staged-recipes team.

Cheers and thank you for contributing to this community effort!

[vnlitvinov]

FYI: dm-tree was added in #13248 and is available now: https://anaconda.org/conda-forge/dm-tree

[bollwyvl]

Thanks for the update!

[vnlitvinov]

@bollwyvl please let me know if you wish to become a co-maintaner for dm-tree (or make a PR adding yourself there), I took tests section from your PR after all 😃

[conda-forge-linter]

Hi! This is the friendly automated conda-forge-linting service.

I just wanted to let you know that I linted all conda-recipes in your PR (recipes/dm-acme, recipes/dm-env) and found it was in an excellent condition.

[bollwyvl]

@vnlitvinov i am generally looking to not take on another exotic build chain, but feel free to ping me if you run into issues over on tree!

Honestly, I had forgotten that tree was where all the complexity was, and that env and acme can be noarch now. Also a ton of the extras are available now, so could add, but would be happy to see these go through in their current form. Also all the LICENSE PRs landed, so this is now a really simple PR.

Anyhow, hoping for good results on CI!

[bollwyvl]

@conda-forge/staged-recipes @conda-forge/help-python (finally) ready for review!

[vnlitvinov]

i am generally looking to not take on another exotic build chain, but feel free to ping me if you run into issues over on tree!

Same here :) I needed dm-tree as a dependency for #11160 (of which I only need ray-core, but for the sake of completeness I'm going to finish packaging the whole project).

[bollwyvl]

You're a hero!