From 92fe7d3725f7785278c4bd2dfd5ad81e290827f5 Mon Sep 17 00:00:00 2001
From: Kovid Goyal <kovid@kovidgoyal.net>
Date: Sun, 31 Oct 2010 12:01:16 -0600
Subject: [PATCH] Amazon metadata download plugin: Improved parsing of broken
 HTML

---
 src/calibre/ebooks/metadata/amazon.py | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/src/calibre/ebooks/metadata/amazon.py b/src/calibre/ebooks/metadata/amazon.py
index e61e0b2748..5d7d0358f0 100644
--- a/src/calibre/ebooks/metadata/amazon.py
+++ b/src/calibre/ebooks/metadata/amazon.py
@@ -9,6 +9,7 @@ Fetch metadata using Amazon AWS
 import sys, re
 
 from lxml import html
+from lxml.html import soupparser
 
 from calibre import browser
 from calibre.ebooks.metadata import check_isbn
@@ -71,7 +72,7 @@ def get_metadata(br, asin, mi):
         return False
     raw = xml_to_unicode(raw, strip_encoding_pats=True,
             resolve_entities=True)[0]
-    root = html.fromstring(raw)
+    root = soupparser.fromstring(raw)
     ratings = root.xpath('//form[@id="handleBuy"]/descendant::*[@class="asinReviewsSummary"]')
     if ratings:
         pat = re.compile(r'([0-9.]+) out of (\d+) stars')