Update ltr_retriever to 2.9.8

[BiocondaBot]

Update ltr_retriever: 2.9.5 → 2.9.8

Info	Link or Description
Recipe	recipes/ltr_retriever (click to view/edit other files)
Summary	Sensitive and accurate identification of LTR retrotransposons
Home	https://github.com/oushujun/LTR_retriever
Releases	https://github.com/oushujun/LTR_retriever/tags
Author	@oushujun

---

This pull request was automatically generated (see docs).

[oushujun]

@Juke34 Can you please help to fix this? GPT suggests adding this line, but I am unsure if this fit:

build:
  number: 0
  noarch: generic
  run_exports:
    - {{ pin_subpackage('ltr_retriever', max_pin='x.x.x') }}

[oushujun]

@mencian thanks for fixing the recipe! I tested with new requirements (loosening version limits for rmblast and libstcxx-ng) via conda install and it worked well, however, I need to specify channels of bioconda and conda-forge to solve. You have all channel specifications removed but still passing the test, are channels not required by bioconda recipes anymore? Thanks again.

[mencian]

however, I need to specify channels of bioconda and conda-forge to solve.

Hi @oushujun, there should be no need to specify the bioconda and conda-forge channels in meta.yaml because the channel priorities have already been configured to use: conda-forge --> bioconda --> defaults (see the Bioconda usage docs).

Will the package not install locally if you do not explicitly list the bioconda and conda-forge channels?

[mencian]

@BiocondaBot please fetch artifacts

[BiocondaBot]

Package(s) built on Azure are ready for inspection:

Arch	Package	Zip File
noarch	ltr_retriever-2.9.8-hdfd78af_0.tar.bz2	LinuxArtifacts

---

You may also use conda to install these after downloading and extracting the appropriate zip file. From the LinuxArtifacts or OSXArtifacts directories:

conda install -c ./packages <package name>

---

Docker image(s) built (images are in the LinuxArtifacts zip file above):

Package	Tag	Install with docker
ltr_retriever	2.9.8--hdfd78af_0	show gzip -dc LinuxArtifacts/images/ltr_retriever:2.9.8--hdfd78af_0.tar.gz | docker load

[oushujun]

@mencian I didn't manually add channels to my ~/.condarc, so it looks like:

channels:
  - bioconda
  - defaults

That's why I need to manually add channels for mamba to solve:
mamba create -n LTR_retriever2.9.8 perl-text-soundex perl cd-hit "repeatmasker<4.1.5" rmblast libstdcxx-ng rmblast tesorter -c bioconda -c conda-forge

I also tested the Azure-built package ltr_retriever-2.9.8-hdfd78af_0.tar.bz2, it requires manually adding the two channels to solve:
mamba install -c ./LinuxArtifacts/packages ltr_retriever -c bioconda -c conda-forge

Will it be more robust if the required channels are included in the recipe?

[oushujun]

@mencian If you can add back the channels, I can try again. Thank you!

[mencian]

@oushujun I can certainly add the channels back into the recipe but it seems to me your situation is an isolated case. Would it be too much trouble to ask you to add conda-forge to your .condarc?

There is really no reason for conda-forge not to be in your .condarc as bioconda recipes depend on many of the generic packages in conda-forge. You also no longer have to keep specifying -c conda-forge -c bioconda everytime you install a package.

[oushujun]

@mencian I don't think my case is unique because obviously conda-forge is not the default channel when you install conda (well, it is the case for miniforge), thus Conda suggests you add these channels to your .condarc after installation. This may be related to channel priorities while solving. Specifying channels and their orders in each recipe will avoid these potential issues and enhance the robustness on the users' end.

[mencian]

@BiocondaBot please fetch artifacts

[BiocondaBot]

Package(s) built on Azure are ready for inspection:

Arch	Package	Zip File
noarch	ltr_retriever-2.9.8-hdfd78af_0.tar.bz2	LinuxArtifacts

---

You may also use conda to install these after downloading and extracting the appropriate zip file. From the LinuxArtifacts or OSXArtifacts directories:

conda install -c ./packages <package name>

---

Docker image(s) built (images are in the LinuxArtifacts zip file above):

Package	Tag	Install with docker
ltr_retriever	2.9.8--hdfd78af_0	show gzip -dc LinuxArtifacts/images/ltr_retriever:2.9.8--hdfd78af_0.tar.gz | docker load

[mencian]

@oushujun I've added the list of channels back into the recipe with a few modifications to priority: conda-forge -> bioconda -> defaults. Please try installing the package locally and see if everything works as expected.

[oushujun]

@mencian Thank you for adding the channels. I tried the latest build and it still has version conflicts:

$ conda install -c ./LinuxArtifacts/packages/ ltr_retriever
Collecting package metadata (current_repodata.json): done
Solving environment: failed with initial frozen solve. Retrying with flexible solve.
Solving environment: failed with repodata from current_repodata.json, will retry with next repodata source.
Collecting package metadata (repodata.json): done
Solving environment: failed with initial frozen solve. Retrying with flexible solve.
Solving environment: \ 
Found conflicts! Looking for incompatible packages.
This can take several minutes.  Press CTRL-C to abort.
failed                                                                                                                                                        

UnsatisfiableError: The following specifications were found to be incompatible with your system:

  - feature:/linux-64::__glibc==2.35=0
  - ltr_retriever -> libstdcxx-ng -> __glibc[version='>=2.17']

Your installed version is: 2.35

Adding the conda-forge channel had it solved:

$ conda install -c ./LinuxArtifacts/packages/ ltr_retriever -c conda-forge

Also adding the conda-forge channel to ~/.condarc had it solved. Do you know why conda is not recognizing channel specifications in the recipe?

Shujun

[oushujun]

I learned that the local channel specification supersed the channels specified in the recipe. So there's no problem with this recipe now. Thank you! @mencian

[oushujun]

@BiocondaBot please merge

[mencian]

I learned that the local channel specification supersed the channels specified in the recipe. So there's no problem with this recipe now. Thank you! @mencian

If you do not mind, I am going to remove the channels in the recipe to clean it up a bit, and the channels section is redundant anyways.