From b9428d9af77a1ea050115c9ac9a1bd7acb0421a4 Mon Sep 17 00:00:00 2001
From: TKE <tobias.kessels@certbw.de>
Date: Wed, 23 Dec 2020 11:48:59 +0100
Subject: [PATCH] Add geurls.py

geturls.py prints links from html file. html can be a file on disk or
url starting with http/https
---
 geturls.py | 31 +++++++++++++++++++++++++++++++
 1 file changed, 31 insertions(+)
 create mode 100755 geturls.py

diff --git a/geturls.py b/geturls.py
new file mode 100755
index 0000000..be24cf1
--- /dev/null
+++ b/geturls.py
@@ -0,0 +1,31 @@
+#!/usr/bin/env python3
+import sys
+from bs4 import BeautifulSoup
+
+if sys.argv[1].startswith("http://") or sys.argv[1].startswith("https://"):
+    import requests
+    response = requests.get(sys.argv[1])
+    data = response.content
+else:
+    with open(sys.argv[1],'rt',encoding='ISO-8859-1') as f:
+        data=f.read()
+
+page=str(BeautifulSoup(data,features="lxml"))
+
+def getURL(page):
+    start_link = page.find("a href")
+    if start_link == -1:
+        return None, 0
+    start_quote = page.find('"', start_link)
+    end_quote = page.find('"', start_quote + 1)
+    url = page[start_quote + 1: end_quote]
+    return url, end_quote
+
+
+while True:
+    url, n = getURL(page)
+    page = page[n:]
+    if url:
+        print(url)
+    else:
+        break