Datascript query builder

This syntax is inspired by MongoDB query syntax

• pagination
• querying

This Query builder consists of the following:

• QueryBuilder
  • Datoms
  • Entity
  • Pull
  • Query
• Result

Datoms is used to access entity data directly (via index). We use it to find all ids for a given Entity class. Entity is used to access entity data directly. Pull can use the Pull one and Pull many APIs to retrieve specific Entity attribute data for one or more entities (by Id). Query is a general purpose query, used for finding Entity data that match specific criteria/constraints (ie. predicates etc.).

The QueryBuilder is a facade to all the different query variants.

• entities(params)
• byId(params)
• query(params, options)

Result is used to pre-process a query result from datascript/datomic DB. A result is typically in the form: [['kris', 32]] or [['name', 'kris', 'age', 32]] however, we would often like it in JSON map form, like {name: 'kris', 'age': 32}. Result is aimed to facilitate this enrichment and also pagination (limiting window/size of results returned) and sorting.

In package.json

  "dqb": "datascript-query-builder"

Alternatively clone this repo and npm link it locally when "hacking".

API usage:

import {QueryBuilder, Result} from 'dqb';

// first 10 entities older than 32 years, sorted by name
var query = {
  age: {$gt: 32},
  $limit: 10,
  $sort: {name: 1}
};

// Create Datalog, ie Datascript/Datomic query from FeatherJS query Object
// Search :person entities only...
var result = new QueryBuilder('person', query).build();

// prepare result using special pagination params ($limit, ...)
return new Result(result).build();

For contexts where $ is not commonly used, you can use alternative special key indicators, such as: .

var query = {
  age: {'.gt': 32},
  '.limit': 10,
  '.sort': {':item/name': 1}
};

You can also supply your pass regular expression for matching special keys, via options.specialKey and even a custom options.specialAttrib function to gain full power.

var qb = new QueryBuilder('person', options);

With custom special key detection.

var qb = new QueryBuilder('person', {
  // specialKey: /^(\$|\.)/,
  specialAttrib: (name) => {
    return name.match(/^\$/);
  }
});

The Datascript Query builder was initially built for use in feathers-datascript, a FeathersJS compatible DB driver/adapter.

Query :person entities matching criteria

// Query person entities matching criteria
var q1 = {age: {$gt: 32}, 'last-name': 'Johnson'};
var datalogQuery = qb.query(q1).build();
var result = conn.d.q(datalogQuery).then(qResult =>{
  return new Result(qResult).build();
});

Pull :person entity 27

// Pull :person entity 27
// Also see: http://augustl.com/blog/2013/find_by_id_in_datomic/
var personPull = qb.byId({id: 27}, options).build();
var params = {$skip: 20, $limit: 20};
var data = conn.d.pull(personPull).then(pulled => {
  return new Result(pulled, params).build();  
});

Fetch all :person entities via index

// Fetch all :person entities via index
// See: http://augustl.com/blog/2013/datomic_direct_index_lookup/
var datomsQuery = qb.datoms('id', options).build();
var entities = conn.d.datoms(datomsQuery).then(index => {
  var datoms = Entity.unpack(index);
  return conn.d.entity(datoms);
});

Get entity by lookup ref

// Get entity by lookup ref: email
var lookupRef = qb.entity({email: '[email protected]'}, options).build();
var person = conn.d.entity(lookupRef).then(entity => {
  return entity;
});

This is still early alpha. Please help out ;)

Write the code in src. Run npm run build to compile the source into a distributable format.

Put unit tests in test/unit. The npm test command runs the tests using Node.

• npm test - Lint the library and tests, then run the unit tests
• npm run lint - Lint the source and unit tests
• npm run watch - Continuously run the unit tests as you make changes to the source and test files themselves
• npm run test-browser - Build the library for use with the browser spec runner. Changes to the source will cause the runner to automatically refresh.
• npm run build - Lint then build the library
• npm run coverage - Generate a coverage report

The browser spec runner can be opened in a browser to run your tests. For it to work, you must first run npm run test-browser. This will set up a watch task that will automatically refresh the tests when your scripts, or the tests, change.