package module
v0.0.0-...-6a20a75 Latest Latest

This package is not in the latest version of its module.

Go to latest
Published: Mar 17, 2019 License: Apache-2.0 Imports: 3 Imported by: 2


We are looking for contributors! Please check the ROADMAP to see how you can help ❤️

❤️ This is a brand new project, let's give it some star and watch to see how it will develop, the open source community is driven by you !

GraphQuery CircleCI Go Report Card Build Status Coverage Status GoDoc Gitter chat


GraphQuery is a query language and execution engine tied to any backend service. It is back-end language independent.

Related Projects:



GraphQuery is an easy to use query language, it has built-in Xpath/CSS/Regex/JSONpath selectors and enough built-in text processing functions.
The most amazing thing is that you can use the minimalist GraphQuery syntax to get any data structure you want.


Use GraphQuery to let you unify text parsing logic on any backend language.
You won't need to find implementations of Xpath/CSS/Regex/JSONpath selectors between different languages ​​and get familiar with their syntax or explore their compatibility.

Multiple selector syntax support

You can use GraphQuery to parse any text and use your skilled selector. GraphQuery currently supports the following selectors:

  1. Jsonpath for parsing JSON strings
  2. Xpath and CSS for parsing XML/HTML
  3. Regular expressions for parsing any text.

You can use these selectors in any combination in GraphQuery. The rich built-in selector provides great flexibility for your parsing.

Complete function

Graphquery has some built-in text processing functions like trim, template, replace. If you think these functions don't meet your needs, you can register new custom functions in the pipeline.

Clear data structure & Concise grammar

With GraphQuery, you won't need to look for parsing libraries when parsing text, nor do you need to write complex nesting and traversal. Simple and clear GraphQuery syntax gives you a clear picture of the data structure.


As you can see from the comparison chart above, the syntax of GraphQuery is so simple that even if you are in touch with it for the first time, you can still understand its meaning and get started quickly.

Mature error message

Whether it is a syntax error or a function call error, there will be detailed error messages for you to debug.


The above example is executed in GraphQuery-PlayGround. We can see that when an error occurs, GraphQuery can return detailed error information to help developers locate the wrong location.

At the same time, GraphQuery is also very easy to integrate into your backend data system (any backend language), let's continue to look down.

Getting Started

GraphQuery consists of query language and pipelines. To guide you through each of these components, we've written an example designed to illustrate the various pieces of GraphQuery. This example is not comprehensive, but it is designed to quickly introduce the core concepts of GraphQuery. The premise of the example is that we want to use GraphQuery to query for information about library books.

1. First example
<!-- Great book. -->
<book id="b0836217462" available="true">
    <title lang="en">Being a Dog Is a Full-Time Job</title>
    <quote>I'd dog paddle the deepest ocean.</quote>
    <author id="CMS">
        <?echo "go rocks"?>
        <name>Charles M Schulz</name>
    <character id="PP">
        <name>Peppermint Patty</name>
        <qualification>bold, brash and tomboyish</qualification>
    <character id="Snoopy">
        <qualification>extroverted beagle</qualification>

Faced with such a text structure, we naturally think of extracting the following data structure from the text :

    character [{

This is perfect, when you know the data structure you want to extract, you have actually succeeded 80%, the above is the data structure we want, we call it DDL (Data Definition Language) for the time being. let's see how GraphQuery does it:

    bookID `css("book");attr("id")`
    title `css("title")`
    isbn `xpath("//isbn")`
    quote `css("quote")`
    language `css("title");attr("lang")`
    author `css("author")` {
        name `css("name")`
        born `css("born")`
        dead `css("dead")`
    character `xpath("//character")` [{
        name `css("name")`
        born `css("born")`
        qualification `xpath("qualification")`

As you can see, the syntax of GraphQuery adds some strings wrapped in ` to the DDL. These strings wrapped by ` are called Pipeline. We will introduce Pipeline later. Let's first take a look at what data GraphQuery engine returns to us.

    "bookID": "b0836217462",
    "title": "Being a Dog Is a Full-Time Job",
    "isbn": "0836217462",
    "quote": "I'd dog paddle the deepest ocean.",
    "language": "en",
    "author": {
        "born": "1922-11-26",
        "dead": "2000-02-12",
        "name": "Charles M Schulz"
    "character": [
            "born": "1966-08-22",
            "name": "Peppermint Patty",
            "qualification": "bold, brash and tomboyish"
            "born": "1950-10-04",
            "name": "Snoopy",
            "qualification": "extroverted beagle"

Wow, it's wonderful. Just like what we want.
We call the above example Example1, now let's have a brief look at what pipeline is.

