Skip to content

Latest commit

 

History

History
62 lines (41 loc) · 1.72 KB

TODO.md

File metadata and controls

62 lines (41 loc) · 1.72 KB

TODO

doc noodlin hash. Samples:

# noodlin param gum.firstname=mark
# noodlin param gum.lastname=plaksin
# noodlin param gum.address.street='110 orchard knob ln'
# noodlin param gum.address.city='athens'
# noodlin param gum.address.state='ga'
# noodlin param gum.address.zipcode='30605'

Um, better hash sample in options. (gum?!)

  • Add /magic/ as an alias for /_/

  • Fix curl examples in README

  • Rubocop

  • Think about ES mappings and DTRT. ** Handle node (etc) names containing dashes and other separators that the ES indexer keys off of.

  • noo-alike

  • content-type

  • noodle display arguments to support:

  • values only
  • csv
  • sort
  • column -t

  • Make Noodle::Search idiom/sugar for finding a single node by name

  • fail more gracefully when ES isn't running

  • 'noodlin help' and 'noodle _ help'

  • Add versioning

  • Make uniqueness validators for :name in Node and Option.

  • Put text/plain output in column -t format

  • Pretty JSON format a la Elasticsearch

  • Set content type and return JSON by default. Support returning "pretty" output as text/plain and make it match what existing Noodle returns

  • Looks like ES limits simple search to 10000

Perf notes

  1. On a random machine It takes about 3 minutes to make 1000 nodes each with a basic set of facts. Slow!
  2. On the same random machine it takes about 10 seconds to dump 2000 nodes to json (noodle fqdn=~. json). Somewhat faster than "dirty-room" Noodle which takes close to a minute to dump 1500 nodes to json.
  3. And about 50 seconds to dump 10000 nodes to json.
  4. No doubt it can be faster!

Maybe

  • Put API under /api/v1, but have a convenience /_/ alias for searching