lucene

README

custom version

Added custom driver with clickhouse syntax

go-lucene

A lucene parser written in go with no dependencies.

With this package you can quickly integrate lucene style searching inside your app and generate sql filters for a particular query. There are no external dependencies and the grammar fully supports Apache Lucene 9.4.2.

Out of the box go-lucene support postgres compliant sql generation but it can be extended to support different flavors of sql (or no sql) as well.

Usage

// suppose you want a query for a red apple that is not a honey crisp and is younger than 5 months old
myQuery := `color:red AND (NOT type:"honey crisp" OR age_in_months:[5 TO *])`
expression, err := lucene.Parse(myQuery)
if err != nil {
    // handle error
}

filter, err := driver.NewPostgresDriver().Render(expression)
if err != nil {
    // handle error
}

SQLTemplate := `
SELECT *
FROM apples
WHERE %s
LIMIT 10;
`
sqlQuery := fmt.Sprintf(SQLTemplate, filter)

// sqlQuery is:
`
SELECT *
FROM apples
WHERE
    color = red
    AND (
      NOT(type = 'honey crisp')
      OR age_in_months >= 5
    )
LIMIT 10;
`

Extending with a custom driver

Just embed the Base driver in your custom driver and override the RenderFN's with your own custom rendering functions. Please contribute drivers back so others can use it too :).

import (
    "github.com/grindlemire/go-lucene/pkg/driver"
    "github.com/grindlemire/go-lucene/pkg/lucene/expr"
)

// MyDriver ...
type MyDriver struct {
	Base
}

// NewMyDriver ...
func NewMyDriver() MyDriver {
	fns := map[expr.Operator]driver.RenderFN{
        // suppose we have our own literal rendering function
		expr.Literal: myLiteral,
	}

	for op, sharedFN := range driver.Shared {
		_, found := fns[op]
		if !found {
			fns[op] = sharedFN
		}
	}

	return MyDriver{
		driver.Base{
			renderFNs: fns,
		},
	}
}

// myLiteral ...
func myLiteral(left, right string) (string, error) {
    // ....
}

Documentation

Index

• func Parse(input string, opts ...opt) (e *expr.Expression, err error)
• func TestPostgresSQLEndToEnd(t *testing.T)
• func WithDefaultField(field string) opt

Functions

func Parse

func Parse(input string, opts ...opt) (e *expr.Expression, err error)

Parse will parse using a buffer and the shift reduce algorithm. It scales rather well since it is a one pass algorithm with no backtracking.

func TestPostgresSQLEndToEnd

func TestPostgresSQLEndToEnd(t *testing.T)

func WithDefaultField

func WithDefaultField(field string) opt