opts configures the two persistent datastores. Possible key-value pairs are:

• location (String, required) - Base directory in which both datastores will be kept.
• keyField (String, required) - JSONpath specifying the field in each document to be used as a key in the document store.
• indexMap (Array | Object, required) - JSONpaths specifying the fields in each document to index for fulltext searching. See index mapping below for details.
• deletable (Boolean, optional) - Whether documents should be deletable. Setting to true increases index size. Default: false.
• ngramLength (Integer | Array, optional) - ngram length(s) to use when building index.

It is quite rare that all fields in a database should be exposed to the user search. More often, we want to allow the user to search certain fields, but retrieve the full document for each result. The indexMap option allows you to specify how to index documents.

There are two ways to tell yuno how to index:

The simple option - an array of fields to index. The contents of each field will be passed through the default Natural Language Processing pipeline before being added to the search index.

To fine-tune the processing on a per-field basis, pass an Object where each key is a field to index. Values can be one of:

• true/false whether to apply the default NLP pipeline
• function a custom processing function.

Custom processing take the field value as a single argument, and their return value (either a string or an array) will be tokenised and added to the index.

db.add(documents, options, callback)

• documents, array of JSON-able objects to store
• options optional, can override the database-wide indexMap option
• callback, function to call on completion, with a single argument to be passed an error if there was one

e.g.

or using a custom indexMap:

### Delete documents

db.del(documents, callback)

• documents, document (object), id (string), or array of documents or ids
• callback, function to call on completion, with a single argument to be passed an error if there was one

e.g.

db.search(query, opts, callback)

Returns a cursor that can be used to page through the results. By default the pageSize is 50.

• query, string search query
• opts, (optional) options object
• callback, function to call on completion. Takes two arguments:
  • err error or null
  • results object containing the result metadata and hits

e.g.

yuno has a minimal command-line interface that can be used to create a database from a file containing JSON objects.

Install the CLI:

Create a new database:

The JSON data file must contain JSON objects, rather than an array. For example:

You can provide database options as a JSON file using the --opts argument:

Where the options JSON looks like:

yuno is being built to serve my use-case of embedding pre-made databases in electron apps. If you have another use-case and would like features added, please open an issue to discuss it - I'm happy to add things that will be widely useful.

Contributions are very welcome. Please open an issue to discuss any changes you would like to PR, or mention in an existing issue that you plan to work on it.

Ideas for improving performance are particularly welcome.

https://creativecommons.org/publicdomain/zero/1.0/

yuno is public domain code. Do whatever you want with it. Credit would be appreciated, but it's not required.