rk_pve/mp4/libav.py

# People's Video Editor: high quality, GPU accelerated mp4 editor
# Copyright (C) 2025 Roz K <roz@rozk.net>
#
# This file is part of People's Video Editor.
#
# People's Video Editor is free software: you can redistribute it and/or modify it under the terms of the
# GNU General Public License as published by the Free Software Foundation, either version 3 of the License,
# or (at your option) any later version.
#
# People's Video Editor is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY;
# without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
# See the GNU General Public License for more details.
#
# You should have received a copy of the GNU General Public License along with People's Video Editor.
# If not, see <https://www.gnu.org/licenses/>.

import errno
import ctypes

_avutil = ctypes.cdll.LoadLibrary('libavutil.so')
_avformat = ctypes.cdll.LoadLibrary('libavformat.so')
_avcodec = ctypes.cdll.LoadLibrary('libavcodec.so')

def _errtag(a, b, c, d):
    return -(ord(a) | (ord(b) << 8) | (ord(c) << 16) | (ord(d) << 24))

if errno.EAGAIN < 0:
    AVERROR_EAGAIN = errno.EAGAIN
else:
    AVERROR_EAGAIN = -errno.EAGAIN

AVERROR_EOF = _errtag('E', 'O', 'F', ' ')

_AV_ERROR_MAX_STRING_SIZE = 64

AVMEDIA_TYPE_UNKNOWN = -1
AVMEDIA_TYPE_VIDEO = 0
AVMEDIA_TYPE_AUDIO = 1
AVMEDIA_TYPE_DATA = 2
AVMEDIA_TYPE_SUBTITLE = 3
AVMEDIA_TYPE_ATTACHMENT = 4

AV_NUM_DATA_POINTERS = 8

c_uint8_p = ctypes.POINTER(ctypes.c_uint8)
c_uint8_pp = ctypes.POINTER(c_uint8_p)

class AVRational(ctypes.Structure):
    _fields_ = [
        ("num", ctypes.c_int),
        ("den", ctypes.c_int)]

class AVFrame(ctypes.Structure):
    _fields_ = [
        ("data", c_uint8_p * AV_NUM_DATA_POINTERS),
        ("linesize", ctypes.c_int * AV_NUM_DATA_POINTERS),
        ("extended_data", c_uint8_pp),
        ("width", ctypes.c_int),
        ("height", ctypes.c_int),
        ("nb_samples", ctypes.c_int),
        ("format", ctypes.c_int),
        ("key_frame", ctypes.c_int),
        ("pict_type", ctypes.c_int),
        ("sample_aspect_ratio", AVRational),
        ("pts", ctypes.c_int64),
        ("pkt_dts", ctypes.c_int64),
        ("time_base", AVRational)]
        # ...

AVFrame_p = ctypes.POINTER(AVFrame)
AVFrame_pp = ctypes.POINTER(AVFrame_p)

class AVCodecParameters(ctypes.Structure):
    pass

AVCodecParameters_p = ctypes.POINTER(AVCodecParameters)

class AVStream(ctypes.Structure):
    _fields_ = [
        ("av_class", ctypes.c_void_p),
        ("index", ctypes.c_int),
        ("id", ctypes.c_int),
        ("codecpar", AVCodecParameters_p),
        ("priv_data", ctypes.c_void_p),
        ("time_base", AVRational)]
        # ...

AVStream_p = ctypes.POINTER(AVStream)
AVStream_pp = ctypes.POINTER(AVStream_p)

class AVFormatContext(ctypes.Structure):
    _fields_ = [
        ("av_class", ctypes.c_void_p),
        ("iformat", ctypes.c_void_p),
        ("oformat", ctypes.c_void_p),
        ("priv_data", ctypes.c_void_p),
        ("pb", ctypes.c_void_p),
        ("ctx_flags", ctypes.c_int),
        ("nb_streams", ctypes.c_uint),
        ("streams", AVStream_pp)]
        # ...

AVFormatContext_p = ctypes.POINTER(AVFormatContext)
AVFormatContext_pp = ctypes.POINTER(AVFormatContext_p)

class AVPacket(ctypes.Structure):
    _fields_ = [
        ("buf", ctypes.c_void_p),
        ("pts", ctypes.c_int64),
        ("dts", ctypes.c_int64),
        ("data", ctypes.c_void_p),
        ("size", ctypes.c_int),
        ("stream_index", ctypes.c_int)]
        # ...

AVPacket_p = ctypes.POINTER(AVPacket)
AVPacket_pp = ctypes.POINTER(AVPacket_p)

class AVCodec(ctypes.Structure):
    _fields_ = [
        ("name", ctypes.c_char_p),
        ("long_name", ctypes.c_char_p)]
        # ...

