Half finished selectors

Author: udan-jayanith

benchmarks/benchmark_test.go

@@ -6,7 +6,9 @@ import(
 	"net/http"
 	"time"
 )
-
+/*
+Adapted from [GoQuery example](https://github.com/PuerkitoBio/goquery?tab=readme-ov-file#examples)
+*/
 func TestFetchPostCovers(t *testing.T){
 	res, err := http.Get("https://www.metalsucks.net/")
 	if err != nil {

node-tree.go

@@ -219,18 +219,7 @@ But this is not the case for QuerySearch, QuerySelector and QuerySelectorAll.
 */
 // Adapted from [https://developer.mozilla.org/en-US/docs/Web/API/Element/closest](MDN Element: closest() method)
 func (node *Node) Closest(query string) *Node {
-	queryTokens := TokenizeQuery(query)
 	traverser := NewTraverser(node)
-	for traverser.GetCurrentNode() != nil {
-		if matchQueryTokens(traverser.GetCurrentNode(), queryTokens) {
-			break
-		}
-
-		if traverser.GetCurrentNode().GetPreviousNode() == nil {
-			traverser.SetCurrentNodeTo(traverser.GetCurrentNode().GetParent())
-		}else{
-			traverser.Previous()
-		}
-	}
+	
 	return traverser.GetCurrentNode()
 }

node-tree_test.go

@@ -114,6 +114,7 @@ func TestRemoveNode(t *testing.T){
 	//t.Log(GoHtml.NodeTreeToHTML(article))
 }
 
+/*
 func TestClosest(t *testing.T){
 	node, err := testFile4NodeTree()
 	if err != nil{
@@ -132,4 +133,5 @@ func TestClosest(t *testing.T){
 	}
 
 
-}
+}
+*/

querying.go

@@ -1,7 +1,7 @@
 package GoHtml
 
 import (
-	"iter"
+	//"iter"
 	"strings"
 )
 
@@ -103,6 +103,7 @@ func (node *Node) GetElementsById(idName string) NodeList {
 /*
 QuerySearch tokenizes the query string and search for nodes that matches with the right most query token. After matching right most query it proceeds to match nodes parents nodes for left over tokens and then passed that node to (yield/range). QuerySearch search the whole node tree for matches unless yield get canceled or range iterator get cancel.
 */
+/*
 func QuerySearch(node *Node, query string) iter.Seq[*Node] {
 	traverser := NewTraverser(node)
 	return func(yield func(node *Node) bool) {
@@ -180,3 +181,5 @@ func (node *Node) QuerySelectorAll(query string) NodeList {
 	}
 	return nodeList
 }
+
+*/

querying_test.go

@@ -138,6 +138,7 @@ func TestGetElementsById(t *testing.T) {
 	}
 }
 
+/*
 func TestSelectorTokenizer(t *testing.T) {
 	stack := linkedliststack.New()
 	stack.Push("article .content")
@@ -209,3 +210,5 @@ func TestQuerySelectorAll(t *testing.T) {
 		}
 	}
 }
+
+*/

