Cleanup of external assets

content/blog/2017-11-24-weatherballoon-lost.md

@@ -14,7 +14,7 @@ aliases:
 
 A news article was written about a project I was involved in. It's even headed by a photo I took!
 
-![](https://smartcdn.gprod.postmedia.digital/v1/dynamic_resize/sws_path/suns-prod-images/1298003385469_ORIGINAL.jpg)
+![](/images/posts/weatherballoon/927942227749416960-DOC2hmHVwAI8aXW.jpg)
 
 From the **London Free Press**:
 
@@ -34,4 +34,4 @@ Once found on the American side of the border, a corresponding article was writt
 
 ## The balloon on the other side
 
-![](https://bloximages.chicago2.vip.townnews.com/tiogapublishing.com/content/tncms/assets/v3/editorial/1/38/1384aff6-d120-11e7-bd7f-e7d82bd3ef3c/5a1826bd8b199.image.jpg)
+![](/images/posts/weatherballoon/5a1826bd8b199.webp)

content/blog/2022-06-26-rickrolling-the-internet.md

@@ -101,7 +101,7 @@ Generally, such a setup would involve daisy-chaining routers physically in your
 
 Conveniently for my wallet, Linux machines provide something called [Tun/Tap Interfaces](https://en.wikipedia.org/wiki/TUN/TAP). These virtual network interfaces allow programs to pretend to be one of many other computers on the network and act as if they were real hosts. When a program registers one of these interfaces with the kernel, it gets raw access to either the 2<sup>nd</sup> or 3<sup>rd</sup> OSI layer of the network stack in the form of a raw stream.
 
-![Tun/Tap in the OSI stack](https://upload.wikimedia.org/wikipedia/commons/thumb/a/af/Tun-tap-osilayers-diagram.png/400px-Tun-tap-osilayers-diagram.png)
+![Tun/Tap in the OSI stack](/images/posts/rickroll-ipv6/400px-Tun-tap-osilayers-diagram.png)
 
 I chose to register a Tun interface, and control things at the Internet Protocol level. This choice was mainly due to simplicity, as I really don't care about hardware addresses.
 

content/blog/2022-12-07-gamels.md

@@ -21,6 +21,6 @@ Linux gamers may be familiar with digging around the `copmpatdata` and `shaderca
 
 `gamels` solves this problem by querying the Steam API to see if any directory names match known app ids. If so, the name of the corresponding game is shown in the listing.
 
-![A screenshot of gamels in action](https://raw.githubusercontent.com/ewpratten/gamels/master/steamdeck-screenshot.png)
+![A screenshot of gamels in action](/images/posts/gamels/steamdeck-screenshot.png)
 
 Installation instructions and source code are available on [GitHub](https://github.com/ewpratten/gamels).

content/blog/2024-01-15-digest-bot.md

@@ -16,7 +16,6 @@ Each morning, shortly after I've woken up, I receive an email.
 <pre style="border-left:2px solid gray;border-radius:0;max-width:100%;overflow-x:scroll;">
 Subject: Your digest for Wednesday, January 03
 From: Digest Bot &lt;redacted@redacted.com&gt;
-
 Good morning, this is your daily digest for Wednesday January 03, 2024.
 
 Daytime weather is expected to be cloudy. 30 percent chance of flurries

sass/elements/table.scss

@@ -14,8 +14,20 @@ table {
     padding: 0;
     padding: 6px 13px;
     border: 1px solid #d0d7de;
+
+    &:has(> audio) {
+      padding: 0;
+      margin: 0;
+    }
+
+    audio {
+      display: block;
+      width: 100%;
+      min-width: 120px;
+    }
   }
 
+
   th {
     font-weight: 600;
   }

scripts/find_external_assets.py

@@ -0,0 +1,37 @@
+import re
+from pathlib import Path
+
+REPO_ROOT = Path(__file__).parent.parent
+
+# Find all MD and HTML files
+md_files = list(REPO_ROOT.rglob("*.md"))
+html_files = list(REPO_ROOT.rglob("*.html"))
+
+# Ignore any files in the `public` directory
+md_files = [f for f in md_files if "public" not in f.parts]
+html_files = [f for f in html_files if "public" not in f.parts]
+
+# Result storage
+external_assets = set()
+
+# Find Markdown images
+for file in md_files:
+    body = file.read_text()
+    for match in re.finditer(r"!\[.*?\]\((.*?)\)", body):
+        link = match.group(1)
+        if link.startswith("http"):
+            external_assets.add((file, link))
+
+# Search HTML
+for file in html_files:
+    body = file.read_text()
+    for match in re.finditer(r'src="(.*?)"', body):
+        link = match.group(1)
+        if link.startswith("http"):
+            external_assets.add((file, link))
+                
+# Print all external assets
+for file_path, link in external_assets:
+    # Strip the prefix off the file path
+    file_path = file_path.relative_to(REPO_ROOT)
+    print(f"{file_path}:\t{link}")