fix: run html.unescape on the title

some titles contains things like '&#8211' so we run html.unescape to
remove them.

1 files changed, 2 insertions, 1 deletions

diff --git a/extract-title b/extract-title
index 46b4d3630..334ad9946 100755
--- a/extract-title
+++ b/extract-title
@@ -2,13 +2,14 @@
 import os
 import sys
 import json
+import html
 
 for slug in sys.argv[1:]:
     data_path = os.path.join("events", slug, "data.json")
     title_path = os.path.join("events", slug, "title")
     with open(data_path, "r") as f:
         data = json.load(f)
-    title = data["title"]["rendered"]
+    title = html.unescape(data["title"]["rendered"])
     if os.path.exists(title_path):
         continue
     with open(title_path, "w") as f: