yta-video-opengl 0.0.11__py3-none-any.whl → 0.0.13__py3-none-any.whl

yta_video_opengl/utils.py

@@ -1,8 +1,10 @@
+from yta_video_opengl.t import fps_to_time_base
 from yta_validation import PythonValidator
 from av.container import InputContainer
 from av.video.stream import VideoStream
 from av.audio.stream import AudioStream
 from av.video.frame import VideoFrame
+from quicktions import Fraction
 from typing import Union
 
 import av
@@ -310,9 +312,13 @@ def iterate_stream_frames_demuxing(
                 pts_to_index(frame.pts, time_base, average_rate)
             )
 
+# TODO: These methods below have to be
+# removed because we created the new T
+# class that is working with Fractions
+# to be precise
 def t_to_pts(
-    t: float,
-    stream_time_base: 'Fraction'
+    t: Union[int, float, Fraction],
+    stream_time_base: Fraction
  ) -> int:
     """
     Transform a 't' time moment (in seconds) to
@@ -323,8 +329,8 @@ def t_to_pts(
 
 def pts_to_index(
     pts: int,
-    stream_time_base: 'Fraction',
-    fps: float
+    stream_time_base: Fraction,
+    fps: Union[float, Fraction]
 ) -> int:
     """
     Transform a 'pts' packet timestamp to a 
@@ -334,8 +340,8 @@ def pts_to_index(
 
 def index_to_pts(
     index: int,
-    stream_time_base: 'Fraction',
-    fps: float
+    stream_time_base: Fraction,
+    fps: Union[float, Fraction]
 ) -> int:
     """
     Transform a frame index into a 'pts' packet
@@ -345,10 +351,165 @@ def index_to_pts(
 
 def pts_to_t(
     pts: int,
-    stream_time_base: 'Fraction'
+    stream_time_base: Fraction
 ) -> float:
     """
     Transform a 'pts' packet timestamp to a 't'
     time moment.
     """
-    return pts * stream_time_base
+    return pts * stream_time_base
+
+
+# TODO: Move this to another utils
+def get_silent_audio_frame(
+    sample_rate: int,
+    layout = 'stereo',
+    number_of_samples: int = 1024,
+    format = 's16'
+):
+    # TODO: This could be a utils or something to
+    # directly transform format into dtype
+    dtype = {
+        's16': np.int16,
+        'flt': np.float32,
+        'fltp': np.float32
+    }.get(format, None)
+
+    if dtype is None:
+        raise Exception(f'The format "{format}" is not accepted.')
+
+    number_of_channels = len(av.AudioLayout(layout).channels)
+    # TODO: I think the option above is better
+    # number_of_channels = (
+    #     2
+    #     if layout == 'stereo' else
+    #     1
+    # )
+
+    # For packed (or planar) formats we apply:
+    # (1, samples * channels). This is the same
+    # amount of data but planar, in 1D only
+    # TODO: This wasn't in the previous version
+    # and it was working, we were sending the
+    # same 'number_of_samples' even when 'fltp'
+    # that includes the 'p'
+    # TODO: This is making the audio last 2x
+    # if 'p' in format:
+    #     number_of_samples *= number_of_channels
+
+    silent_array = np.zeros((number_of_channels, number_of_samples), dtype = dtype)
+    frame = av.AudioFrame.from_ndarray(silent_array, format = format, layout = layout)
+    frame.sample_rate = sample_rate
+    
+    return frame
+
+def get_black_background_video_frame(
+    size: tuple[int, int] = (1920, 1080),
+    format: str = 'rgb24',
+    pts: Union[int, None] = None,
+    time_base: Union[Fraction, None] = None
+):
+    """
+    Get a pyav VideoFrame that is a completely black
+    frame. If the 'pts' and/or 'time_base' parameters
+    are provided, they will be set to the frame that
+    is returned with them. If not, remember to set
+    later because they are needed to be sent to the
+    pyav muxer.
+    """
+    frame = av.VideoFrame.from_ndarray(
+        # TODO: What if we want alpha (?)
+        # Size must be inverted
+        array = np.zeros((size[1], size[0], 3), dtype = np.uint8),
+        format = format
+    )
+
+    frame.pts = (
+        pts
+        if pts is not None else
+        frame.pts
+    )
+
+    frame.time_base = (
+        time_base
+        if time_base is not None else
+        frame.time_base
+    )
+
+    return frame
+
+def get_audio_frame_pts_range(
+    frame: av.AudioFrame,
+    do_in_seconds: bool = False
+):
+    """
+    Get the [start_pts, end_pts) range of the
+    pyav AudioFrame, or in seconds if the
+    'do_in_seconds' parameter is True.
+    """
+    if frame.pts is None:
+        raise Exception('No "pts" found.')
+
+    # First and last sample. Remember that
+    # the last one is not included
+    start_pts = frame.pts
+    end_pts = frame.pts + frame.samples
+
+    # Time base for the seconds conversion
+    time_base = (
+        frame.time_base
+        if frame.time_base else
+        Fraction(1, frame.sample_rate)
+    )
+
+    start_time = (
+        float(start_pts * time_base)
+        if do_in_seconds else
+        start_time
+    )
+    end_time = (
+        float(end_pts * time_base)
+        if do_in_seconds else
+        end_time
+    )
+
+    return (
+        start_time,
+        end_time
+    )
+
+def audio_frames_and_remainder_per_video_frame(
+    # TODO: Maybe force 'fps' as int (?)
+    video_fps: Union[float, Fraction],
+    sample_rate: int, # audio_fps
+    number_of_samples_per_audio_frame: int
+) -> tuple[int, int]:
+    """
+    Get how many full silent audio frames we
+    need and the remainder for the last one
+    (that could be not complete), according
+    to the parameters provided.
+
+    This method returns a tuple containing
+    the number of full silent audio frames
+    we need and the number of samples we need
+    in the last non-full audio frame.
+    """
+    # Video frame duration (in seconds)
+    time_base = fps_to_time_base(video_fps)
+    sample_rate = Fraction(int(sample_rate), 1)
+
+    # Example:
+    # 44_100 / 60 = 735  ->  This means that we
+    # will have 735 samples of sound per each
+    # video frame
+    # The amount of samples per frame is actually
+    # the amount of samples we need, because we
+    # are generating it...
+    samples_per_frame = sample_rate * time_base
+    # The 'nb_samples' is the amount of samples
+    # we are including on each audio frame
+    full_audio_frames_needed = samples_per_frame // number_of_samples_per_audio_frame
+    remainder = samples_per_frame % number_of_samples_per_audio_frame
+    
+    return int(full_audio_frames_needed), int(remainder)

