Defer XML string cleaning to improve performance (#1511)

* perf: Attempt to parse XML strings without cleaning (which is expensive) before trying again with cleaning

* refactor: Removed unused XML import

* refactor: Make use of the new parseXMLString function in myplex.py

* perf: Only encode strings to utf-8 once in the event a parse error occurs in parseXMLStrings

* refacotr: Updated the parseXMLString function to more closely follow the logic that previously existed

* refactor: Use explicit imports and reorder imports according to the repo's style guidelines

* fix: Bad exception handler

Author: eliasbenb

plexapi/client.py

@@ -197,8 +197,7 @@ def query(self, path, method=None, headers=None, timeout=None, **kwargs):
                 raise NotFound(message)
             else:
                 raise BadRequest(message)
-        data = utils.cleanXMLString(response.text).encode('utf8')
-        return ElementTree.fromstring(data) if data.strip() else None
+        return utils.parseXMLString(response.text)
 
     def sendCommand(self, command, proxy=None, **params):
         """ Convenience wrapper around :func:`~plexapi.client.PlexClient.query` to more easily

plexapi/media.py

@@ -1,7 +1,7 @@
 # -*- coding: utf-8 -*-
-import xml
 from pathlib import Path
 from urllib.parse import quote_plus
+from xml.etree import ElementTree
 
 from plexapi import log, settings, utils
 from plexapi.base import PlexObject
@@ -1077,7 +1077,7 @@ def select(self):
         data = f'{key}?url={quote_plus(self.ratingKey)}'
         try:
             self._server.query(data, method=self._server._session.put)
-        except xml.etree.ElementTree.ParseError:
+        except ElementTree.ParseError:
             pass
 
     @property

plexapi/myplex.py

@@ -4,7 +4,6 @@
 import threading
 import time
 from urllib.parse import parse_qsl, urlencode, urlsplit, urlunsplit
-from xml.etree import ElementTree
 
 import requests
 
@@ -250,8 +249,7 @@ def query(self, url, method=None, headers=None, timeout=None, **kwargs):
             return response.json()
         elif 'text/plain' in response.headers.get('Content-Type', ''):
             return response.text.strip()
-        data = utils.cleanXMLString(response.text).encode('utf8')
-        return ElementTree.fromstring(data) if data.strip() else None
+        return utils.parseXMLString(response.text)
 
     def ping(self):
         """ Ping the Plex.tv API.
@@ -1879,8 +1877,7 @@ def _query(self, url, method=None, headers=None, **kwargs):
             codename = codes.get(response.status_code)[0]
             errtext = response.text.replace('\n', ' ')
             raise BadRequest(f'({response.status_code}) {codename} {response.url}; {errtext}')
-        data = response.text.encode('utf8')
-        return ElementTree.fromstring(data) if data.strip() else None
+        return utils.parseXMLString(response.text)
 
 
 def _connect(cls, url, token, session, timeout, results, i, job_is_done_event=None):

plexapi/server.py

@@ -2,7 +2,6 @@
 import os
 from functools import cached_property
 from urllib.parse import urlencode
-from xml.etree import ElementTree
 
 import requests
 
@@ -768,8 +767,7 @@ def query(self, key, method=None, headers=None, params=None, timeout=None, **kwa
                 raise NotFound(message)
             else:
                 raise BadRequest(message)
-        data = utils.cleanXMLString(response.text).encode('utf8')
-        return ElementTree.fromstring(data) if data.strip() else None
+        return utils.parseXMLString(response.text)
 
     def search(self, query, mediatype=None, limit=None, sectionId=None):
         """ Returns a list of media items or filter categories from the resulting

plexapi/utils.py

@@ -17,12 +17,12 @@
 from hashlib import sha1
 from threading import Event, Thread
 from urllib.parse import quote
+from xml.etree import ElementTree
 
 import requests
 from requests.status_codes import _codes as codes
 
 from plexapi.exceptions import BadRequest, NotFound, Unauthorized
-
 try:
     from tqdm import tqdm
 except ImportError:
@@ -718,3 +718,14 @@ def sha1hash(guid):
 
 def cleanXMLString(s):
     return _illegal_XML_re.sub('', s)
+
+
+def parseXMLString(s: str):
+    """ Parse an XML string and return an ElementTree object. """
+    if not s.strip():
+        return None
+    try:  # Attempt to parse the string as-is without cleaning (which is expensive)
+        return ElementTree.fromstring(s.encode('utf-8'))
+    except ElementTree.ParseError:  # If it fails, clean the string and try again
+        cleaned_s = cleanXMLString(s).encode('utf-8')
+        return ElementTree.fromstring(cleaned_s) if cleaned_s.strip() else None