From 63b6942f6b2649c70c30cdced6c033ff2607724f Mon Sep 17 00:00:00 2001
From: Nick White <git@njw.name>
Date: Mon, 21 Mar 2022 16:26:55 +0000
Subject: rescribe: move getBookIdFromUrl() to gbook.go, and add tests for it

---
 cmd/rescribe/gbook.go | 47 +++++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 47 insertions(+)

(limited to 'cmd/rescribe/gbook.go')

diff --git a/cmd/rescribe/gbook.go b/cmd/rescribe/gbook.go
index f0e9e1e..fe2f4b8 100644
--- a/cmd/rescribe/gbook.go
+++ b/cmd/rescribe/gbook.go
@@ -14,6 +14,7 @@ import (
 	"os"
 	"os/exec"
 	"path"
+	"regexp"
 	"strings"
 	"unicode"
 
@@ -206,3 +207,49 @@ func getGoogleBook(ctx context.Context, gbookcmd string, id string, basedir stri
 
 	return dir, nil
 }
+
+// getBookIdFromUrl returns a 12 character Google Book ID from
+// a Google URL, or an error if one can't be found.
+// Example URLs:
+// https://books.google.it/books?id=QjQepCuN8JYC
+// https://www.google.it/books/edition/_/VJbr-Oe2au0C
+func getBookIdFromUrl(url string) (string, error) {
+	lurl := strings.ToLower(url)
+	if len(url) == 12 && !strings.ContainsAny(url, "?/:") {
+		return url, nil
+	}
+
+	matchUrl, err := regexp.MatchString("https://www.google.[^\\/]*/books/", url)
+	if err != nil {
+		return "", err
+	}
+
+	if strings.HasPrefix(lurl, "https://books.google") {
+		start := strings.Index(lurl, "?id=")
+		if start == -1 {
+			start = strings.Index(lurl, "&id=")
+		}
+
+		if start >= 0 {
+			start += 4
+			if len(url[start:]) < 12 {
+				return "", fmt.Errorf("Could not find book ID in URL")
+			}
+			return url[start : start+12], nil
+		}
+
+		return "", fmt.Errorf("Could not find book ID in URL")
+	}
+	if matchUrl == true {
+		start := strings.Index(lurl, "edition/_/")
+
+		if start >= 0 {
+			start += 10
+			if len(url[start:]) < 12 {
+				return "", fmt.Errorf("Could not find book ID in URL")
+			}
+			return url[start : start+12], nil
+		}
+	}
+	return "", fmt.Errorf("Could not find book ID in URL")
+}
-- 
cgit v1.2.1-24-ge1ad