From 6c59d7cf61e6a7b7426f7e680a381ac1923bdee0 Mon Sep 17 00:00:00 2001
From: Markus Birth <mbirth@gmail.com>
Date: Sun, 22 Sep 2019 02:27:44 +0200
Subject: [PATCH] Replaces find720 script with Python version.

---
 find720.py | 83 ++++++++++++++++++++++++++++++++++++++++++++++++++++++
 find720.sh | 15 ----------
 2 files changed, 83 insertions(+), 15 deletions(-)
 create mode 100755 find720.py
 delete mode 100755 find720.sh

diff --git a/find720.py b/find720.py
new file mode 100755
index 0000000..77ae109
--- /dev/null
+++ b/find720.py
@@ -0,0 +1,83 @@
+#!/usr/bin/env python3
+# -*- coding: utf-8 -*-
+
+import json
+import os
+from pprint import pprint
+import subprocess
+import sys
+
+VIDEO_EXTS = ["avi", "mp4", "mkv", "wmv", "mov", "m4v", "flv", "webm"]
+MAX_HEIGHT = 720    # Only list videos with height (=shortest side) of more than this
+
+# https://stackoverflow.com/questions/3844430/how-to-get-the-duration-of-a-video-in-python
+def ffprobe(filepath):
+    command = [
+        "ffprobe",
+        "-loglevel",  "quiet",
+        "-print_format", "json",
+        "-show_format",
+        "-show_streams",
+        filepath
+    ]
+
+    pipe = subprocess.Popen(command, stdout=subprocess.PIPE, stderr=subprocess.STDOUT)
+    out, err = pipe.communicate()
+    return json.loads(out)
+
+def nice_size(bytesize):
+    units = ["B", "K", "M", "G", "T"]
+    bs = float(bytesize)
+    u = 0
+    while bs > 1200:
+        bs = bs / 1024
+        u += 1
+    return "{:.2f}{}".format(bs, units[u])
+
+large_vids = []
+
+for root, dirs, files in os.walk("."):
+    for f in files:
+        ext = f.split(".")[-1]
+        if not ext in VIDEO_EXTS:
+            continue
+        filepath = "{}/{}".format(root, f)
+        meta = ffprobe(filepath)
+        #pprint(meta)
+        vmeta = None
+        if not "streams" in meta:
+            print("ERROR reading {}.".format(filepath), file=sys.stderr)
+            continue
+        for s in meta["streams"]:
+            if s["codec_type"] == "video":
+                vmeta = s
+        if vmeta is None:
+            print("No video stream found in {}.".format(filepath), file=sys.stderr)
+            continue
+        #pprint(vmeta)
+        smaller = vmeta["height"]
+        if vmeta["width"] < smaller:
+            smaller = vmeta["width"]
+
+        if smaller > MAX_HEIGHT:
+            record = {
+                "filepath": filepath,
+                "width": vmeta["width"],
+                "height": vmeta["height"],
+                "filesize": meta["format"]["size"],
+            }
+            large_vids.append(record)
+            print("+", end="", file=sys.stderr, flush=True)
+        else:
+            print(".", end="", file=sys.stderr, flush=True)
+
+print(" done.", file=sys.stderr)
+
+# Sort in descending order with largest file first
+large_vids.sort(key=lambda x: int(x["filesize"]), reverse=True)
+
+#pprint(large_vids)
+
+for v in large_vids:
+    print(v["filepath"])
+    print("{:=4}x{:=4} {} {}".format(v["width"], v["height"], nice_size(v["filesize"]), v["filepath"]), file=sys.stderr)
diff --git a/find720.sh b/find720.sh
deleted file mode 100755
index 3452ba7..0000000
--- a/find720.sh
+++ /dev/null
@@ -1,15 +0,0 @@
-#!/bin/bash
-# Find video files with >720 height
-find . -type f \( -iname "*.wmv" -or -iname "*.mp4" -or -iname "*.avi" -or -iname "*.mkv" -or -iname "*.flv" -or -iname "*.mov" -or -iname "*.m4v" -or -iname "*.webm" \) -print0 | while read -d $'\0' i; do
-    DIM=$(ffprobe -v error -select_streams v:0 -show_entries stream=width,height -of csv=s=x:p=0 "$i")
-    DIMS=($(echo $DIM | tr "x" " "))
-    NARROW=${DIMS[0]}
-    if [ ${DIMS[1]} -lt ${DIMS[0]} ]; then
-        NARROW=${DIMS[1]}
-    fi
-    if [ $NARROW -gt 720 ]; then
-        echo "$i"
-    else
-        echo -n "." > /dev/stderr
-    fi
-done