From 7a73c6adf7b85bab08dea56ee32f3b2026e416e0 Mon Sep 17 00:00:00 2001
From: David Beitey <david@davidjb.com>
Date: Mon, 4 May 2020 18:45:40 +1000
Subject: [PATCH] scrape: extract strings from new non-text tags (#35021)

With the upgrade to beautifulsoup4 to 4.9.0 (#34007), certain tags
(`<style>`, `<script>` and `<template>`) are no longer treated as having
text content (see
https://www.crummy.com/software/BeautifulSoup/bs4/doc/#comments-and-other-special-strings
and reported bug https://bugs.launchpad.net/beautifulsoup/+bug/1868861)
meaning the content of these types of tags became inaccessible to HA.

Where the previous code could access `.text` on the tag, bs4 4.9 now
yields an empty string; these types of tags require accesing `.string`
instead.  This PR checks the tag name (which will aalways be lowercase
given how the parser works;
https://www.crummy.com/software/BeautifulSoup/bs4/doc/#other-parser-problems)
and applies this different access strategy to get the content of the
HTML tag.  All other tags are handled in the original manner.
---
 homeassistant/components/scrape/sensor.py | 6 +++++-
 1 file changed, 5 insertions(+), 1 deletion(-)

diff --git a/homeassistant/components/scrape/sensor.py b/homeassistant/components/scrape/sensor.py
index 13d99a0cb8f..4119f7e4c6b 100644
--- a/homeassistant/components/scrape/sensor.py
+++ b/homeassistant/components/scrape/sensor.py
@@ -132,7 +132,11 @@ class ScrapeSensor(Entity):
             if self._attr is not None:
                 value = raw_data.select(self._select)[self._index][self._attr]
             else:
-                value = raw_data.select(self._select)[self._index].text
+                tag = raw_data.select(self._select)[self._index]
+                if tag.name in ("style", "script", "template"):
+                    value = tag.string
+                else:
+                    value = tag.text
             _LOGGER.debug(value)
         except IndexError:
             _LOGGER.error("Unable to extract data from HTML")