perf(parsing.utils): ensure_plain: use run_async only if processing HTML

Utilizing aio_helper.run_async() actually means that the function call
will be queued until the ThreadPoolExecutor is free to pick it up. The
queue can be extremely long under high loads.

Preventing ensure_plain from utilizing aio_helper.run_async() when
unneeded, so that the pressure of ThreadPoolExecutor won't be too high.

Signed-off-by: Rongrong <[email protected]>

src/parsing/utils.py

@@ -15,7 +15,7 @@
 #  along with this program.  If not, see <https://www.gnu.org/licenses/>.
 
 from __future__ import annotations
-from typing import Optional, Sequence, Union, Final, Iterable, Awaitable
+from typing import Optional, Sequence, Union, Final, Iterable
 
 import re
 import string
@@ -166,13 +166,17 @@ async def html_validator(html: str) -> str:
     return await run_async(_html_validator, html, prefer_pool='thread')
 
 
-def ensure_plain_sync(s: str, enable_emojify: bool = False) -> str:
+def _bs_html_get_text(s: str) -> str:
+    return BeautifulSoup(s, 'lxml').get_text()
+
+
+async def ensure_plain(s: str, enable_emojify: bool = False) -> str:
     if not s:
         return s
     s = stripAnySpace(
         replaceSpecialSpace(
             replaceInvalidCharacter(
-                BeautifulSoup(s, 'lxml').get_text()
+                await run_async(_bs_html_get_text, s, prefer_pool='thread')
                 if '<' in s and '>' in s
                 else unescape(s)
             )
@@ -181,10 +185,6 @@ def ensure_plain_sync(s: str, enable_emojify: bool = False) -> str:
     return emojify(s) if enable_emojify else s
 
 
-def ensure_plain(s: str, enable_emojify: bool = False) -> Awaitable[str]:
-    return run_async(ensure_plain_sync, s, enable_emojify, prefer_pool='thread')
-
-
 async def parse_entry(entry, feed_link: Optional[str] = None):
     class EntryParsed:
         content: str = ''