[Debian-med-packaging] Bug#961013: diamond-aligner: Diamond faill with the error: Assertion `index >= 0 && index < size()' failed.

[Andreas Tille]

Control: forwarded -1 https://github.com/bbuchfink/diamond/issues/351

Hi Marcelo,

thank you for the effort you have spent into this issue.  I' watching
the upstream issue and due to the first line in this mail also the
Debian Bug Tracking system will do so.

On Tue, May 19, 2020 at 08:19:44AM -0300, Marcelo Luiz de Laia wrote:
> Package: diamond-aligner
> Version: 0.9.32-2
> Severity: important
> 
> Dear Maintainer,
> 
> I run diamond like this:
> 
> diamond blastx --threads 2 --outfmt 5 --max-target-seqs 5 --evalue 1e-5 --db
> /home/marcelo/Diamond_DB/nr --query /home/marcelo/Google\
> Drive/Mlaia/NoISeq/ExonID_GDE.fa -o /home/marcelo/Google\
> Drive/Mlaia/Blastn/blastx.exons.diamond.xml
> 
> And I got this error:
> 
> marcelo at marcelo:~/Google Drive/Mlaia/Blastn$ diamond blastx --outfmt 5 --max-
> target-seqs 5 --evalue 1e-5 --db /media/marcelo/Seagate\ Expansion\
> Drive/NCBI_DB/Diamond/nr --query /home/marcelo/Google\
> Drive/Mlaia/NoISeq/ExonID_GDE.fa -o /home/marcelo/Google\
> Drive/Mlaia/Blastn/blastx.exons.diamond.xml
> diamond v0.9.32.133 (C) Max Planck Society for the Advancement of Science
> Documentation, support and updates available at http://www.diamondsearch.org
> 
> #CPU threads: 4
> Scoring parameters: (Matrix=BLOSUM62 Lambda=0.267 K=0.041 Penalties=11/1)
> Temporary directory: /home/marcelo/Google Drive/Mlaia/Blastn
> Opening the database...  [0s]
> #Target sequences to report alignments for: 5
> Reference = /media/marcelo/Seagate Expansion Drive/NCBI_DB/Diamond/nr.dmnd
> Sequences = 282858011
> Letters = 101765175729
> Block size = 2000000000
> Opening the input file...  [0s]
> Opening the output file...  [0s]
> Loading query sequences...  [0.046s]
> Masking queries... diamond: /build/diamond-aligner-jUHUMq/diamond-
> aligner-0.9.32/src/lib/Eigen/src/Core/DenseCoeffsBase.h:408:
> Eigen::DenseCoeffsBase<Derived, 1>::Scalar& Eigen::DenseCoeffsBase<Derived,
> 1>::operator[](Eigen::Index) [with Derived = Eigen::Array<float, -1, 1>;
> Eigen::DenseCoeffsBase<Derived, 1>::Scalar = float; Eigen::Index = long int]:
> Assertion `index >= 0 && index < size()' failed.
> diamond: /build/diamond-aligner-jUHUMq/diamond-
> aligner-0.9.32/src/lib/Eigen/src/Core/DenseCoeffsBase.h:408:
> Eigen::DenseCoeffsBase<Derived, 1>::Scalar& Eigen::DenseCoeffsBase<Derived,
> 1>::operator[](Eigen::Index) [with Derived = Eigen::Array<float, -1, 1>;
> Eigen::DenseCoeffsBase<Derived, 1>::Scalar = float; Eigen::Index = long int]:
> Assertion `index >= 0 && index < size()' failed.
> Abortado

(Nice, "Abortado" sounds way more kind than "Aborted" ;-)) )

> marcelo at marcelo:~/Google Drive/Mlaia/Blastn$
> 
> I opened a issue on diamond support forum at
> https://github.com/bbuchfink/diamond/issues/351
> 
> Please, could you see that thread? May be this bug is only in the Debian
> version, because I run the github version 0.32 and it worked out of the box.

Thanks a lot.  That's very helpful information.  Same question as inside
the github issue:  Are you able to provide the dataset you have used?
Additional question:  Does the call work with *other* datasets you might
have used?

We try to debug the issue and for sure the Debian packaged version
should provide the same results the code that is fetched from upstream.
We should definitely add this somehow in an autopkgtest.

One quick idea came up that we might link against Debian packaged
libeigen - but when checking this is actually not the case.  This is bad
Debian policy wise but at least for the case of debugging this issue its 
good to work on the very same code basis.

Kind regards

      Andreas.


-- 
http://fam-tille.de