Tag new release and create Bioconda recipe

[heuermh]

I may have done something wrong building a Docker image (been spoiled by Bioconda/Biocontainers for too long :smile:), it works ok interactively

$ docker run -it heuermh/seqwish-dev /bin/bash
root@c523ee06a258:/seqwish# seqwish --help
  seqwish {OPTIONS}

    seqwish: a variation graph inducer
...

$ docker run -it heuermh/seqwish-dev seqwish
  seqwish {OPTIONS}

    seqwish: a variation graph inducer

but not when run by Nextflow (copied from .command.run)

$ docker run -i \
  -v /Users/heuermh/working/viral-assembly/work:/Users/heuermh/working/viral-assembly/work \
  -v "$PWD":"$PWD" \
  -w "$PWD" \
  --entrypoint /bin/bash heuermh/seqwish-dev \
  -c "/bin/bash -ue /Users/heuermh/working/viral-assembly/work/2a/bec35575e4f91b50b4a41bb6a73e61/.command.sh"

/Users/heuermh/working/viral-assembly/work/2a/bec35575e4f91b50b4a41bb6a73e61/.command.sh:
line 2: seqwish: command not found

Docker file, Docker Hub repo

Perhaps the best thing to do would be to tag a new release and create a Bioconda recipe.

[ekg]

Ok I can do that. Will in the morning.

On Mon, Mar 16, 2020, 15:28 Michael L Heuer notifications@github.com wrote:

I may have done something wrong building a Docker image (been spoiled by Bioconda/Biocontainers for too long 😄), it works ok interactively

$ docker run -it heuermh/seqwish-dev /bin/bash

root@c523ee06a258:/seqwish# seqwish --help

seqwish {OPTIONS}

seqwish: a variation graph inducer

...

$ docker run -it heuermh/seqwish-dev seqwish

seqwish {OPTIONS}

seqwish: a variation graph inducer

but not when run by Nextflow (copied from .command.run)

$ docker run -i \

-v /Users/heuermh/working/viral-assembly/work:/Users/heuermh/working/viral-assembly/work \

-v "$PWD":"$PWD" \

-w "$PWD" \

--entrypoint /bin/bash heuermh/seqwish-dev \

-c "/bin/bash -ue /Users/heuermh/working/viral-assembly/work/2a/bec35575e4f91b50b4a41bb6a73e61/.command.sh"

/Users/heuermh/working/viral-assembly/work/2a/bec35575e4f91b50b4a41bb6a73e61/.command.sh:

line 2: seqwish: command not found

Docker file https://github.com/heuermh/docker-linuxbrew-bio/blob/master/seqwish-dev/Dockerfile, Docker Hub repo https://hub.docker.com/repository/docker/heuermh/seqwish-dev

Perhaps the best thing to do would be to tag a new release and create a Bioconda recipe.

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/ekg/seqwish/issues/34, or unsubscribe https://github.com/notifications/unsubscribe-auth/AABDQEM2HPMRKP7WAYKINF3RHYZPJANCNFSM4LMJAQZA .

[ekg]

FYI we (me and @urbanslug) have found two bugs with the current master. But an unstable release can be cut.

On Mon, Mar 16, 2020, 22:37 Erik Garrison erik.garrison@gmail.com wrote:

Ok I can do that. Will in the morning.

On Mon, Mar 16, 2020, 15:28 Michael L Heuer notifications@github.com wrote:

I may have done something wrong building a Docker image (been spoiled by Bioconda/Biocontainers for too long 😄), it works ok interactively

$ docker run -it heuermh/seqwish-dev /bin/bash

root@c523ee06a258:/seqwish# seqwish --help

seqwish {OPTIONS}

seqwish: a variation graph inducer

...

$ docker run -it heuermh/seqwish-dev seqwish

seqwish {OPTIONS}

seqwish: a variation graph inducer

but not when run by Nextflow (copied from .command.run)

$ docker run -i \

-v /Users/heuermh/working/viral-assembly/work:/Users/heuermh/working/viral-assembly/work \

-v "$PWD":"$PWD" \

-w "$PWD" \

--entrypoint /bin/bash heuermh/seqwish-dev \

-c "/bin/bash -ue /Users/heuermh/working/viral-assembly/work/2a/bec35575e4f91b50b4a41bb6a73e61/.command.sh"

/Users/heuermh/working/viral-assembly/work/2a/bec35575e4f91b50b4a41bb6a73e61/.command.sh:

line 2: seqwish: command not found

Docker file https://github.com/heuermh/docker-linuxbrew-bio/blob/master/seqwish-dev/Dockerfile, Docker Hub repo https://hub.docker.com/repository/docker/heuermh/seqwish-dev

Perhaps the best thing to do would be to tag a new release and create a Bioconda recipe.