AVCodec_p = ctypes.POINTER(AVCodec)
AVCodec_pp = ctypes.POINTER(AVCodec_p)

class AVCodecContext(ctypes.Structure):
    pass

AVCodecContext_p = ctypes.POINTER(AVCodecContext)
AVCodecContext_pp = ctypes.POINTER(AVCodecContext_p)

_avutil.av_strerror.restype = ctypes.c_int
_avutil.av_strerror.argtypes = [
    ctypes.c_int, # errno
    ctypes.c_char_p, # errbuf
    ctypes.c_size_t] # errbuff_size

def strerror(errno):
    errbuf = ctypes.create_string_buffer(_AV_ERROR_MAX_STRING_SIZE)
    _avutil.av_strerror(errno, errbuf, _AV_ERROR_MAX_STRING_SIZE)
    return errbuf.value.decode("utf-8")

_avutil.av_frame_alloc.restype = AVFrame_p
_avutil.av_frame_alloc.argtypes = None

def frame_alloc():
    return _avutil.av_frame_alloc()

_avutil.av_frame_free.restype = None
_avutil.av_frame_free.argtypes = [AVFrame_pp]

def frame_free(frame):
    _avutil.av_frame_free(ctypes.byref(frame))

_avformat.avformat_alloc_context.restype = AVFormatContext_p
_avformat.avformat_alloc_context.argtypes = None

def format_alloc_context():
    return _avformat.avformat_alloc_context()

_avformat.avformat_free_context.restype = None
_avformat.avformat_free_context.argtypes = [AVFormatContext_p]

def format_free_context(context):
    _avformat.avformat_free_context(context)

_avformat.avformat_open_input.restype = ctypes.c_int
_avformat.avformat_open_input.argtypes = [
    AVFormatContext_pp,
    ctypes.c_char_p, # url
    ctypes.c_void_p, # format
    ctypes.POINTER(ctypes.c_void_p)] # options

def format_open_input(context, url):
    return _avformat.avformat_open_input(ctypes.byref(context), url.encode('ascii', 'ignore'), None, None)

_avformat.avformat_close_input.restype = None
_avformat.avformat_close_input.argtypes = [AVFormatContext_pp]

def format_close_input(context):
    _avformat.avformat_close_input(ctypes.byref(context))

_avformat.avformat_find_stream_info.restype = ctypes.c_int
_avformat.avformat_find_stream_info.argtypes = [
    AVFormatContext_p,
    ctypes.POINTER(ctypes.c_void_p)] # options

def format_find_stream_info(context):
    return _avformat.avformat_find_stream_info(context, None)

_avformat.av_find_best_stream.restype = ctypes.c_int
_avformat.av_find_best_stream.argtypes = [
    AVFormatContext_p,
    ctypes.c_int, # type
    ctypes.c_int, # wanted stream
    ctypes.c_int, # related stream
    AVCodec_pp,
    ctypes.c_int] # flags

def format_find_best_stream(context, type):
    codec = AVCodec_p()
    index = _avformat.av_find_best_stream(context, type, -1, -1, ctypes.byref(codec), 0)
    return index, codec

_avformat.av_packet_alloc.restype = AVPacket_p
_avformat.av_packet_alloc.argtypes = None

def packet_alloc():
    return _avformat.av_packet_alloc()

_avformat.av_packet_free.restype = None
_avformat.av_packet_free.argtypes = [AVPacket_pp]

def packet_free(packet):
    _avformat.av_packet_free(ctypes.byref(packet))

_avformat.av_read_frame.restype = ctypes.c_int
_avformat.av_read_frame.argtypes = [AVFormatContext_p, AVPacket_p]

def read_frame(context, packet):
    return _avformat.av_read_frame(context, packet)

_avcodec.avcodec_alloc_context3.restype = AVCodecContext_p
_avcodec.avcodec_alloc_context3.argtypes = [AVCodec_p]

def codec_alloc_context(codec):
    return _avcodec.avcodec_alloc_context3(codec)

_avcodec.avcodec_free_context.restype = None
_avcodec.avcodec_free_context.argtypes = [AVCodecContext_pp]

def codec_free_context(context):
    _avcodec.avcodec_free_context(ctypes.byref(context))

_avcodec.avcodec_parameters_to_context.restype = ctypes.c_int
_avcodec.avcodec_parameters_to_context.argtypes = [AVCodecContext_p, AVCodecParameters_p]

def codec_parameters_to_context(context, parameters):
    return _avcodec.avcodec_parameters_to_context(context, parameters)

_avcodec.avcodec_open2.restype = ctypes.c_int
_avcodec.avcodec_open2.argtypes = [
    AVCodecContext_p,
    AVCodec_p,
    ctypes.POINTER(ctypes.c_void_p)] # options

def codec_open(context, codec):
    return _avcodec.avcodec_open2(context, codec, None)

_avcodec.avcodec_send_packet.restype = ctypes.c_int
_avcodec.avcodec_send_packet.argtypes = [AVCodecContext_p, AVPacket_p]

def codec_send_packet(context, packet):
    return _avcodec.avcodec_send_packet(context, packet)

_avcodec.avcodec_receive_frame.restype = ctypes.c_int
_avcodec.avcodec_receive_frame.argtypes = [AVCodecContext_p, AVFrame_p]

def codec_receive_frame(context, frame):
    return _avcodec.avcodec_receive_frame(context, frame)