[Chaoss-software] [Meeting item] Collaboration of projects within the Software TC
D M German
dmg at turingmachine.org
Mon Oct 23 01:23:24 UTC 2017
Thanks Jesus, now I get it. Thanks for the long explanation.
Jesus> In short, we have a level, which is Perceval, which just extracts
Jesus> information from the APIs and produces JSON documents. Those are
Jesus> available directly from Perceval, as Python dictionaries, or from an
Jesus> ElasticSearch database, as JSON documents (that's what we call "raw
Jesus> indexes"). Other databases are perfectly possible, and easy to
Jesus> implement: just upload the documents to them. This layer tries to mimic
Jesus> as much as possible the original data source, so from a semantic point
Jesus> of view is not a "layer": it just provides more convenient availability
Jesus> of data.
So Perceval is the universal fact extractor. and it generates some json
files that then can be read by the next stage in the process. This is
quite nice because it makes the user not worry of the intrinsic format
of the data (and retrieval) but only of reading a JSON file (its
output). Very nice.
I guess this is a good model (inefficient in some cases, but flexible).
In MSR we never go to agree in data models. I doubt we will agree here
either. I suspect that at the end, we just need to know what data (and
in which format) each tool creates/generates and adjust our tools
accordingly.
I, for example, favour SQL (specially sqlite) as a method of
representation and manipulation.
--dmg
--
Daniel M. German "El pasado es una engañosa utopía
y nada fue tan grato o tan terrible
Germán Dehesa -> como lo recordamos en el presente."
http://turingmachine.org/
http://silvernegative.com/
dmg (at) uvic (dot) ca
replace (at) with @ and (dot) with .
More information about the Chaoss-software
mailing list