View difference between Paste ID: EbHZq4uU and HVJkzHiL

-#!/usr/bin/env python3
+#!/usr/bin/env python3
-import subprocess
+import subprocess
-import tempfile
+import tempfile
-import os
+import os
-import os.path
+import os.path
-import collections
+import collections
-import json
+import json
-import sys
+import sys
-import multiprocessing
+import multiprocessing
-import struct
+import struct
-
+
-ProcessedTrack = collections.namedtuple("ProcessedTrack", ["dfpwm_file", "metadata"])
+ProcessedTrack = collections.namedtuple("ProcessedTrack", ["dfpwm_file", "metadata"])
-
+
-def convert_wav_dfpwm(infile, outfile):
+def convert_wav_dfpwm(infile, outfile):
-    subprocess.run(["java", "-jar", "LionRay.jar", infile, outfile])
+    subprocess.run(["java", "-jar", "LionRay.jar", infile, outfile])
-
+
-def convert_any_wav(infile, outfile):
+def convert_any_wav(infile, outfile):
-    subprocess.run(["ffmpeg", "-hide_banner", "-i", infile, "-ac", "1", outfile], stderr=subprocess.PIPE)
+    subprocess.run(["ffmpeg", "-hide_banner", "-i", infile, "-ac", "1", outfile], stderr=subprocess.PIPE)
-
+
-def read_meta(path):
+def read_meta(path):
-    proc = subprocess.run(["ffprobe", "-v", "quiet", "-print_format", "json", "-show_format", "-show_streams", path], stdout=subprocess.PIPE)
+    proc = subprocess.run(["ffprobe", "-v", "quiet", "-print_format", "json", "-show_format", "-show_streams", path], stdout=subprocess.PIPE)
-    data = json.loads(proc.stdout)
+    data = json.loads(proc.stdout)
-    meta = {}
+    meta = {}
-    # These are the two locations I've found tags in in my not very thorough testing
+    # These are the two locations I've found tags in in my not very thorough testing
-    try:
+    try:
-        meta.update(data["format"]["tags"])
+        meta.update(data["format"]["tags"])
-    except KeyError: pass
+    except KeyError: pass
-    try:
+    try:
-        meta.update(data["streams"][0]["tags"])
+        meta.update(data["streams"][0]["tags"])
-    except KeyError: pass
+    except KeyError: pass
-    # lowercase all keys because in Opus files these seem to be uppercase sometimes
+    # lowercase all keys because in Opus files these seem to be uppercase sometimes
-    return { k.lower(): v for k, v in meta.items() }
+    return { k.lower(): v for k, v in meta.items() }
-
+
-def process_file(filename):
+def process_file(filename):
-    meta = read_meta(filename)
+    meta = read_meta(filename)
-    wav_dest = tempfile.mktemp(".wav")
+    wav_dest = tempfile.mktemp(".wav")
-    convert_any_wav(filename, wav_dest)
+    convert_any_wav(filename, wav_dest)
-    dfpwm_dest = tempfile.mktemp(".dfpwm")
+    dfpwm_dest = tempfile.mktemp(".dfpwm")
-    convert_wav_dfpwm(wav_dest, dfpwm_dest)
+    convert_wav_dfpwm(wav_dest, dfpwm_dest)
-    os.remove(wav_dest)
+    os.remove(wav_dest)
-    print(filename)
+    print(filename)
-    return ProcessedTrack(dfpwm_dest, {
+    return ProcessedTrack(dfpwm_dest, {
-        "title": meta["title"],
+        "title": meta["title"],
-        "artist": meta.get("artist", None) or meta.get("artists", None),
+        "artist": meta.get("artist", None) or meta.get("artists", None),
-        "album": meta.get("album", None)
+        "album": meta.get("album", None)
-    })
+    })
-
+
-def read_binary(filename):
+def read_binary(filename):
-    with open(filename, "rb") as f:
+    with open(filename, "rb") as f:
-        return f.read()
+        return f.read()
-
+
-def process_dir(dirname):
+def process_dir(dirname):
-    files = list(map(lambda file: os.path.join(dirname, file), os.listdir(dirname)))
+    files = list(map(lambda file: os.path.join(dirname, file), os.listdir(dirname)))
-    with multiprocessing.Pool(8) as p:
+    with multiprocessing.Pool(8) as p:
-        tracks = p.map(process_file, files)
+        tracks = p.map(process_file, files)
-    tape_image = b""
+    tape_image = b""
-    tracks_meta = []
+    tracks_meta = []
-    for track in tracks:
+    for track in tracks:
-        track.metadata["start"] = len(tape_image)
+        track.metadata["start"] = len(tape_image)
-        data = read_binary(track.dfpwm_file)
+        data = read_binary(track.dfpwm_file)
-        os.remove(track.dfpwm_file)
+        os.remove(track.dfpwm_file)
-        track.metadata["end"] = track.metadata["start"] + len(data)
+        track.metadata["end"] = track.metadata["start"] + len(data)
-        tape_image += data
+        tape_image += data
-        tracks_meta.append(track.metadata)
+        tracks_meta.append(track.metadata)
-    meta = json.dumps({ "tracks": tracks_meta }).encode("utf-8")
+    # dump in a compact format to save space
-    assert(len(meta) < 65536)
+    meta = json.dumps({ "tracks": tracks_meta }, separators=(',', ':')).encode("utf-8")
-    # new format - 0x54 marker byte, then metadata length as 2-byte big endian integer, then metadata, then concatenated DFPWM files
+    assert(len(meta) < 65536)
-    # start is now not an absolute position but just how far after the metadata it is
+    # new format - 0x54 marker byte, then metadata length as 2-byte big endian integer, then metadata, then concatenated DFPWM files
-    tape_image = b"\x54" + struct.pack(">H", len(meta)) + meta + tape_image
+    # start is now not an absolute position but just how far after the metadata it is
-    with open("tape.bin", "wb") as f:
+    tape_image = b"\x54" + struct.pack(">H", len(meta)) + meta + tape_image
-        f.write(tape_image)
+    with open("tape.bin", "wb") as f:
-
+        f.write(tape_image)
     # Tape lengths are measured in minutes. 6000 bytes are played per second because they use a 48000Hz sample rate and DFPWM is somehow 1 bit per sample.
     length_minutes = len(tape_image) / (6000*60)
     print(length_minutes, "minute tape required")
 
 process_dir(sys.argv[1])