autocleanup of manually deleted images from media

1 files changed, 6 insertions, 0 deletions

diff --git a/bot.py b/bot.py
index edf04c1..a012827 100644
--- a/bot.py
+++ b/bot.py
@@ -254,12 +254,15 @@ vector_db_ready = False  # gets set to True after first rebuild so we don't keep
 
 saved_media = set()  # file names inside MEDIA_DIR
 
+kept_media_lines = []
 with MEDIA_FILE.open("r", encoding="utf-8") as f:
     for line in f:
         if "saved media:" in line:
             name = line.split("saved media:", 1)[1].strip()
             if (MEDIA_DIR / name).is_file():  # old entries were urls (those expire), they don't match a file so they get dropped
                 saved_media.add(name)
+                kept_media_lines.append(line)
+MEDIA_FILE.write_text("".join(kept_media_lines), encoding="utf-8")  # rewritten without the entries of deleted files
 
 async def steal_media(url: str, name: str):
     # discord cdn links are signed and die after a day or so, so we keep our own copy
@@ -428,6 +431,9 @@ def sync_reactions(guild):
             if n not in data[key]:
                 data[key][n] = ""
                 changed = True
+    for n in [n for n in data["media"] if n not in saved_media]:  # file got deleted, drop its entry too
+        del data["media"][n]
+        changed = True
     if changed:
         REACTIONS_FILE.write_text(json.dumps(data, indent=2, ensure_ascii=False) + "\n", encoding="utf-8")