selectors.go

@@ -0,0 +1,175 @@
+package GoHtml
+
+import (
+	"strings"
+)
+
+type BasicSelector int
+
+const (
+	Id BasicSelector = iota
+	Class
+	Tag
+)
+
+type Selector struct {
+	selector     string
+	selectorName string
+	selectorType BasicSelector
+}
+
+func matchNode(node *Node, basicSelectorName string, basicSelectorType BasicSelector) bool {
+	if basicSelectorName == ""{
+		return true
+	}else if node == nil {
+		return false
+	}
+
+	switch basicSelectorType {
+	case Id:
+		idName, _ := node.GetAttribute("id")
+		return idName == basicSelectorName
+	case Class:
+		classList := NewClassList()
+		classList.DecodeFrom(node)
+		return classList.Contains(basicSelectorName)
+	case Tag:
+		return node.GetTagName() == basicSelectorName
+	}
+	return false
+}
+
+func NewSelector(selector string) Selector {
+	selector = strings.TrimSpace(selector)
+	selectorStruct := Selector{}
+	if len(selector) == 0 || (selector[0] == '.' || selector[0] == '#') && len(selector) <= 1 {
+		return selectorStruct
+	}
+
+	switch selector[0] {
+	case '.':
+		selectorStruct.selectorType = Class
+	case '#':
+		selectorStruct.selectorType = Id
+	default:
+		selectorStruct.selectorType = Tag
+	}
+
+	selectorStruct.selector = selector
+	if selectorStruct.selectorType != Tag {
+		selectorStruct.selectorName = selector[1:]
+	} else {
+		selectorStruct.selectorName = selector
+	}
+	return selectorStruct
+}
+
+type Combinator int
+
+const (
+	Descendant Combinator = iota
+	Child
+	NextSibling
+	SubsequentSibling
+	//if no combinator
+	NoneCombinator
+)
+
+type CombinatorEl struct {
+	Type      Combinator
+	Selector1 Selector
+	Selector2 Selector
+}
+
+func TokenizeSelectorsAndCombinators(selector string) []CombinatorEl {
+	list := make([]CombinatorEl, 0, 1)
+	slice := strings.SplitSeq(selector, " ")
+	currentCombinator := *new(CombinatorEl)
+	currentCombinator.Selector1 = NewSelector("")
+	for str := range slice {
+		if strings.TrimSpace(str) == "" {
+			continue
+		}
+
+		switch str {
+		case "+":
+			currentCombinator.Type = NextSibling
+		case ">":
+			currentCombinator.Type = Child
+		case "~":
+			currentCombinator.Type = SubsequentSibling
+		default:
+			newSelector := NewSelector(str)
+			currentCombinator.Selector2 = newSelector
+			list = append(list, currentCombinator)
+			currentCombinator = *new(CombinatorEl)
+			currentCombinator.Selector1 = newSelector
+		}
+
+	}
+
+	if len(list) == 1 {
+		list[0].Type = NoneCombinator
+	}
+
+	return list
+}
+
+func (ce *CombinatorEl) IsMatchingNode(node *Node) bool {
+	switch ce.Type {
+	case Descendant:
+		return ce.isDescended(node)
+	case Child:
+		return ce.isDirectChild(node)
+	case NextSibling:
+		return ce.isNextSibling(node)
+	case SubsequentSibling:
+		return ce.isSubsequentSibling(node)
+	case NoneCombinator:
+		return matchNode(node, ce.Selector2.selectorName, ce.Selector2.selectorType)
+	}
+	return false
+}
+
+// isDescended returns wether the given node is a ce.Selector2 and descended of ce.Selector1.
+func (ce *CombinatorEl) isDescended(node *Node) bool {
+	if !matchNode(node, ce.Selector2.selectorName, ce.Selector2.selectorType) {
+		return false
+	}
+
+	parentNode := node.GetParent()
+	for parentNode != nil && !matchNode(parentNode, ce.Selector1.selectorName, ce.Selector1.selectorType) {
+		parentNode = parentNode.GetParent()
+	}
+	return parentNode != nil
+}
+
+// isDirectChild returns whether the given node is a direct child of ce.Selector1 and node is of ce.Selector2
+func (ce *CombinatorEl) isDirectChild(node *Node) bool {
+	if node == nil {
+		return false
+	}
+
+	return matchNode(node, ce.Selector2.selectorName, ce.Selector2.selectorType) && matchNode(node.GetParent(), ce.Selector1.selectorName, ce.Selector1.selectorType)
+}
+
+// isNextSibling return whether the given node is of ce.Selector2 and next sibling of ce.Selector1
+func (ce *CombinatorEl) isNextSibling(node *Node) bool {
+	if node == nil {
+		return false
+	}
+
+	return matchNode(node, ce.Selector2.selectorName, ce.Selector2.selectorType) && matchNode(node.GetPreviousNode(), ce.Selector1.selectorName, ce.Selector1.selectorType)
+}
+
+func (ce *CombinatorEl) isSubsequentSibling(node *Node) bool {
+	if !matchNode(node, ce.Selector2.selector, ce.Selector2.selectorType) {
+		return false
+	}
+
+	traverser := NewTraverser(node)
+	for traverser.GetCurrentNode() != nil && !matchNode(traverser.GetCurrentNode(), ce.Selector1.selector, ce.Selector1.selectorType) {
+		traverser.Previous()
+	}
+	return matchNode(traverser.GetCurrentNode(), ce.Selector1.selector, ce.Selector1.selectorType)
+}

