From 95e2b086f37c2ed4f0f2dfa154521a1782eba5aa Mon Sep 17 00:00:00 2001
From: Tristan Smith <tristan.smith@gmail.com>
Date: Fri, 5 Apr 2024 22:49:10 -0400
Subject: [PATCH] Add sizegetter.py to calculate total size of files containing
 a specific keyword

---
 sizegetter.py | 29 +++++++++++++++++++++++++++++
 zipremove.py  |  0
 zipsize       | 17 +++++++++++++++++
 3 files changed, 46 insertions(+)
 create mode 100644 sizegetter.py
 create mode 100644 zipremove.py
 create mode 100644 zipsize

diff --git a/sizegetter.py b/sizegetter.py
new file mode 100644
index 0000000..356bdaa
--- /dev/null
+++ b/sizegetter.py
@@ -0,0 +1,29 @@
+import re
+import sys
+
+def calculate_total_size(filename, keyword):
+    pattern = re.compile(rf'{re.escape(keyword)}.*?(\d+(?:\.\d+)?)\s(MiB|GiB)')
+    total_size_mib = 0
+
+    with open(filename, 'r', encoding='utf-8') as file:
+        for line in file:
+            match = pattern.search(line)
+            if match:
+                size, unit = match.groups()
+                size = float(size)
+                if unit == "GiB":
+                    size *= 1024  # Convert GiB to MiB
+                total_size_mib += size
+
+    total_size_gib = total_size_mib / 1024
+    return total_size_mib, total_size_gib
+
+if __name__ == "__main__":
+    if len(sys.argv) < 3:
+        print("Usage: python script.py <path to file> <keyword>")
+        sys.exit(1)
+
+    filename = sys.argv[1]
+    keyword = sys.argv[2]
+    total_size_mib, total_size_gib = calculate_total_size(filename, keyword)
+    print(f"Total size of files containing '{keyword}': {total_size_gib:.2f} GiB ({total_size_mib:.0f} MiB)")
diff --git a/zipremove.py b/zipremove.py
new file mode 100644
index 0000000..e69de29
diff --git a/zipsize b/zipsize
new file mode 100644
index 0000000..98ac49e
--- /dev/null
+++ b/zipsize
@@ -0,0 +1,17 @@
+#!/bin/bash
+
+awk 'BEGIN{CONVFMT="%.2f"} 
+{
+    if ($1 ~ /^[0-9]+$/) {
+        byte=$1; 
+        if (byte >= 1073741824) 
+            {printf("%.2f GB\t%s\n", byte/1073741824, $4)}
+        else if (byte >= 1048576) 
+            {printf("%.2f MB\t%s\n", byte/1048576, $4)}
+        else if (byte >= 1024) 
+            {printf("%.2f KB\t%s\n", byte/1024, $4)}
+        else 
+            {print $1 " B\t" $4}
+    } 
+    else {print $0}
+}'