From 30756fc46267faad0af1fa7537d1e6045a34b982 Mon Sep 17 00:00:00 2001
From: "A.M. Rowsell" <amrowsell@frozenelectronics.ca>
Date: Mon, 3 Feb 2025 14:34:15 -0500
Subject: [PATCH] Added HTML filter, removed duplicate titles in embed, better
 output

---
 discorss.py | 23 ++++++++++++++++-------
 1 file changed, 16 insertions(+), 7 deletions(-)

diff --git a/discorss.py b/discorss.py
index c720c68..46224b3 100755
--- a/discorss.py
+++ b/discorss.py
@@ -14,25 +14,30 @@ import requests
 import feedparser
 from pathlib import Path
 import json
-# import datetime
 import time
 import os
+import re
 
 config_file_path = r"/etc/discorss.conf"
 # config_file_path = r"discorss.conf"
-# log_file_path = r"/var/log/discorss"
-log_file_path = r"./log"
+log_file_path = r"/var/log/discorss"
+# log_file_path = r"./log"
 log_file_name = r"/app.log"
+# Yes, I know you "can't parse HTML with regex", but
+# just watch me.
+html_filter = re.compile(r"\<\/?([A-Za-z \"\=])*\>")
 
 
 def get_description(feed):
     try:
         temporary_string = str(feed.entries[0]["summary_detail"]["value"])
+        temporary_string = html_filter.sub("", temporary_string)
         desc = (
             temporary_string[:150] if len(temporary_string) > 150 else temporary_string
         )
     except KeyError:
         temporary_string = str(feed.entries[0]["description"])
+        temporary_string = html_filter.sub("", temporary_string)
         desc = (
             temporary_string[:150] if len(temporary_string) > 150 else temporary_string
         )
@@ -59,7 +64,7 @@ def main():
         feed = feedparser.parse(hook["url"])
         published_time = time.mktime(feed.entries[0]["published_parsed"])
         published_time = published_time + hook["offset"]
-        print(feed.entries[0]["published"], published_time, now)
+        print("Parsing feed {}...".format(hook["name"]))
         # Generate the webhook
         webhook = {
             "embeds": [
@@ -74,7 +79,7 @@ def main():
                     "author": {"name": str(hook["name"]), "url": str(hook["siteurl"])},
                     "fields": [
                         {
-                            "name": str(feed.entries[0]["title"]),
+                            "name": "Excerpt from post:",
                             "value": get_description(feed),
                         }
                     ],
@@ -92,8 +97,12 @@ def main():
             r = requests.post(
                 hook["webhook"], data=webhook_string, headers=custom_header
             )
-            if r.status_code != '200':
-                print("Error {} while trying to post {}".format(r.status_code, hook["webhook"]))
+            if r.status_code != "200":
+                print(
+                    "Error {} while trying to post {}".format(
+                        r.status_code, hook["webhook"]
+                    )
+                )
         app_config["lastupdate"] = now
         with open(config_file_path, "w") as config_file:
             json.dump(app_config, config_file, indent=4)