From 228f71708bae41878fd994788dd886956d695ff1 Mon Sep 17 00:00:00 2001
From: ed <s@ocv.me>
Date: Sun, 8 Aug 2021 02:47:04 +0200
Subject: [PATCH] improve youtube collector/parser

---
 bin/mtag/res/yt-ipr.user.js | 32 ++++++++++++++++++++++++++++++++
 bin/mtag/yt-ipr.py          | 18 ++++++++----------
 2 files changed, 40 insertions(+), 10 deletions(-)
 create mode 100644 bin/mtag/res/yt-ipr.user.js

diff --git a/bin/mtag/res/yt-ipr.user.js b/bin/mtag/res/yt-ipr.user.js
new file mode 100644
index 00000000..b81947e0
--- /dev/null
+++ b/bin/mtag/res/yt-ipr.user.js
@@ -0,0 +1,32 @@
+// ==UserScript==
+// @name    youtube-playerdata-hub
+// @match   https://youtube.com/*
+// @match   https://*.youtube.com/*
+// @version 1.0
+// @grant   GM_addStyle
+// ==/UserScript==
+
+function main() {
+    var sent = {};
+    function send(txt) {
+        if (sent[txt])
+            return;
+
+        fetch('https://127.0.0.1:3923/playerdata?_=' + Date.now(), { method: "PUT", body: txt });
+        console.log('[yt-ipr] yeet %d bytes', txt.length);
+        sent[txt] = 1;
+    }
+
+    function collect() {
+        setTimeout(collect, 60 * 1000);
+        var pd = document.querySelector('ytd-watch-flexy');
+        if (pd)
+            send(JSON.stringify(pd.playerData));
+    }
+    setTimeout(collect, 5000);
+}
+
+var scr = document.createElement('script');
+scr.textContent = '(' + main.toString() + ')();';
+(document.head || document.getElementsByTagName('head')[0]).appendChild(scr);
+console.log('[yt-ipr] a');
diff --git a/bin/mtag/yt-ipr.py b/bin/mtag/yt-ipr.py
index 3f4c2f6d..a7eba6bf 100644
--- a/bin/mtag/yt-ipr.py
+++ b/bin/mtag/yt-ipr.py
@@ -2,6 +2,7 @@
 
 import re
 import sys
+import gzip
 import json
 from datetime import datetime
 
@@ -14,20 +15,17 @@ example usage:
        :c,mtp=yt-id,yt-title,yt-author,yt-channel,yt-views,yt-private,yt-expires=bin/mtag/yt-ipr.py
        :c,mte=yt-id,yt-title,yt-author,yt-channel,yt-views,yt-private,yt-expires
 
-quick userscript to push them across:
-  console.log('a');
-  setTimeout(function() {
-    for (var scr of document.querySelectorAll('script[nonce]'))
-      if (scr.innerHTML.indexOf('manifest.googlevideo.com/api/manifest')>0)
-        fetch('https://127.0.0.1:3923/playerdata', {method:"PUT", body: scr.innerHTML});
-  }, 10*1000);
-
+see res/yt-ipr.user.js for the example userscript to go with this
 """
 
 
 def main():
-    with open(sys.argv[1], "r", encoding="utf-8") as f:
-        txt = f.read()
+    try:
+        with gzip.open(sys.argv[1], "rt", encoding="utf-8", errors="replace") as f:
+            txt = f.read()
+    except:
+        with open(sys.argv[1], "r", encoding="utf-8", errors="replace") as f:
+            txt = f.read()
 
     txt = "{" + txt.split("{", 1)[1]