For help on pull and push run script/views (pull|push) without a database name.
USAGE
exit
end
if ARGV.length == 1
case command
when "pull"
puts <<-PULL
script/views pull my-database-name [id prefix (default is _design)]
I will automagically create a "views" directory in your current working directory if none exists.
Then I copy the design documents and views into a directory structure like:
./views/my-design-doc/view-name-map.js
./views/my-design-doc/view-name-reduce.js
If your view names don't end in "map" or "reduce" I'll add those suffixes as a pull. On push I'll put them in new locations corresponding to these new names (overwriting the old design documents). I'm opinionated, but if these conventions don't work for you, the source code is right here.
PULL
when "push"
puts <<-PUSH
script/views push my-database-name
I'll push all the files in your views directory to the specified database. Because CouchDB caches the results of view calculation by function content, there's no performance penalty for duplicating the map function twice, which I'll do if you have a reduce function. This makes it possible to browse the results of just the map, which can be useful for both queries and debugging.