start rewriting readability

2021-03-30 11:47:36 +01:00 · 2021-03-30 11:47:36 +01:00 · e5920259b6
commit e5920259b6
parent 8c44d2fc87
8 changed files with 238 additions and 69 deletions
--- a/src/htmlutil/query_test.go
+++ b/src/htmlutil/query_test.go
@ -0,0 +1,64 @@
+package htmlutil
+
+import (
+	"strings"
+	"testing"
+
+	"golang.org/x/net/html"
+)
+
+func TestQuery(t *testing.T) {
+	node, _ := html.Parse(strings.NewReader(`
+		<!DOCTYPE html>
+		<html lang="en">
+		<head>
+			<meta charset="UTF-8">
+			<title></title>
+		</head>
+		<body>
+			<div>
+				<p>test</p>
+			</div>
+		</body>
+		</html>
+	`))
+	nodes := Query(node, "p")
+	match := (
+		len(nodes) == 1 &&
+		nodes[0].Type == html.ElementNode &&
+		nodes[0].Data == "p")		
+	if !match {
+		t.Fatalf("incorrect match: %#v", nodes)
+	}
+}
+
+func TestQueryMulti(t *testing.T) {
+	node, _ := html.Parse(strings.NewReader(`
+		<!DOCTYPE html>
+		<html lang="en">
+		<head>
+			<meta charset="UTF-8">
+			<title></title>
+		</head>
+		<body>
+			<p>foo</p>
+			<div>
+				<p>bar</p>
+				<span>baz</span>
+			</div>
+		</body>
+		</html>
+	`))
+	nodes := Query(node, "p , span")
+	match := (
+		len(nodes) == 3 &&
+		nodes[0].Type == html.ElementNode && nodes[0].Data == "p" &&
+		nodes[1].Type == html.ElementNode && nodes[1].Data == "p" &&
+		nodes[2].Type == html.ElementNode && nodes[2].Data == "span")	
+	if !match {
+		for i, n := range nodes {
+			t.Logf("%d: %s", i, HTML(n))
+		}
+		t.Fatal("incorrect match")
+	}
+}