🔄 Sync from monorepo

wearemojo/mojo@c6b8b60
wearemojo · Mar 20, 2024 · 17e8524 · 17e8524
1 parent 25b41a8
commit 17e8524
Show file tree

Hide file tree

Showing 3 changed files with 95 additions and 11 deletions.
diff --git a/lib/discourse/discourse.go b/lib/discourse/discourse.go
@@ -2,6 +2,7 @@ package discourse
 
 import (
 	"net/http"
+	"net/url"
 	"time"
 
 	"github.com/wearemojo/mojo-public-go/lib/httpclient"
@@ -12,12 +13,12 @@ import (
 const ErrEmptyParam = merr.Code("empty_param")
 
 type Client struct {
-	BaseURL string
+	BaseURL *url.URL
 
 	apiKey string
 }
 
-func NewClient(baseURL, apiKey string) *Client {
+func NewClient(baseURL *url.URL, apiKey string) *Client {
 	return &Client{
 		BaseURL: baseURL,
 
@@ -32,7 +33,7 @@ type IdentifiedClient struct {
 func (c *Client) identifiedClient(header http.Header) *IdentifiedClient {
 	return &IdentifiedClient{
 		client: jsonclient.NewClient(
-			c.BaseURL,
+			c.BaseURL.String(),
 			httpclient.NewClient(10*time.Second, roundTripper{header}),
 		),
 	}

diff --git a/lib/discourseemoji/html_img.go → lib/discourseemoji/unmunge.go b/lib/discourseemoji/html_img.go → lib/discourseemoji/unmunge.go
@@ -1,14 +1,15 @@
 package discourseemoji
 
 import (
+	"net/url"
 	"slices"
 	"strings"
 
 	"github.com/wearemojo/mojo-public-go/lib/slicefn"
 	"golang.org/x/net/html"
 )
 
-func ReplaceHTMLImagesWithEmojis(src string) (string, error) {
+func UnmungeCookedHTML(src string, baseURL *url.URL) (string, error) {
 	doc, err := html.Parse(strings.NewReader(src))
 	if err != nil {
 		return "", err
@@ -17,9 +18,17 @@ func ReplaceHTMLImagesWithEmojis(src string) (string, error) {
 	var body *html.Node
 	var fn func(*html.Node) *html.Node
 	fn = func(node *html.Node) *html.Node {
-		if node.Type == html.ElementNode && node.Data == "body" {
+		switch {
+		case node.Type == html.ElementNode && node.Data == "body":
 			body = node
-		} else if node.Type == html.ElementNode && node.Data == "img" {
+		case node.Type == html.ElementNode && node.Data == "a" && baseURL != nil:
+			if idx := slicefn.FindIndex(node.Attr, func(a html.Attribute) bool { return a.Key == "href" }); idx != -1 {
+				if url, _ := url.Parse(node.Attr[idx].Val); url != nil {
+					url = baseURL.ResolveReference(url)
+					node.Attr[idx].Val = url.String()
+				}
+			}
+		case node.Type == html.ElementNode && node.Data == "img":
 			class, ok1 := slicefn.Find(node.Attr, func(a html.Attribute) bool { return a.Key == "class" })
 			alt, ok2 := slicefn.Find(node.Attr, func(a html.Attribute) bool { return a.Key == "alt" })
 			shortcode, ok3 := strings.CutPrefix(alt.Val, ":")

diff --git a/lib/discourseemoji/html_img_test.go → lib/discourseemoji/unmunge_test.go b/lib/discourseemoji/html_img_test.go → lib/discourseemoji/unmunge_test.go
@@ -1,40 +1,55 @@
 package discourseemoji
 
 import (
+	"net/url"
 	"strings"
 	"testing"
 
 	"github.com/matryer/is"
 )
 
