From 6725a9af634ae459d71e3f20cfc7af90d6b5d992 Mon Sep 17 00:00:00 2001
From: Alon Levy <alon@greenvibe.io>
Date: Sun, 31 Dec 2023 13:40:47 +0200
Subject: [PATCH] decode.py: rewrite to use iterators

---
 decode.py | 62 ++++++++++++++++++++++++++++---------------------------
 1 file changed, 32 insertions(+), 30 deletions(-)
 mode change 100644 => 100755 decode.py

diff --git a/decode.py b/decode.py
old mode 100644
new mode 100755
index 3a4f809..aa6152c
--- a/decode.py
+++ b/decode.py
@@ -1,3 +1,4 @@
+#!/usr/bin/env python3
 import argparse
 import os
 import subprocess
@@ -25,7 +26,7 @@ basename = os.path.splitext(os.path.basename(input_file))[0]
 
 # Read packets from a pcap file
 scanner = pcapng.scanner.FileScanner(open(input_file, "rb"))
-blocks = list(tqdm(scanner))
+blocks = tqdm(scanner)
 
 # Helper function to safely get an attribute from an object
 def tryget(obj, att):
@@ -34,17 +35,22 @@ def tryget(obj, att):
     return None
 
 
-# Create a DataFrame with packet lengths
-df = pd.DataFrame(
-    [{"index": i, "length": tryget(obj, "packet_len")} for i, obj in enumerate(blocks)]
-)
+
+def rightsize(it):
+    for i, obj in enumerate(it):
+        if not hasattr(obj, 'packet_len'):
+            continue
+        len = obj.packet_len
+        if len != 6972:
+            continue
+        yield obj.packet_data
 
 
-# Filter and extract data packets of a specific length
-data = [blocks[i] for i in df[df.length == 6972.0].index]
+def removestart(it):
+    "Remove the UDP header from the packets"
+    for x in it:
+        yield x[0x2A:]
 
-# Remove the UDP header from the packets
-raw = [d.packet_data[0x2A:] for d in data]
 
 
 # Function to parse packet data
@@ -59,38 +65,34 @@ def parse(data):
     return ret
 
 
-# Parse each packet and create a DataFrame
-df = pd.DataFrame([parse(d) for d in raw])
-df2 = df[[c for c in df.columns if c != "data"]]
+def parsed(it):
+    for x in it:
+        yield parse(x)
 
 
 # Function to group data into frames
-def getframes(df):
-    frames = []
+def frames(it):
     current = []
-    for i, row in df.iterrows():
-        if row.part == 0:
+    for obj in it:
+        if obj['part'] == 0:
             if len(current) > 0:
-                frames.append(current)
+                yield b"".join(current)
             current = []
-        current.append(row["data"])
+        current.append(obj["data"])
     if len(current) > 0:
-        frames.append(current)
-    return [b"".join(parts) for parts in frames]
+        yield b"".join(current)
 
 
-# Function to convert binary frame data into images
-def image16(frame, width, height, pixelformat=">H"):
-    return [
-        Image.fromarray(np.frombuffer(frame, dtype=pixelformat).reshape(width, height))
-        for frame in frames
-        if len(frame) == 2 * width * height
-    ]
+def iterimages(it, width, height, pixelformat=">H"):
+    for frame in it:
+        if len(frame) != width * height * 2:  # 16 bpp
+            continue
+        yield Image.fromarray(np.frombuffer(frame, dtype=pixelformat).reshape(width, height))
 
 
 # Get frames and convert them to images
-frames = getframes(df)
-images = image16(frames, 384, 288)
+frames = frames(parsed(removestart(rightsize(blocks))))
+images = iterimages(it=frames, width=384, height=288)
 
 # Create the directory for frames if not exists
 frame_dir = f"frames/{basename}"
@@ -98,7 +100,7 @@ if not os.path.exists(frame_dir):
     os.makedirs(frame_dir)
 
 # Save each image as a PNG file
-for i, img in enumerate(tqdm(images)):
+for i, img in enumerate(images):
     img.save(f'frames/{basename}/{basename}_{i:04}.png')
 
 # Produce a video from the saved images