it.skrape.selects.CssSelectable.kt Maven / Gradle / Ivy

Go to download

Show more of this group Show more artifacts with this name
Show all versions of skrapeit-html-parser Show documentation

A Kotlin-based testing/scraping/parsing library providing the ability to analyze and extract data from HTML (server & client-side rendered). It places particular emphasis on ease of use and a high level of readability by providing an intuitive DSL. First and foremost it aims to be a testing lib, but it can also be used to scrape websites in a convenient fashion.

There is a newer version: 1.3.0-alpha.2

Show newest version

package it.skrape.selects

import it.skrape.SkrapeItDsl
import org.jsoup.nodes.Element
import java.util.*

@Suppress("TooManyFunctions")
@SkrapeItDsl
public abstract class CssSelectable {
    public abstract val toCssSelector: String

    internal abstract fun applySelector(rawCssSelector: String): List

    public fun  selection(cssSelector: String, init: CssSelector.() -> T): T =
            CssSelector(rawCssSelector = cssSelector, doc = this).init()

    /**
     * Will create a CssSelector scope to calculate a css selector
     * @param init block for configuring the CSS-Selector that will be considered during calculation
     * @return T
     */
    public operator fun  String.invoke(init: CssSelector.() -> T): T =
            [email protected](this, init)

    public open fun makeDefault(cssSelector: String): DocElement {
        return DocElement(Element("${UUID.randomUUID()}"))
    }

    /**
     * Will pick all occurrences of elements that are matching the CSS-Selector
     * @see Overview of CSS-Selectors for further information.
     * @param cssSelector that represents an CSS-Selector
     * @return T
     */
    public infix fun findAll(cssSelector: String): List =
            this.applySelector(cssSelector)

    public fun findByIndex(index: Int, cssSelector: String = ""): DocElement =
            findAll(cssSelector).getOrElse(index) { makeDefault(cssSelector) }

    public operator fun Int.invoke(cssSelector: String = ""): DocElement =
            findByIndex(this, cssSelector)

    public fun findBySelectorMatching(regex: Regex): List =
            [email protected]("*").filter { it.ownCssSelector.matches(regex) }

    public operator fun Regex.invoke(): List =
            findBySelectorMatching(this)

    /**
     * Will pick the first occurrence of an element that
     * is matching the CSS-Selector from a parsed document and invoke it to a lambda function.
     * @see Overview of CSS-Selectors for further information.
     * @param cssSelector that represents an CSS-Selector
     * @return T
     */
    public infix fun findFirst(cssSelector: String): DocElement =
            findByIndex(0, cssSelector)

    public fun findSecond(cssSelector: String = ""): DocElement =
            findByIndex(1, cssSelector)

    public fun findThird(cssSelector: String = ""): DocElement =
            findByIndex(2, cssSelector)

    public fun findLast(cssSelector: String = ""): DocElement =
            findAll(cssSelector).last()

    public fun findSecondLast(cssSelector: String = ""): DocElement =
            findAll(cssSelector).let { it.getOrElse(it.lastIndex -1) { makeDefault(cssSelector) } }

    /**
     * Will pick all occurrences of elements that are matching the CSS-Selector
     * @see Overview of CSS-Selectors for further information.
     * @param cssSelector that represents an CSS-Selector
     * @return T
     */
    public fun  findAll(cssSelector: String = "", init: List.() -> T): T =
            findAll(cssSelector).init()

    public fun  findByIndex(index: Int, cssSelector: String = "", init: DocElement.() -> T): T =
            findByIndex(index, cssSelector).init()

    public operator fun  Int.invoke(cssSelector: String = "", init: DocElement.() -> T): T =
            this(cssSelector).init()

    public fun  findBySelectorMatching(regex: Regex, init: List.() -> T): T =
            findBySelectorMatching(regex).init()

    public operator fun  Regex.invoke(init: List.() -> T): T =
            this().init()

    /**
     * Will pick the first occurrence of an element that
     * is matching the CSS-Selector from a parsed document and invoke it to a lambda function.
     * @see Overview of CSS-Selectors for further information.
     * @param cssSelector that represents an CSS-Selector
     * @return T
     */
    public fun  findFirst(cssSelector: String = "", init: DocElement.() -> T): T =
            findFirst(cssSelector).init()

    public fun  findSecond(cssSelector: String = "", init: DocElement.() -> T): T =
            findSecond(cssSelector).init()

    public fun  findThird(cssSelector: String = "", init: DocElement.() -> T): T =
            findThird(cssSelector).init()

    public fun  findLast(cssSelector: String = "", init: DocElement.() -> T): T =
            findLast(cssSelector).init()

    public fun  findSecondLast(cssSelector: String = "", init: DocElement.() -> T): T =
            findSecondLast(cssSelector).init()
}