-func TestReplaceHTMLImagesWithEmojis(t *testing.T) {
+func urlMustParse(s string) *url.URL {
+	u, err := url.Parse(s)
+	if err != nil {
+		panic(err)
+	}
+	return u
+}
+
+func TestUnmungeCookedHTML(t *testing.T) {
 	tests := []struct {
-		name     string
-		input    string
-		expected string
+		name         string
+		inputBaseURL *url.URL
+		inputSource  string
+		expected     string
 	}{
 		{
 			"upside_down_face",
+			nil,
 			`<img src="https://emoji.discourse-cdn.com/twitter/upside_down_face.png?v=12" title=":upside_down_face:" class="emoji" alt=":upside_down_face:" loading="lazy" width="20" height="20">`,
 			"🙃",
 		},
 		{
 			"only-emoji end",
+			nil,
 			`<img src="https://emoji.discourse-cdn.com/twitter/upside_down_face.png?v=12" title=":upside_down_face:" class="emoji only-emoji" alt=":upside_down_face:" loading="lazy" width="20" height="20">`,
 			"🙃",
 		},
 		{
 			"only-emoji start",
+			nil,
 			`<img src="https://emoji.discourse-cdn.com/twitter/upside_down_face.png?v=12" title=":upside_down_face:" class="only-emoji emoji" alt=":upside_down_face:" loading="lazy" width="20" height="20">`,
 			"🙃",
 		},
 		{
 			"only-only-emoji",
+			nil,
 			`<img src="https://emoji.discourse-cdn.com/twitter/upside_down_face.png?v=12" title=":upside_down_face:" class="only-emoji" alt=":upside_down_face:" loading="lazy" width="20" height="20">`,
 			`<img src="https://emoji.discourse-cdn.com/twitter/upside_down_face.png?v=12" title=":upside_down_face:" class="only-emoji" alt=":upside_down_face:" loading="lazy" width="20" height="20"/>`,
 		},
 		{
 			"two paragraphs",
+			nil,
 			strings.TrimSpace(`
 				<p>
 					<img src="https://emoji.discourse-cdn.com/twitter/upside_down_face.png?v=12" title=":upside_down_face:" class="emoji" alt=":upside_down_face:" loading="lazy" width="20" height="20">
@@ -76,6 +91,7 @@ func TestReplaceHTMLImagesWithEmojis(t *testing.T) {
 		},
 		{
 			"unrecognized",
+			nil,
 			strings.TrimSpace(`
 				<p>
 					<img src="https://emoji.discourse-cdn.com/twitter/upside_down_face.png?v=12" title=":upside_down_face:" class="emoji" alt=":upside_down_face:" loading="lazy" width="20" height="20">
@@ -91,13 +107,71 @@ func TestReplaceHTMLImagesWithEmojis(t *testing.T) {
 				</p>
 			`),
 		},
+		{
+			"urls",
+			nil,
+			strings.TrimSpace(`
+				<p>
+					<img src="https://emoji.discourse-cdn.com/twitter/upside_down_face.png?v=12" title=":upside_down_face:" class="emoji" alt=":upside_down_face:" loading="lazy" width="20" height="20">
+					<img src="https://emoji.discourse-cdn.com/blah/raised_back_of_hand/2.png?v=12" title=":blah:t2:" class="emoji" alt=":blah:t2:" loading="lazy" width="20" height="20">
+					<img src="https://emoji.discourse-cdn.com/apple/cold_face.png?v=12" title=":cold_face:" class="emoji" alt=":cold_face:" loading="lazy" width="20" height="20">
+				</p>
+			`),
+			strings.TrimSpace(`
+				<p>
+					🙃
+					<img src="https://emoji.discourse-cdn.com/blah/raised_back_of_hand/2.png?v=12" title=":blah:t2:" class="emoji" alt=":blah:t2:" loading="lazy" width="20" height="20"/>
+					🥶
+				</p>
+			`),
+		},
+		{
+			"preserves links without a base URL",
+			nil,
+			strings.TrimSpace(`
+				<a href="https://example.com">example</a>
+				<a href="/relative">relative</a>
+				<a href="anchor">anchor</a>
+				<a href="mailto:[email protected]">email</a>
+				<a href="?query">query</a>
+				<a href="#fragment">fragment</a>
+			`),
+			strings.TrimSpace(`
+				<a href="https://example.com">example</a>
+				<a href="/relative">relative</a>
+				<a href="anchor">anchor</a>
+				<a href="mailto:[email protected]">email</a>
+				<a href="?query">query</a>
+				<a href="#fragment">fragment</a>
+			`),
+		},
+		{
+			"updates links correctly with a base URL",
+			urlMustParse("https://example.invalid/testing/foo"),
+			strings.TrimSpace(`
+				<a href="https://example.com">example</a>
+				<a href="/relative">relative</a>
+				<a href="anchor">anchor</a>
+				<a href="mailto:[email protected]">email</a>
+				<a href="?query">query</a>
+				<a href="#fragment">fragment</a>
+			`),
+			strings.TrimSpace(`
+				<a href="https://example.com">example</a>
+				<a href="https://example.invalid/relative">relative</a>
+				<a href="https://example.invalid/testing/anchor">anchor</a>
+				<a href="mailto:[email protected]">email</a>
+				<a href="https://example.invalid/testing/foo?query">query</a>
+				<a href="https://example.invalid/testing/foo#fragment">fragment</a>
+			`),
+		},
 	}
 
 	for _, test := range tests {
 		t.Run(test.name, func(t *testing.T) {
 			is := is.New(t)
 
-			res, err := ReplaceHTMLImagesWithEmojis(test.input)
+			res, err := UnmungeCookedHTML(test.inputSource, test.inputBaseURL)
 
 			is.NoErr(err)
 			is.Equal(res, test.expected)