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/ekg/seqwish/issues/34, or unsubscribe https://github.com/notifications/unsubscribe-auth/AABDQEM2HPMRKP7WAYKINF3RHYZPJANCNFSM4LMJAQZA .

[heuermh]

Thanks! Or if using the last tag v0.2 is better, I don't mind either way.

[foreveremain]

hi - any progress on that bioconda recipe :) just trying a build on osx so may be able to add another platform.

[ekg]

Use v0.2.1. I've just tagged it.

On Fri, Mar 27, 2020 at 12:58 PM Jim Procter notifications@github.com wrote:

hi - any progress on that bioconda recipe :) just trying a build on osx so may be able to add another platform.

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/ekg/seqwish/issues/34#issuecomment-604961237, or unsubscribe https://github.com/notifications/unsubscribe-auth/AABDQEPE6XUILLTCQSLWQFDRJSIHTANCNFSM4LMJAQZA .

[foreveremain]

mmapallocator isn't playing on x86_64-apple-darwin13.4.0 - probably something simple (of course) but my clang is rusty - errors all look similar to: ./mmappable_vector.h:117:10: error: no member named '_M_impl' in 'std::__1::vector<int, mmap_allocator_namespace::mmap_allocator >' Base::_M_impl._M_finish = Base::_M_impl._M_start + n;

pausing for now - will pick up again in 24hrs or so unless someone throws me a bone here :)

[ekg]

Can you work in docker?

Is this an issue with clang vs gcc? What compiler and version are you using?

On Fri, Mar 27, 2020, 13:47 Jim Procter notifications@github.com wrote:

mmapallocator isn't playing on x86_64-apple-darwin13.4.0 - probably something simple (of course) but my clang is rusty - errors all look similar to: ./mmappable_vector.h:117:10: error: no member named '_M_impl' in 'std::__1::vector<int, mmap_allocator_namespace::mmap_allocator >' Base::_M_impl._M_finish = Base::_M_impl._M_start + n;

pausing for now - will pick up again in 24hrs or so unless someone throws me a bone here :)

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/ekg/seqwish/issues/34#issuecomment-604981781, or unsubscribe https://github.com/notifications/unsubscribe-auth/AABDQEMUBBYWFHABPKPHVGTRJSN55ANCNFSM4LMJAQZA .

[foreveremain]

Clang 9.0.0 - sticking to a vanilla conda installation, but still running high sierra so base install may be old. I could run in docker of course but prefer not to if this is ending up as a conda recipe :) [edit: conda update to clang 9.0.1]

[heuermh]

Thanks for making the new tag, @ekg!

Since my earlier comment above I fixed the build-from-source Docker image linked above (heuermh/seqwish-dev on Docker Hub) to work correctly with Nextflow. I can also add a new Docker image based on the release tag (heuermh/seqwish) if that helps.

I am not a conda expert, however. The only recipes I have done are for JVM or pyspark applications.

[foreveremain]

@heuermh - appreciate the update. I've given up on the osx build for the moment, since I'm supposed to be doing something else with that machine, but have (I think) a working conda build environment for seqwish now so it should be trivial to create the recipe. I'll come back to the osx build when I've done what I should have been doing in the last 4 hours :)

[heuermh]

Work in progress here https://github.com/bioconda/bioconda-recipes/pull/21199

I am at a similar place now with the Bioconda build on OSX

...
16:50:46 BIOCONDA INFO (ERR) /Users/distiller/project/miniconda/conda-bld/seqwish_1585612045569/work/deps/mmap_allocator/mmappable_vector.h:117:10: error: no member named '_M_impl' in 'std::__1::vector<int, mmap_allocator_namespace::mmap_allocator<int> >'
16:50:46 BIOCONDA INFO (ERR)                         Base::_M_impl._M_finish = Base::_M_impl._M_start + n;
16:50:46 BIOCONDA INFO (ERR)                               ^

https://circleci.com/gh/bioconda/bioconda-recipes/101384

[heuermh]

Fixed downstream, though the Bioconda recipe does not build on OSX.