yta_video_opengl/video.py

@@ -1,7 +1,8 @@
 from yta_video_opengl.reader import VideoReader
 from yta_video_opengl.writer import VideoWriter
-from yta_video_opengl.utils import iterate_stream_frames_demuxing
+from yta_video_opengl.t import T
 from yta_validation import PythonValidator
+from quicktions import Fraction
 from typing import Union
 
 
@@ -26,7 +27,8 @@ class Video:
         This timestamp is used to read the video
         file source.
         """
-        return int(self.start / self.reader.time_base)
+        # TODO: What if 'time_base' is None (?)
+        return T(self.start, self.reader.time_base).truncated_pts
     
     @property
     def end_pts(
@@ -40,7 +42,8 @@ class Video:
         file source.
         """
         return (
-            int(self.end / self.reader.time_base)
+            # TODO: What if 'time_base' is None (?)
+            T(self.end, self.reader.time_base).truncated_pts
             # TODO: What do we do if no duration (?)
             if self.duration is not None else
             None
@@ -54,7 +57,8 @@ class Video:
         The start packet time stamp (pts), needed 
         to optimize the packet iteration process.
         """
-        return int(self.start / self.reader.audio_time_base)
+        # TODO: What if 'audio_time_base' is None (?)
+        return T(self.start, self.reader.audio_time_base).truncated_pts
     
     @property
     def audio_end_pts(
@@ -65,7 +69,8 @@ class Video:
         optimize the packet iteration process.
         """
         return (
-            int(self.end / self.reader.audio_time_base)
+            # TODO: What if 'audio_time_base' is None (?)
+            T(self.end, self.reader.audio_time_base).truncated_pts
             # TODO: What do we do if no duration (?)
             if self.duration is not None else
             None
@@ -74,7 +79,7 @@ class Video:
     @property
     def duration(
         self
-    ):
+    ) -> Fraction:
         """
         The duration of the video.
         """
@@ -83,7 +88,7 @@ class Video:
     @property
     def number_of_frames(
         self
-    ):
+    ) -> Union[int, None]:
         """
         The number of frames of the video.
         """
@@ -117,8 +122,8 @@ class Video:
     def __init__(
         self,
         filename: str,
-        start: float = 0.0,
-        end: Union[float, None] = None
+        start: Union[int, float, Fraction] = 0.0,
+        end: Union[int, float, Fraction, None] = None
     ):
         self.filename: str = filename
         """
@@ -130,12 +135,12 @@ class Video:
         """
         The pyav video reader.
         """
-        self.start: float = start
+        self.start: Fraction = Fraction(start)
         """
         The time moment 't' in which the video
         should start.
         """
-        self.end: Union[float, None] = (
+        self.end: Union[Fraction, None] = Fraction(
             # TODO: Is this 'end' ok (?)
             self.reader.duration
             if end is None else
@@ -145,12 +150,80 @@ class Video:
         The time moment 't' in which the video
         should end.
         """
+
+    # TODO: We need to implement the 'get_frame'
+    # methods because this Video can be subclipped
+    # and have a 'start' and' end' that are 
+    # different from [0, end)
+    def _get_real_t(
+        self,
+        t: Union[int, float, Fraction]
+    ) -> Fraction:
+        """
+        Get the real 't' time moment based on the
+        video 'start' and 'end'. If they were 
+        asking for the t=0.5s but our video was
+        subclipped to [1.0, 2.0), the 0.5s must be
+        actually the 1.5s of the video because of
+        the subclipped time range.
+        """
+        t = self.start + t
         
