David Crossley wrote: > Robert Burrell Donkin wrote: > > David Crossley wrote: > > > > > > Data files are also generated so that other tools can easily > > > re-use the metadata for the set of projects currently in incubation. > > > > > > * clutch.txt - list of project names, one per line. > > > * Structured list of various details. Not yet happening > > > - need to decide which format (N3, DOAP, etc.) and which > > > attributes to be stored. > > > > RDFa has the advantage that the X(HT)ML document can be viewed > > adequately in a browser > > I had wondered about that too. Is it easy to generate > that output without the support of an rdf module? > Note that i am trying to keep the dependencies of > Clutch to an absolute minimum, so that any incubator > committer can run it. > > One approach that i have been considering is to dump > the data from clutch to a Python "pickle" file, then > run a separate process which connects to people.apache.org > (like you do with the "audit" scanner) where we could > have the necessary Python modules installed to easily > generate other data output files.
Now done the next step: Serialise certain attributes to a pickle file in incubator SVN at public/trunk/clutch.pkl Uses binary pickle which also stops unnecessary svn commit diff emails. A separate python script 'clutch2data.py' can utilise that data to create whatever xml data structure we decide. This de-couples the tasks of collecting the data and producing the special output. Not yet sure if that is the correct approach -David --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
