Convert to/from Mash sketch format

[lskatz]

Hi @ctb, I was wondering if you could make a Sourmash subcommand to convert back and forth between YAML and the format used for Mash sketches. This would help increase compatibility between Mash and Sourmash.

[ctb]

Hi Lee, the code is already there! Sort of. See the 'import_csv' command of sourmash. Also see: marbl/Mash#27 (comment) Disclaimer: this has been back burnered, but the code should still work. You might also be interested in https://github.com/dib-lab/sourmash/blob/dd18fdd89d4a061b8fca2ee7e529b0064066704b/utils/compute-dna-mh-another-way.py which shows you how to construct a JSON (we switched recently :) signature from any collection of hashes.

[lskatz]

Ok. So to import a mash sketch to Sourmash, I would use the mash info -d command to make a dump and then sourmash import_csv?

[ctb]

On Mon, Jan 09, 2017 at 08:24:10AM -0800, Lee Katz wrote: Ok. So to import a mash sketch to Sourmash, I would use the `mash info -d` command to make a dump and then `sourmash import_csv`?

Yes, I think. I'm not sure where the mash support for this stands, but the sourmash side of things is in our tests so that code should work! If you run into mash trouble let me know and I'll see what I can do.

[lskatz]

Thanks!

[ctb]

(and if it works and you get a chance to report back, that'd be nice too :)

[ctb]

I've hacked together a script, utils/mash-to-sourmash-sig.py, that reads the output of 'mash info -d' and outputs it to a sourmash signature. See #121. It's hacky for the moment but seems to basically work - I'll improve it before #121 is committed, and add some tests.

Refs:

• Exporting mash hashes for interoperability marbl/Mash#27
• Convert to/from Mash sketch format #99
• Getting hashes out of .msh marbl/Mash#35

[lskatz]

Hi @ctb , is there a new version of this command line script for converting JSON<=>MSH? I don't see this script in the latest version but I would like to use it again.

Traceback (most recent call last):
  File "/scicomp/home/gzu2/src/mashtree/bin/mash-to-sourmash-sig.py", line 20, in <module>
    E = sourmash_lib.Estimators(ksize=ksize, n=num, is_protein=False)
AttributeError: 'module' object has no attribute 'Estimators'

[ctb]

Yes! Please see sourmash signature import and export!!

…

-- Titus Brown, ctbrown@ucdavis.edu<mailto:ctbrown@ucdavis.edu> On Jan 15, 2019, at 11:27 AM, Lee Katz <notifications@github.com<mailto:notifications@github.com>> wrote: Hi @ctb<https://github.com/ctb> , is there a new version of this command line script for converting JSON<=>MSH? I don't see this script in the latest version but I would like to use it again. — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub<#99 (comment)>, or mute the thread<https://github.com/notifications/unsubscribe-auth/AADHSL0C6VuiVM3h4Tn8fSdZV-U4ohvcks5vDiuvgaJpZM4LeUNF>.

[lskatz]

Oh that's great! However, I got an error Unrecognized command for sig, import, and signature

Unrecognized command

sourmash <command> [<args>]

** Commands include:

compute <filenames>         Compute MinHash signatures for sequences in files.
compare <filenames.sig>     Compute similarity matrix for multiple signatures.
search <query> <against>    Search a signature against a list of signatures.
plot <matrix>               Plot a distance matrix made by 'compare'.
gather                      Search a metagenome signature for multiple
                                 non-overlapping matches.

** Taxonomic classification utilities:

   Run 'sourmash lca' for the taxonomic classification routines.

** Sequence Bloom Tree (SBT) utilities:

index                   Index a collection of signatures for fast searching.
sbt_combine             Combine multiple SBTs into a new one.
categorize              Identify best matches for many signatures using an SBT.
watch                   Classify a stream of sequences.

** Other information:

info                        Sourmash version and other information.

Use '-h' to get subcommand-specific help, e.g.

sourmash compute -h

** Documentation is available at https://sourmash.readthedocs.io/

[lskatz]

sourmash version 2.0.0a11
- loaded from path: /scicomp/home/gzu2/bin/anaconda2/lib/python2.7/site-packages/sourmash

[ctb]

Ahh, sorry, the very latest isn’t released yet on conda or pypi, I guess! You can still use the script from an earlier version just fine. We will release 2.0 very soon.

…

-- Titus Brown, ctbrown@ucdavis.edu<mailto:ctbrown@ucdavis.edu> On Jan 15, 2019, at 11:37 AM, Lee Katz <notifications@github.com<mailto:notifications@github.com>> wrote: sourmash version 2.0.0a11 - loaded from path: /scicomp/home/gzu2/bin/anaconda2/lib/python2.7/site-packages/sourmash — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub<#99 (comment)>, or mute the thread<https://github.com/notifications/unsubscribe-auth/AADHSBEf38RjSKtiyrRIePbVxJc6slgwks5vDi4MgaJpZM4LeUNF>.