selectors_test.go

@@ -0,0 +1,13 @@
+package GoHtml_test
+
+import(
+	"testing"
+	"github.com/udan-jayanith/GoHTML"
+)
+
+func TestTokenizeSelector(t *testing.T){
+	slice := GoHtml.TokenizeSelectorsAndCombinators(".class-1 > .class-2 + .class-3 a")
+	for _, el := range slice{
+		t.Log(el)
+	}
+}

tokenizer.go

@@ -8,7 +8,6 @@ import (
 	"golang.org/x/net/html"
 )
 
-
 // Tokenizer contains a *html.Tokenizer.
 type Tokenizer struct {
 	z *html.Tokenizer
@@ -26,7 +25,7 @@ func (t *Tokenizer) Advanced() html.TokenType {
 	return t.z.Next()
 }
 
-// CurrentNode returns the current node. 
+// CurrentNode returns the current node.
 // Returned value can be nil regardless of tt.
 func (t *Tokenizer) GetCurrentNode() *Node {
 	currentToken := t.z.Token()
@@ -85,7 +84,7 @@ func (ntb *NodeTreeBuilder) WriteNodeTree(node *Node, tt html.TokenType) {
 		if node == nil {
 			return
 		}
-		
+
 		if isTopNode(ntb.currentNode, ntb.stack) {
 			ntb.currentNode.AppendChild(node)
 		} else {
@@ -120,83 +119,4 @@ func isTopNode(node *Node, stack *linkedliststack.Stack) bool {
 
 	topNode := val.(*Node)
 	return topNode == node
-}
-
-// QueryToken types
-const (
-	Id int = iota
-	Tag
-	Class
-)
-
-// QueryToken store data about basic css selectors(ids, classes, tags).
-type QueryToken struct {
-	Type         int
-	SelectorName string
-	Selector     string
-}
-
- /*
-TokenizeQuery tokenizes the query and returns a list of QueryToken.
-
-query should be of only consists of class, tag and/or id. This applies to every function that accepts a parameter name query.
-query should not consists of css selectors, Combinators and separators.
-*/
-func TokenizeQuery(query string) []QueryToken {
-	slice := make([]QueryToken, 0, 1)
-	if strings.TrimSpace(query) == "" {
-		return slice
-	}
-
-	iter := strings.SplitSeq(query, " ")
-	for sec := range iter {
-		token := QueryToken{}
-		switch sec {
-		case "", " ", ".", "#":
-			continue
-		}
-
-		switch string(sec[0]) {
-		case ".":
-			token.Type = Class
-			token.SelectorName = sec[1:]
-		case "#":
-			token.Type = Id
-			token.SelectorName = sec[1:]
-		default:
-			token.Type = Tag
-			token.SelectorName = sec
-		}
-		token.Selector = sec
-		slice = append(slice, token)
-	}
-
-	return slice
-}
-
-// matchQueryTokens returns wether the queryTokens match given the node. 
-func matchQueryTokens(node *Node, queryTokens []QueryToken) bool {
-	if len(queryTokens) == 0 {
-		return false
-	}
-	classList := NewClassList()
-	classList.DecodeFrom(node)
-	for _, token := range queryTokens {
-		switch token.Type {
-		case Id:
-			idName, _ := node.GetAttribute("id")
-			if token.SelectorName != idName {
-				return false
-			}
-		case Tag:
-			if node.GetTagName() != token.SelectorName {
-				return false
-			}
-		case Class:
-			if !classList.Contains(token.SelectorName) {
-				return false
-			}
-		}
-	}
-	return true
 }