From d4c8e4d2da17783ffcb9744f162ca0465e862e14 Mon Sep 17 00:00:00 2001
From: Evan Reichard <evan@reichard.io>
Date: Sun, 11 Aug 2024 11:02:46 -0400
Subject: [PATCH] fix(search): broken parser & download source

---
 search/anna.go   | 38 ++++++++++++++++++++++++++++++++++++--
 search/search.go |  4 ++--
 2 files changed, 38 insertions(+), 4 deletions(-)

diff --git a/search/anna.go b/search/anna.go
index c0a9eb2..3473c7e 100644
--- a/search/anna.go
+++ b/search/anna.go
@@ -3,11 +3,14 @@ package search
 import (
 	"fmt"
 	"io"
+	"regexp"
 	"strings"
 
 	"github.com/PuerkitoBio/goquery"
 )
 
+var commentRE = regexp.MustCompile(`(?s)<!--(.*?)-->`)
+
 func parseAnnasArchiveDownloadURL(body io.ReadCloser) (string, error) {
 	// Parse
 	defer body.Close()
@@ -25,6 +28,35 @@ func parseAnnasArchiveDownloadURL(body io.ReadCloser) (string, error) {
 	return downloadURL, nil
 }
 
+// getAnnasArchiveBookSelection parses potentially commented out HTML. For some reason
+// Annas Archive comments out blocks "below the fold". They aren't rendered until you
+// scroll. This attempts to parse the commented out HTML.
+func getAnnasArchiveBookSelection(rawBook *goquery.Selection) *goquery.Selection {
+	rawHTML, err := rawBook.Html()
+	if err != nil {
+		return rawBook
+	}
+
+	strippedHTML := strings.TrimSpace(rawHTML)
+	if !strings.HasPrefix(strippedHTML, "<!--") || !strings.HasSuffix(strippedHTML, "-->") {
+		return rawBook
+	}
+
+	allMatches := commentRE.FindAllStringSubmatch(strippedHTML, -1)
+	if len(allMatches) != 1 || len(allMatches[0]) != 2 {
+		return rawBook
+	}
+
+	captureGroup := allMatches[0][1]
+	docReader := strings.NewReader(captureGroup)
+	doc, err := goquery.NewDocumentFromReader(docReader)
+	if err != nil {
+		return rawBook
+	}
+
+	return doc.Selection
+}
+
 func parseAnnasArchive(body io.ReadCloser) ([]SearchItem, error) {
 	// Parse
 	defer body.Close()
@@ -36,18 +68,20 @@ func parseAnnasArchive(body io.ReadCloser) ([]SearchItem, error) {
 	// Normalize Results
 	var allEntries []SearchItem
 	doc.Find("form > div.w-full > div.w-full > div > div.justify-center").Each(func(ix int, rawBook *goquery.Selection) {
+		rawBook = getAnnasArchiveBookSelection(rawBook)
+
 		// Parse Details
 		details := rawBook.Find("div:nth-child(2) > div:nth-child(1)").Text()
 		detailsSplit := strings.Split(details, ", ")
 
 		// Invalid Details
-		if len(detailsSplit) < 3 {
+		if len(detailsSplit) < 4 {
 			return
 		}
 
 		language := detailsSplit[0]
 		fileType := detailsSplit[1]
-		fileSize := detailsSplit[2]
+		fileSize := detailsSplit[3]
 
 		// Get Title & Author
 		title := rawBook.Find("h3").Text()
diff --git a/search/search.go b/search/search.go
index 02e4015..ebba671 100644
--- a/search/search.go
+++ b/search/search.go
@@ -57,9 +57,9 @@ type sourceDef struct {
 var sourceDefs = map[Source]sourceDef{
 	SOURCE_ANNAS_ARCHIVE: {
 		searchURL:         "https://annas-archive.org/search?index=&q=%s&ext=epub&sort=&lang=en",
-		downloadURL:       "http://libgen.li/ads.php?md5=%s",
+		downloadURL:       "http://library.lol/fiction/%s",
 		parseSearchFunc:   parseAnnasArchive,
-		parseDownloadFunc: parseAnnasArchiveDownloadURL,
+		parseDownloadFunc: parseLibGenDownloadURL,
 	},
 	SOURCE_LIBGEN_FICTION: {
 		searchURL:         "https://libgen.is/fiction/?q=%s&language=English&format=epub",