reds

reds is a light-weight Redis search for node.js. This module was originally developed to provide search capabilities for Kue a priority job queue, however it is very much a light general purpose search library that could be integrated into a blog, a documentation server, etc.

Upgrading

Version 1.0.0 is syntactically compatible with previous versions of reds (0.2.5). However, natural has been updated. Documents indexed with older installs of reds (using natural v0.2.0) may need to be re-indexed to avoid some edge cases.

Installation

  $ npm install reds

Example

The first thing you'll want to do is create a Search instance, which allows you to pass a key, used for namespacing within Redis so that you may have several searches in the same db. You may specify your own node_redis instance with the reds.setClient function.

var search = reds.createSearch('pets');

reds acts against arbitrary numeric or string based ids, so you could utilize this library with essentially anything you wish, even combining data stores. The following example just uses an array for our "database", containing some strings, which we add to reds by calling Search#index() padding the body of text and an id of some kind, in this case the index.

1var strs = [];
2strs.push('Tobi wants four dollars');
3strs.push('Tobi only wants $4');
4strs.push('Loki is really fat');
5strs.push('Loki, Jane, and Tobi are ferrets');
6strs.push('Manny is a cat');
7strs.push('Luna is a cat');
8strs.push('Mustachio is a cat');
9
10strs.forEach(function(str, i){ search.index(str, i); });

To perform a query against reds simply invoke Search#query() with a string, and pass a callback, which receives an array of ids when present, or an empty array otherwise.

1search
2  .query(query = 'Tobi dollars', function(err, ids){
3    if (err) throw err;
4    console.log('Search results for "%s":', query);
5    ids.forEach(function(id){
6      console.log('  - %s', strs[id]);
7    });
8    process.exit();
9  });

By default reds performs an intersection of the search words. The previous example would yield the following output since only one string contains both "Tobi" and "dollars":

Search results for "Tobi dollars":
  - Tobi wants four dollars

We can tweak reds to perform a union by passing either "union" or "or" to Search#type() in reds.search() between Search#query() and Search#end(), indicating that any of the constants computed may be present for the id to match.

1search
2  .query(query = 'tobi dollars')
3  .type('or')
4  .end(function(err, ids){
5    if (err) throw err;
6    console.log('Search results for "%s":', query);
7    ids.forEach(function(id){
8      console.log('  - %s', strs[id]);
9    });
10    process.exit();
11  });

The union search would yield the following since three strings contain either "Tobi" or "dollars":

Search results for "tobi dollars":
  - Tobi wants four dollars
  - Tobi only wants $4
  - Loki, Jane, and Tobi are ferrets

API

1reds.createSearch(key)
2Search#index(text, id[, fn])
3Search#remove(id[, fn]);
4Search#query(text[, type][, fn]); // will return a `Query` instance
5
6Query#between(start, stop)
7Query#type(type)
8Query#end(fn)

Examples:

1var search = reds.createSearch('misc');
2search.index('Foo bar baz', 'abc');
3search.index('Foo bar', 'bcd');
4search.remove('bcd');
5search.query('foo bar').end(function(err, ids){});

Extending reds

Starting in 1.0.0, you can easily extend and expand how reds functions. When creating a new search, supply an object as the second argument. There are currently three properties that can be configured:

• nlpProcess the natural language processing function. You can alter how the words are processed (split, stemmed, and converted to metaphones) using this function.
• writeIndex how the items are written to the index.
• removeIndex how the items are removed from the index.

See the lib/reds.js file for the implementation of each. Please keep in mind that changing these functions may invalidate your previously stored index.

1reds.createSearch('pets', {
2  nlpProcess  : yourNlpProcessingFunction,
3  writeIndex  : yourWriteIndexFunction,
4  removeIndex : yourRemoveIndexFunction
5});

About

Currently reds strips stop words and applies the metaphone and porter stemmer algorithms to the remaining words before mapping the constants in Redis sets. For example the following text:

Tobi is a ferret and he only wants four dollars

Converts to the following constant map:

1{
2  Tobi: 'TB',
3  ferret: 'FRT',
4  wants: 'WNTS',
5  four: 'FR',
6  dollars: 'DLRS'
7}

This also means that phonetically similar words will match, for example "stefen", "stephen", "steven" and "stefan" all resolve to the constant "STFN". Reds takes this further and applies the porter stemming algorithm to "stem" words, for example "counts", and "counting" become "count".

Consider we have the following bodies of text:

Tobi really wants four dollars
For some reason tobi is always wanting four dollars

The following search query will then match both of these bodies, and "wanting", and "wants" both reduce to "want".

tobi wants four dollars

Benchmarks

Nothing scientific but preliminary benchmarks show that a small 1.6kb body of text is currently indexed in ~6ms, or 163 ops/s. Medium bodies such as 40kb operate around 6 ops/s, or 166ms.

Querying with a multi-word phrase, and an index containing ~3500 words operates around 5300 ops/s. Not too bad.

If working with massive documents, you may want to consider adding a "keywords" field, and simply indexing it's value instead of multi-megabyte documents.

License

(The MIT License)

Copyright (c) 2011 TJ Holowaychuk <tj@vision-media.ca>

Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the 'Software'), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions:

The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software.

THE SOFTWARE IS PROVIDED 'AS IS', WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.