Point a gardener at couchdb and it will install, update, and run node processes that design-docs and couchapps depend on.
npm install gardener -g
gardener http://admin:pass@localhost:5984
This tells the gardener to watch all the databases on your local couch. But what does it do?
Simple. It looks for node_modules on design docs. Like this:
{
"id": "_design/twitter_things",
"views" : {
"map" : "by_date" : "function(doc) { emit(doc.date, null); }"
}
"node_modules" : "twitter-loader,worker-generate-thumbnails@0.4.6"
}
In this case, gardener will npm install twitter-loader and worker-generate-thumbnails and run them. Gardener passes them the couch url, and optional username and password. It will look something like this:
$ gardener http://localhost:5984
info: [gardener] polling couch for design doc changes.
info: [gardener] installing twitter-loader-1.0.1.tgz start_immediate=true, module_digest=md5-COu0+gC6Cvk+UPOB3yz6iQ==, module_name=twitter-loader-1.0.1.tgz, package_version=1.0.1, local_name=aHR0cDovL2xvY2FsaG9zdDo1OTg0L3Rlc3QvX2Rlc2lnbi9tdWNreW11Y2s=, ddoc_url=http://localhost:5984/test/_design/myapp, db_url=http://localhost:5984/test
npm http GET https://registry.npmjs.org/twitter-loader
npm http 200 https://registry.npmjs.org/twitter-loader/twitter-loader-1.0.1.tgz
npm WARN package.json twitter-loader@1.0.1 No README.md file found!
twitter-loader@1.0.1 installed_packages/node_modules/twitter-loader
info: [gardener] starting package [twitter-loader] in working_dir/aHR0cDovL2xvY2FsaG9zdDo1OTg0L3Rlc3QvX2Rlc2lnbi9tdWNreW11Y2s=
info: [twitter-loader] Twitter loader is dumping tweets into http://localhost:5984/test
Yay, twitter loader has been installed, run, and hooked up to the right db.
Create a directory called 'twitter-loader'. Add server.js to it that looks like:
// the gardener will set the COUCH_URL variable. Now do what you want!
var db_url = process.env.COUCH_URL || 'http://localhost:5984/test';
// do something and save to couch
load_all_of_twitter(db_url);
Add a package.json file. npm publish.
(When publishing try to use the couchdb-worker tag so it is easy for others to find and resuse).
Use some existing workers on npm: tagged couchdb-worker on npm
Here are just some ideas, you can extrapolate:
- use existing workers from hoodie or null2/jo
- Make a queue service
- RSS feed archiving
- download all the things
- Add a pixel-tracker and store in couch
- Add chained map reduce to couchdb with something like couch-incarnate
- and much more!
Usage: gardener <url> [OPTIONS]
Available OPTIONS:
--include, -i Comma seperated db and design docs to include.
--exclude, -e Comma seperated db and design docs to exclude.
--web, -w Turn on a local http server and add turn on the /_garden proxy to it from couch ("true" or "false", "true" by default)
--host, -h The host name (or IP address) to bind the http server to. (localhost by default)
--port, -p The port to bind the http server to. (integer, 25984 by default)
--upnp Use upnp to open the port on the firewall, and use the public ip ("true" or "false", "false" by default)
--user The username given to the node module for connecting to couch
--pass The password given to the node module for connecting to couch
--time, -t Polling interval (seconds) to check couch for changes to design docs (integer, 30 by default)
--stdinpass when specified, gardener will read a single line from stdin, which will be used as the password when connecting to couch
see examples for examples of commandline usage.
To reduce the footprint of gardener dependencies, not all are listed in the dependencies section. To use some features you will have to do an addition npm install.
- web, you need to
npm install tako - upnp, you need to
npm install nat-upnp - testing (or make or npm test)
npm install mocha
I am probably doing this wrong. Please let me know a better solution.
gardener follows semver-ftw. Dont think 1.0.0 means production ready yet. There were some breaking changes, so had to move up the major version.