From 5336615375c66000e63a492d9f82a6cb99cf1d4d Mon Sep 17 00:00:00 2001
From: Seth Falco <seth@falco.fun>
Date: Fri, 21 Jul 2023 14:50:00 +0100
Subject: [PATCH] fix: remove html tags in overviews

---
 resources/lib/item_functions.py  |   4 +-
 resources/lib/plainhtmlparser.py | 123 +++++++++++++++++++++++++++++++
 resources/lib/utils.py           |  14 ++++
 3 files changed, 139 insertions(+), 2 deletions(-)
 create mode 100644 resources/lib/plainhtmlparser.py

diff --git a/resources/lib/item_functions.py b/resources/lib/item_functions.py
index 45dd2e1c..d1188a2e 100644
--- a/resources/lib/item_functions.py
+++ b/resources/lib/item_functions.py
@@ -10,7 +10,7 @@
 import xbmcgui
 
 from .utils import (
-    datetime_from_string, get_art_url, image_url, get_current_datetime
+    datetime_from_string, get_art_url, image_url, get_current_datetime, plainify_html
 )
 from .lazylogger import LazyLogger
 
@@ -300,7 +300,7 @@ def extract_item_info(item, gui_options):
         item_details.resume_time = int(reasonable_ticks / 10000)
 
     item_details.series_name = item.get("SeriesName", '')
-    item_details.plot = item.get("Overview", '')
+    item_details.plot = plainify_html(item.get("Overview", ''))
 
     runtime = item.get("RunTimeTicks")
     if item_details.is_folder is False and runtime:
diff --git a/resources/lib/plainhtmlparser.py b/resources/lib/plainhtmlparser.py
new file mode 100644
index 00000000..3c9ee36c
--- /dev/null
+++ b/resources/lib/plainhtmlparser.py
@@ -0,0 +1,123 @@
+from html.parser import HTMLParser
+
+
+class PlainHTMLParser(HTMLParser):
+    '''
+    HTMLParser implementation that strips HTML tags, preserving the content.
+    This is not intended to interpret HTML, nor output sanitized and secure HTML
+    that's safe to use in a web browser.
+
+    This parses a string that may contain HTML, and removes HTML tags, and content
+    that isn't intended for users to read, such as <script> and <style>. It will
+    preserve the content of tags that is semantically intended to be read by the
+    user, such <a>, <p>, and <span>.
+
+    Implementation should always be locale independent. It works with the HTML,
+    not string/ASCII content.
+    '''
+    html_elements = [
+        "html", "base", "head", "link", "meta", "style", "title", "body",
+        "address", "article", "aside", "footer", "header", "h1", "h2", "h3",
+        "h4", "h5", "h6", "hgroup", "main", "nav", "section", "search",
+        "blockquote", "dd", "div", "dl", "dt", "figcaption", "figure", "hr",
+        "li", "menu", "ol", "p", "pre", "ul", "a", "abbr", "b", "bdi", "bdo",
+        "br", "cite", "code", "data", "dfn", "em", "i", "kbd", "mark", "q",
+        "rp", "rt", "ruby", "s", "samp", "small", "span", "strong", "sub",
+        "sup", "time", "u", "var", "wbr", "area", "audio", "img", "map",
+        "track", "video", "embed", "iframe", "object", "picture", "portal",
+        "source", "svg", "math", "canvas", "noscript", "script", "del", "ins",
+        "caption", "col", "colgroup", "table", "tbody", "td", "tfoot", "th",
+        "thead", "tr", "button", "datalist", "fieldset", "form", "input",
+        "label", "legend", "meter", "optgroup", "option", "output", "progress",
+        "select", "textarea", "details", "dialog", "summary", "slot", "template"
+    ]
+    '''
+    List of HTML elements, excluding obsolete and deprecated HTML elements.
+
+    This allows us to seperate stylized text from actual HTML. For example,
+    if a description or movie name contains "<3", it will be preserved because
+    "3" is not the list of known HTML elements.
+
+    See: https://developer.mozilla.org/en-US/docs/Web/HTML/Element
+    '''
+
+    inline_elements = [
+        "a", "b", "em", "i", "s", "span", "strong", "sub", "sup", "u"
+    ]
+    '''
+    List of inline HTML elements. Allows us know which element doesn't need to
+    have whitespace appended.
+    '''
+
+    tag_denylist = [
+        "head", "meta", "style", "canvas", "noscript", "script", "summary"
+    ]
+    '''
+    Denylist of tags with content we don't want to display.
+
+    We drop <summary> because it's part of the <details> tag, but since we'll
+    always display the details anyway, there is no need for the summary of it.
+    '''
+
+    def __init__(self):
+        super().__init__(convert_charrefs=True)
+        self.elements = []
+        self.accumulator = []
+        self.pending_data = []
+        self.result = None
+
+    def handle_starttag(self, tag, _):
+        if self.elements and self.elements[-1] in PlainHTMLParser.tag_denylist:
+            return
+
+        self.handle_pending_data()
+
+        if tag not in PlainHTMLParser.html_elements:
+            self.accumulator.append(self.get_starttag_text())
+            return
+
+        self.elements.append(tag)
+
+    def handle_endtag(self, tag):
+        self.handle_pending_data()
+
+        if self.elements and self.elements[-1] == tag:
+            self.elements.pop()
+            return
+        elif self.elements and self.elements[-1] in PlainHTMLParser.tag_denylist:
+            return
+
+        if tag not in PlainHTMLParser.html_elements:
+            self.accumulator.append(self.get_starttag_text())
+
+    def handle_startendtag(self, tag, _):
+        if self.elements and self.elements[-1] in PlainHTMLParser.tag_denylist:
+            return
+
+        self.handle_pending_data()
+
+        if tag not in PlainHTMLParser.html_elements:
+            self.accumulator.append(self.get_starttag_text())
+
+    def handle_data(self, data):
+        if self.elements and self.elements[-1] in PlainHTMLParser.tag_denylist:
+            return
+
+        self.pending_data.append(data)
+
+    def close(self):
+        super().close()
+        self.handle_pending_data()
+        self.result = "".join(self.accumulator)
+
+    def handle_pending_data(self):
+        if not self.pending_data:
+            return
+
+        data_concat = "".join(self.pending_data)
+
+        if self.accumulator and self.elements and self.elements[-1] not in PlainHTMLParser.inline_elements:
+            data_concat = " " + data_concat
+
+        self.accumulator.append(data_concat)
+        self.pending_data.clear()
diff --git a/resources/lib/utils.py b/resources/lib/utils.py
index 9612e9b8..e6603a40 100644
--- a/resources/lib/utils.py
+++ b/resources/lib/utils.py
@@ -26,6 +26,7 @@
 
 from .lazylogger import LazyLogger
 from .kodi_utils import HomeWindow
+from .plainhtmlparser import PlainHTMLParser
 
 # hack to get datetime strptime loaded
 throwaway = time.strptime('20110101', '%Y%m%d')
@@ -451,3 +452,16 @@ def get_bitrate(enum_value):
                7000, 8000, 9000, 10000, 12000, 14000, 16000, 18000,
                20000, 25000, 30000, 35000, 40000, 100000, 1000000, 2147483]
     return bitrate[int(enum_value) if enum_value else 24] * 1000
+
+
+def plainify_html(body):
+    '''
+    Strip HTML elements from the string, preserving human-readable content.
+    '''
+    if body is None:
+        raise ValueError("body must not be None")
+
+    parser = PlainHTMLParser()
+    parser.feed(body)
+    parser.close()
+    return parser.result