cozy-indexer

0.0.13 • Public • Published

Cozy Node Indexer

Overlay to search-index with support for stemming and internationalization using natural and franc.

API

    OPTIONS =
        docType:
            filter: true
        tags:
            filter: true
        title:
            nGramLength: 1,
            stemming: trueweight: 5fieldedSearch: true
        content:
            nGramLength: {gte: 1lte: 2},
            stemming: trueweight: 1fieldedSearch: false
 
    DOC = {
        "_id":"fake_id_note_eng",
        "docType": "Note",
        "title": "Hello world",
        "content": "This is an english note with some information.",
        "tags": ["tagA""tagged"]
    }
 
    indexer.addBatch [DOC]OPTIONScallback
 
    QUERY_OPTIONS = {
        query:
            "*": "cozy" # search cozy in all fields 
            "title":"indexing" # AND indexing in title 
        pageSize: 5 # 5 docs per page 
        offset: 2 # 2nd page 
        facets:
            docType: {}
        filter:
            date: [ # search something at the end of this or last month 
                ['2015-10-25''2015-10-30'],
                ['2015-11-25''2015-11-30'],
            ]
    }
 
    indexer.search QUERY_OPTIONS(err, results) ->
        results.totalHits === 16 # total number of matching docs 
        results.hits === [{
            {id: "docid"}
            {id: "docid2"}
            ...
        }# ids of the 5 docs in 2nd page 
        results.facets[0=== { key: 'docType'value: [
            { key: 'note'value: 3 },
            { key: 'file'value: 13 }
        } # there is 13 file and 3 note matching our query 

Used libraries limitations and effects

The following problems have been encoutered with search-index

  • No hook for tokenization : force us to .split(natural).join(separator) so that search-index can .split(separator)
  • No OR request : would allow to just stem text and perform a double query if we cant guess the language of the query
  • Fail when docs are not pure javascript object : incompatible with flatiron/cradle in the ds, investigate using the raw options of cradle

The following problems have been encoutered with natural

  • no german stemmer

What about the former indexer

This library is a replacement for it, the former indexer is not used anymore by the Data System and can be safely uninstalled.

Hack

git clone https://github.com/cozy/cozy-indexer.git
cd cozy-indexer
npm install

# make changes in ./src

npm run build

Tests

npm run test

BuildStatus

License

Cozy Data System is developed by Cozy Cloud and distributed under the AGPL v3 license.

What is Cozy?

Cozy Logo

Cozy is a platform that brings all your web services in the same private space. With it, your web apps and your devices can share data easily, providing you with a new experience. You can install Cozy on your own hardware where no one profiles you.

Community

You can reach the Cozy Community by:

  • Chatting with us on IRC #cozycloud on irc.freenode.net
  • Posting on our Forum
  • Posting issues on the Github repos
  • Mentioning us on Twitter

Versions

Current Tags

  • Version
    Downloads (Last 7 Days)
    • Tag
  • 0.0.13
    4
    • latest

Version History

Package Sidebar

Install

npm i cozy-indexer

Weekly Downloads

3

Version

0.0.13

License

none

Last publish

Collaborators

  • mycozycloud