RE: Table Information extraction tools in DebianMed?
Thank you for these information. This was very useful. So I guess the conclusion would be to package Jena and OpenNLP and other Apache related libraries that I use for the start, since they are dependencies of my project. Later, I can possibly package my solution, and since MetaMap tagging is not a core feature, I can package it without it, I don't think it would be a big problem. So for time being, I would leave MetaMap by the side. I will also check what all need to be packaged, what are my dependencies, but they are mainly apache related, apart from Weka that is already packed.
Thank you,
Nikola Milosevic
-----Original Message-----
From: Sascha Steinbiss [mailto:satta@tetrinetsucht.de]
Sent: 20 July 2016 22:26
To: Nikola Milosevic
Cc: Andreas Tille; debian-med@lists.debian.org
Subject: Re: Table Information extraction tools in DebianMed?
Hi Nikola,
> I have used quite a few libraries as .jar files, such as weka, openNLP, JENA, MetaMap, etc. Does what you said mean that I need to package all these .jars before I package mine?
AFAICS weka is already packaged [1].
As far as Jena and openNLP is concerned, these seem to be Apache projects and Apache licensed, which should be fine.
> Also, what happens if some package, such as MetaMap is not open source?
MetaMap (which does provide source code) indeed has a weird license [2]. It doesn’t look too bad at first, but then there are things like:
"Users of the data distributed with MetaMap and MetaMap Tools are responsible for compliance with the UMLS Metathesaurus License Agreement which requires you to respect the copyrights of the constituent vocabularies and to file a brief annual report on your use of the UMLS. You also must have activated a UMLS Terminology Services (UTS) account.”
which IMHO would be a deal breaker license-wise if the ‘data distributed with MetaMap’ are strictly required for your tools to work. Please also see the DFSG guidelines [3]. MetaMap apparently also needs a commercial, non-free Prolog implementation to be built from source [4]. If this can’t be replaced by a free alternative then there’s no luck. Without this you can’t build your MetaMap from source, so it can’t go into Debian.
Cheers
Sascha
[1] https://packages.debian.org/sid/weka
[2] https://metamap.nlm.nih.gov/MMTnCs.shtml
[3] https://www.debian.org/social_contract.html#guidelines
[4] https://metamap.nlm.nih.gov/Sources.shtml
>
> -----Original Message-----
> From: Andreas Tille [mailto:andreas@an3as.eu]
> Sent: 19 July 2016 20:56
> To: debian-med@lists.debian.org
> Subject: Re: Table Information extraction tools in DebianMed?
>
> Hi Nikola,
>
> thanks for your interest in Debian Med.
>
> On Tue, Jul 19, 2016 at 12:15:34PM +0000, Nikola Milosevic wrote:
>>
>> I am a PhD student at the University of Manchester and during my PhD, I have developed a set of tools for processing, annotating and information extraction from tables in biomedical literature (currently supported XML formats of PubMedCentral and DailyMed). The tools can be found on GitHub and create a pipeline (one uses the output of the other):
>> - https://github.com/nikolamilosevic86/TableAnnotator
>> - https://github.com/nikolamilosevic86/Marvin
>> -
>> https://github.com/nikolamilosevic86/WIET---Wizard-for-Information-Ext
>> raction-from-Tables
>>
>> They are still under development, but are now quite stable and ready to be used. I would like to package these tools and include them into DebianMed distribution. I read the following page https://wiki.debian.org/DebianMed/MoM and I would be happy if someone can help me to start.
>
> Fine. You can easily get the next MoM slot with one of the programs of yours - preferably the most basic one. Please note: For Debian packaging you need to get rid of *all* binary code from your source tarball which are also *.jar files. I have not checked yet but usually it happens that you need to start with packaging some Java precondition.
>
> So feel free to add your name to the MoM table and start following the Debian Med policy to create an account on alioth.debian.org to gain commit permissions.
>
> Kind regards
>
> Andreas.
>
> --
> http://fam-tille.de
>
Reply to: