cachecow

Cache Cow

All the caching modules I found on npm were for caching things in memory, or in an external service like Redis. I couldn't find any cache modules on npm that fit my need, so I wrote this one.

Motivation

• Do you have extra hard drive space?
• Store some of your files in the cloud or whatever?
• Want to save bandwidth or time by making fewer fetches to that cloud?

That is the use case this module is designed for. I wanted to use the hard drive to cache files that are in remote storage like Amazon S3.

Features

• When cache reaches 90% of the defined maxsize it deletes files until the cache is 70% of maxsize.
• Returns streams, not buffers, for faster requests.
• How it chooses which files to delete can be customized

Free support

Email me or create an issue if you need help! I'm throwing this up without a lot of documentation, and am very welcome to feedback and suggestions.

Example

# Hypothetical example using CacheCow with an HTTP proxy server
express = require('express')
request = require('request')
CacheCow = require('cachecow')

app = express()

cache = new CacheCow({
  maxsize: 1024*1024*1024*10 # 10GB
  # Define a function to deal with cache misses
  getter: (filename, callback) ->
    console.log "Not found in cache"
    stream = request("http://httpstat.us/#{filename}")
    callback(null, stream)
})

# Retrieve something from cache
app.get "/:filepath", (req, res) ->
  cache.get req.params.filepath, (err, stream) ->
    return res.status(404).send(err.message) if err?
    stream.pipe res

# This scans the cache directory to gather file size information for any files
# already present in the cache directory.
cache.init (err) ->
  console.log err if err?
  app.listen 8080, (err) ->
    console.log err if err?
    console.log "started"

Options

# Default values
cache = new CacheCow
  maxsize: 1024*1024*1 # 1MB
  dir: upath.join(os.tmpdir(), 'cachecow')
  getter: ->
    console.log 'You must define a getter function for cache misses!'
  # Weights used in garbage collection equation
  DeleteByFileSize: 0      # Delete big files first, keep as many files as possible
  DeleteByTotalReads: 0    # Delete rarely used files, keep file that are frequently fetched
  DeleteByLastAccess: 1    # Delete old files, keep files that have been recently fetched