From 96cf231814d27f6dc9451e4eef0b7faaa63fc3c5 Mon Sep 17 00:00:00 2001
From: Peter Kurfer <peter.kurfer@rwe.com>
Date: Fri, 17 Jan 2025 14:45:14 +0100
Subject: [PATCH] feat(comics): add xkcd comics

---
 comics/main.go   | 81 ++++++++++++++++++++++++++++++++++++++----------
 comics/spin.toml |  2 +-
 2 files changed, 66 insertions(+), 17 deletions(-)

diff --git a/comics/main.go b/comics/main.go
index c55d17d..23b8ace 100644
--- a/comics/main.go
+++ b/comics/main.go
@@ -6,6 +6,7 @@ import (
 	"log/slog"
 	"math/rand/v2"
 	"net/http"
+	"net/url"
 
 	"github.com/PuerkitoBio/goquery"
 	spinhttp "github.com/fermyon/spin/sdk/go/v2/http"
@@ -22,8 +23,9 @@ type MonkeyUserEntry struct {
 }
 
 type Comic struct {
-	Title string `json:"title"`
-	URL   string `json:"url"`
+	Source string `json:"source"`
+	Title  string `json:"title"`
+	URL    string `json:"url"`
 }
 
 func init() {
@@ -31,12 +33,69 @@ func init() {
 
 	router.GET("/comics/random", randomComic)
 	router.GET("/comics/monkeyuser", monkeyUserComic)
+	router.GET("/comics/xkcd", xkcdComic)
 
 	spinhttp.Handle(router.ServeHTTP)
 }
 
 func randomComic(w http.ResponseWriter, re *http.Request, params httprouter.Params) {
-	monkeyUserComic(w, re, params)
+	possibleHandlers := []httprouter.Handle{
+		xkcdComic,
+		monkeyUserComic,
+	}
+
+	possibleHandlers[rand.IntN(len(possibleHandlers))](w, re, params)
+}
+
+func xkcdComic(w http.ResponseWriter, re *http.Request, _ httprouter.Params) {
+	pageResponse, err := spinhttp.Get("https://c.xkcd.com/random/comic/")
+	if err != nil {
+		slog.Error("failed to fetch xkcd comic page", slog.String("err", err.Error()))
+		http.Error(w, err.Error(), http.StatusInternalServerError)
+		return
+	}
+
+	slog.Info("Fetched xkcd comic page", slog.Int("status", pageResponse.StatusCode))
+
+	defer pageResponse.Body.Close()
+	if pageResponse.StatusCode != http.StatusOK {
+		http.Error(w, pageResponse.Status, pageResponse.StatusCode)
+		return
+	}
+
+	comicDoc, err := goquery.NewDocumentFromReader(pageResponse.Body)
+	if err != nil {
+		slog.Error("failed to parse comic page", slog.String("err", err.Error()))
+		http.Error(w, err.Error(), http.StatusInternalServerError)
+		return
+	}
+
+	slog.Info("Extracting comic info")
+
+	ref := Comic{
+		Source: "xkcd",
+		Title:  comicDoc.Find("#ctitle").Text(),
+		URL:    comicDoc.Find("#comic > img").AttrOr("src", ""),
+	}
+
+	parsed, err := url.Parse(ref.URL)
+	if err != nil {
+		slog.Error("failed to parse comic URL", slog.String("err", err.Error()))
+		http.Error(w, err.Error(), http.StatusInternalServerError)
+		return
+	}
+
+	if parsed.Scheme == "" {
+		parsed.Scheme = "https"
+		ref.URL = parsed.String()
+	}
+
+	w.Header().Add(contentTypeHeader, "application/json")
+	w.WriteHeader(http.StatusOK)
+	if err := json.NewEncoder(w).Encode(ref); err != nil {
+		slog.Error("failed to encode image ref", slog.String("err", err.Error()))
+		http.Error(w, err.Error(), http.StatusInternalServerError)
+	}
 }
 
 func monkeyUserComic(w http.ResponseWriter, re *http.Request, _ httprouter.Params) {
@@ -86,21 +145,11 @@ func monkeyUserComic(w http.ResponseWriter, re *http.Request, _ httprouter.Param
 	}
 
 	ref := Comic{
-		Title: entry.Title,
+		Source: "monkeyuser",
+		Title:  entry.Title,
+		URL:    fmt.Sprintf("https://www.monkeyuser.com%s", comicDoc.Find("div.content > p > img").AttrOr("src", "")),
 	}
 
-	comicDoc.Find("div.content > p > img").Each(func(_ int, s *goquery.Selection) {
-		for _, node := range s.Nodes {
-			for _, attr := range node.Attr {
-				if attr.Key == "src" {
-					ref.URL = fmt.Sprintf("https://www.monkeyuser.com%s", attr.Val)
-					break
-				}
-			}
-			slog.Info("found URL", slog.String("url", ref.URL))
-		}
-	})
-
 	w.Header().Add(contentTypeHeader, "application/json")
 	w.WriteHeader(http.StatusOK)
 	if err := json.NewEncoder(w).Encode(ref); err != nil {
diff --git a/comics/spin.toml b/comics/spin.toml
index f867fc1..0f08b2d 100644
--- a/comics/spin.toml
+++ b/comics/spin.toml
@@ -12,7 +12,7 @@ component = "comics"
 
 [component.comics]
 source = "main.wasm"
-allowed_outbound_hosts = [ "https://www.monkeyuser.com" ]
+allowed_outbound_hosts = [ "https://www.monkeyuser.com", "https://c.xkcd.com" ]
 [component.comics.build]
 command = "tinygo build -target=wasip1 -gc=leaking -no-debug -scheduler=none -buildmode=c-shared -o main.wasm main.go"
 watch = ["**/*.go", "go.mod"]