by po84 on 4/25/13, 12:11 PM with 0 comments
Say I want to get started building a processing pipeline using the Python scikit-learn package. I know I can probably poke around in Google or the scikit-learn.org site and find a starting example. I would prefer, however, to do what I would normally do using npm for Node modules, or brew for Mac recipies, or PackageManager for SublimeText packages: quickly search on a web site or command line for a relevant boilerplate. Not only would this cut down my time in finding a starting point, the community aspect of it (votes, # of downloads, most depended on, most modified, ...) would provide hints about the quality, something that's hard to deduce from examples cherry-picked from web search.
Concrete examples:
$ boilerplate search "scikit pipeline"
# results here
$ boilerplate fetch scikit-pipeline-boilerplate
# grabs it, unpacks it
$ boilerplate fetch h5bp
# grabs standard h5bp, unpacks it
# etc.