+        if t >= self.end:
+            raise Exception(f'The "t" ({str(t)}) provided is out of range. This video lasts from [{str(self.start)}, {str(self.end)}).')
+        
+        return t
+
+    def get_frame_from_t(
+        self,
+        t: Union[int, float, Fraction]
+    ) -> 'VideoFrame':
+        """
+        Get the video frame with the given 't' time
+        moment, using the video cache system.
+        """
+        return self.reader.video_cache.get_frame_from_t(self._get_real_t(t))
+    
+    def get_audio_frame_from_t(
+        self,
+        t: Union[int, float, Fraction]
+    ) -> 'AudioFrame':
+        """
+        Get the audio frame with the given 't' time
+        moment, using the audio cache system.
+        """
+        return self.reader.audio_cache.get_frame_from_t(self._get_real_t(t))
+    
+    def get_audio_frames_from_t(
+        self,
+        t: Union[int, float, Fraction]
+    ):
+        """
+        Get the sequence of audio frames for the 
+        given video 't' time moment, using the
+        audio cache system.
+
+        This is useful when we want to write a
+        video frame with its audio, so we obtain
+        all the audio frames associated to it
+        (remember that a video frame is associated
+        with more than 1 audio frame).
+        """
+        for frame in self.reader.get_audio_frames_from_t(self._get_real_t(t)):
+            yield frame
+
     def save_as(
         self,
         filename: str
     ) -> 'Video':
-        writer =  VideoWriter(filename)
+        """
+        Save the video locally as the given 'filename'.
+
+        TODO: By now we are doing tests inside so the
+        functionality is a manual test. Use it 
+        carefully.
+        """
+        writer = VideoWriter(filename)
         writer.set_video_stream_from_template(self.reader.video_stream)
         writer.set_audio_stream_from_template(self.reader.audio_stream)
 

yta_video_opengl/writer.py

@@ -7,7 +7,7 @@ from av.video.stream import VideoStream
 from av.audio.stream import AudioStream
 from av.container.output import OutputContainer
 from av import open as av_open
-from fractions import Fraction
+from quicktions import Fraction
 from typing import Union
 
 
@@ -43,6 +43,7 @@ class VideoWriter:
     def set_video_stream(
         self,
         codec_name: Union[str, None],
+        # TODO: Maybe force fps to 'int' (?)
         fps: Union[Fraction, int, float, None],
         size: Union[tuple[int, int], None] = None,
         pixel_format: Union[str, None] = None,
@@ -52,6 +53,10 @@ class VideoWriter:
         Set the video stream, that will overwrite any other
         previous video stream set.
         """
+        # TODO: They say that pyav accepts fps as
+        # Fraction...
+        fps = int(fps)
+
         self.video_stream: VideoStream = self.output.add_stream(
             # TODO: Maybe 'libx264' as default 'codec_name' (?)
             codec_name = codec_name,
@@ -60,7 +65,7 @@ class VideoWriter:
         )
 
         # We need to force this or it will not work
-        self.video_stream.time_base = Fraction(1, int(fps))
+        self.video_stream.time_base = Fraction(1, fps)
 
         if size is not None:
             self.video_stream.width = size[0]
@@ -91,17 +96,27 @@ class VideoWriter:
 
     def set_audio_stream(
         self,
-        codec_name: Union[str, None]
+        codec_name: Union[str, None],
+        fps: Union[int, float, Fraction] = 44_100.0
         # TODO: Add more if needed
     ) -> 'VideoWriter':
         """
         Set the audio stream, that will overwrite any other
         previous audio stream set.
         """
+        # TODO: Check what else we can set
         self.audio_stream: AudioStream = self.output.add_stream(
-            codec_name = codec_name
+            codec_name = codec_name,
+            rate = int(fps)
         )
 
