retronews

commit 0ca9709b3679feda2304460f1c9092eb1f32c569
parent 36b1785c1374dbc540a5da4c34939c96f2922c76
Author: luke8086 <55237178+luke8086@users.noreply.github.com>
Date:   Fri, 29 Jul 2022 16:11:45 +0000

Decode HTML entities in message titles

Diffstat:
M
retronews.py
 | 
8
+++++---

1 file changed, 5 insertions(+), 3 deletions(-)
diff --git a/retronews.py b/retronews.py
@@ -847,7 +847,7 @@ def hn_parse_search_hit(hit: HNSearchHit) -> Message:
         content_location=f"https://news.ycombinator.com/item?id={hit['objectID']}",
         date=datetime.fromtimestamp(hit["created_at_i"]),
         author=hit["author"],
-        title=hit["title"],
+        title=html.unescape(hit["title"]),
         total_comments=(hit["num_comments"] or 0) + 1,
     )
 
@@ -855,6 +855,8 @@ def hn_parse_search_hit(hit: HNSearchHit) -> Message:
 def hn_parse_entry(entry: HNEntry, thread_id: str = "", parent_title: str = "") -> Message:
     thread_id = thread_id or str(entry["id"])
 
+    my_title = html.unescape(entry["title"]) if entry["title"] else None
+
     body = f"<p>{entry['url']}</p>" if entry["url"] else ""
     body = f"{body}{entry['text']}" if entry["text"] else body
 
@@ -864,9 +866,9 @@ def hn_parse_entry(entry: HNEntry, thread_id: str = "", parent_title: str = "") 
         content_location=f"https://news.ycombinator.com/item?id={entry['id']}",
         date=datetime.fromtimestamp(entry["created_at_i"]),
         author=entry["author"] or "unknown",
-        title=entry["title"] or f"Re: {parent_title}",
+        title=my_title or f"Re: {parent_title}",
         body=body,
-        children=[hn_parse_entry(child, thread_id, entry["title"] or parent_title) for child in entry["children"]],
+        children=[hn_parse_entry(child, thread_id, my_title or parent_title) for child in entry["children"]],
     )