2. Pipeline

A pipeline is a collection of functions that use the parent element text as an entry parameter to execute the functions in the collection in sequence. For example, the language field in our previous example is defined as follows:

language `css("title");attr("lang")`

The language is the field name, css("title"); attr("lang") is the pipeline. In this pipeline, GraphQuery first uses the CSS selector to find the title node from the document, and the title node will be obtained. Pass the obtained node into the attr() function and get its lang attribute. The whole process is as follows:

In Example1, we not only use the css and attr functions, but also xpath(). It is easy to associate, Xpath() is to select elements with the Xpath selector. The following is a list of the pipeline functions built into the current version of graphquery:

pipeline prototype example introduce
css css(CSSSelector) css("title") Use CSS selector to select elements
json json(JSONSelector) json("title") Use json path to select elements
xpath xpath(XpathSelector) xpath("//title") Use Xpath selector to select elements
regex regex(RegexSelector) regex("") Use Regex selector to select elements
trim trim() trim() Clear spaces and line breaks before and after the string
template template(TemplateStr) template("[{$}]") Add characters before and after variables
attr attr(AttributeName) attr("lang") Extract the property of the current node
eq eq(Index) eq("0") Take the nth element in the current node collection
string string() string() Extract the current node native string
text text() text() Extract the text of the current node
link link(KeyName) link("title") Returns the current text of the specified key
replace replace(A, B) replace("a", "b") Replace all A in the current node to B
absolute absolute(A) absolute("") Absolute will take A as a reference and absoluteize the current text as a URL

More detailed introduction to pipeline and function, please go to docs.


GraphQuery is currently only native to Golang, but for other languages, it can be invoked as a service.

1. Golang:
go get -u

Create a new go file :

package main

import (


func main() {
	document := `
                <a href="01.html">Page 1</a>
                <a href="02.html">Page 2</a>
                <a href="03.html">Page 3</a>
	expr := "{ anchor `css(\"a\")` [ content `text()` ] }"
	response := graphquery.ParseFromString(document, expr)
	bytes, _ := json.Marshal(response.Data)

Run the go file, the output is as follows :

{"anchor":["Page 1","Page 2","Page 3"]}
2. Other language

We use the HTTP protocol to provide a cross-language solution for developers to query GraphQuery using any back-end language you want to use to access the specified port after starting the service.

GraphQuery-http : Cross language solution for GraphQuery

You can also use RPC for communication, but currently you may need to do this yourself, because the RPC project on GraphQuery is still under development.
At the same time, We welcome the contributors to write native support code for other languages ​​in GraphQuery.

Update Log

  1. Now graphquery.Response is the alias of kernel.Response
  2. Change the type of Response.Errors from []string to kernel.Errors, kernel.Errors implements some common interfaces for error and json
  3. Added three methods for kernel.Response
    3.1 MarshalData() (string, error): Output Response.Data as json string 3.2 JSON() string: Output Response as json string, It is now equivalent to String()
    3.3 Decode(obj interface{}): Map parsing results to a given data structure at a lower cost. You can use it like this:

type Anchor struct {
	Title string `json:"title"`
	URL   string `json:"url"`

func main() {
	document := `
        <a href="1.html">anchor 1</a>
        <a href="2.html">anchor 2</a>
        <a href="3.html">anchor 3</a>
	query := "a `css(\"a\")` [{ title `text();trim()` url  `attr(\"href\")` }]"
	response := graphquery.ParseFromString(document, query)       
    anchors := []*Anchor{}
    // Now you have converted the parsing results to []*Anchor

  1. Any questions in use, please feel free to issue :)




This section is empty.


This section is empty.


func Compile

func Compile(expr []byte) (*kernel.Graph, error)

Compile compiles an expression in the form of a byte array into a parser

func MustCompile

func MustCompile(expr []byte) *kernel.Graph

MustCompile is used to compile expressions, will panic when compile errors


type Response

type Response = kernel.GraphResponse

Response is an alias for kernel.GraphResponse

func ParseFromBytes

func ParseFromBytes(document string, expr []byte) (response *Response)

ParseFromBytes is used to parse documents in string format and expressions in []byte form errors will be recorded in the Errors of Response

func ParseFromString

func ParseFromString(document string, expr string) (response *Response)

ParseFromString will parse documents and expressions directly, and errors will be recorded in the Errors of Response


Path Synopsis

Jump to

Keyboard shortcuts

? : This menu
/ : Search site
f or F : Jump to
y or Y : Canonical URL