+        # audio_stream = output.add_stream("aac", rate=48000)  # codec AAC, 48kHz
+        # # Configurar stream
+        # audio_stream.channels = 2                # número de canales
+        # audio_stream.layout = "stereo"           # layout
+        # audio_stream.sample_rate = 48000         # sample rate
+        # audio_stream.format = "s16"              # formato de las muestras (PCM signed 16-bit)
+
         # TODO: Add more if needed
 
         return self
@@ -127,16 +142,6 @@ class VideoWriter:
 
         return self
 
-        # This below is not working
-        self.audio_stream: AudioStream = self.output.add_stream_from_template(
-            template
-        )
-        # TODO: Is this actually needed (?)
-        # Force this 'rate'
-        self.audio_stream.time_base = Fraction(1, template.codec_context.rate)
-
-        return self
-
     def encode_video_frame(
         self,
         frame: Union[VideoFrame, None] = None
@@ -186,6 +191,10 @@ class VideoWriter:
         """
         ParameterValidator.validate_mandatory_instance_of('packet', packet, Packet)
 
+        # We are ignoring empty packets because they
+        # are used to indicate the end or things like
+        # that, not actual data... But maybe we are 
+        # wrong...
         if packet.size > 0:
             try:
                 self.output.mux(packet)

{yta_video_opengl-0.0.11.dist-info → yta_video_opengl-0.0.13.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.3
 Name: yta-video-opengl
-Version: 0.0.11
+Version: 0.0.13
 Summary: Youtube Autonomous Video OpenGL Module
 Author: danialcala94
 Author-email: danielalcalavalera@gmail.com
@@ -10,6 +10,7 @@ Classifier: Programming Language :: Python :: 3.9
 Requires-Dist: av (>=0.0.1,<19.0.0)
 Requires-Dist: moderngl (>=0.0.1,<9.0.0)
 Requires-Dist: numpy (>=0.0.1,<9.0.0)
+Requires-Dist: quicktions (>=0.0.1,<9.0.0)
 Requires-Dist: yta_timer (>0.0.1,<1.0.0)
 Requires-Dist: yta_validation (>=0.0.1,<1.0.0)
 Requires-Dist: yta_video_frame_time (>=0.0.1,<1.0.0)

yta_video_opengl-0.0.13.dist-info/RECORD

@@ -0,0 +1,21 @@
+yta_video_opengl/__init__.py,sha256=ycAx_XYMVDfkuObSvtW6irQ0Wo-fgxEz3fjIRMe8PpY,205
+yta_video_opengl/classes.py,sha256=t5-Tfc7ecvHl8JlVBp_FVzZT6ole6Ly5-FeBBH7wcxo,37742
+yta_video_opengl/complete/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+yta_video_opengl/complete/timeline.py,sha256=drIeM1xYv-8OM-uwjUifwA5teT5X9LZOEK5Nfzux4FI,9384
+yta_video_opengl/complete/track.py,sha256=MQJPFyZpl1XFRwYqfq32QzzEESCVtPcnz9yhwm7crng,13729
+yta_video_opengl/complete/video_on_track.py,sha256=iDD3wyEHblbRxm1e-clMecx104wUfUwKdCq7ZXSq6CU,5127
+yta_video_opengl/nodes/__init__.py,sha256=TZ-ZO05PZ0_ABq675E22_PngLWOe-_w5s1cLlV3NbWM,3469
+yta_video_opengl/nodes/audio/__init__.py,sha256=4nKkC70k1UgLcCSPqFWm3cKdaJM0KUmQTwGWv1xFarQ,2926
+yta_video_opengl/nodes/video/__init__.py,sha256=gSoaoEmjdQmyRwH18mf5z3NAhap3S0RgbeBbfBXi4jc,132
+yta_video_opengl/nodes/video/opengl.py,sha256=K2pyCJEd9z4gnZqJetKyGPbtHuBzFsx74ZYyzhSqYPo,8510
+yta_video_opengl/reader/__init__.py,sha256=Cxk838QMjIUOTiO2ZyPGkiPojHlX0c7x6iJkm81_UB4,20039
+yta_video_opengl/reader/cache.py,sha256=YYx5AX0BkGj1KtiDePHKj-AZlET4ATTqPB0xbr98WOw,17742
+yta_video_opengl/t.py,sha256=vK23hBQQt_KtBO3ceLQ-SwQbB2jfFsPFX2pYNjYaVwk,5488
+yta_video_opengl/tests.py,sha256=EdTyYtTUd_mj6geWnrvnF-wZSHCKKvhYgiLclkV73O0,26576
+yta_video_opengl/utils.py,sha256=yUi17EjNR4SVpvdDUwUaKl4mBCb1uyFCSGoIX3Zr2F0,15586
+yta_video_opengl/video.py,sha256=ap3H-aA29Yj-SAsTKsNq2gotcBUatQVQ2jmplgKMm9Q,8465
+yta_video_opengl/writer.py,sha256=QwvjQcEkzn1WAVqVTFiI6tYIXJO67LKKUTJGO_eflFM,8893
+yta_video_opengl-0.0.13.dist-info/LICENSE,sha256=6kbiFSfobTZ7beWiKnHpN902HgBx-Jzgcme0SvKqhKY,1091
+yta_video_opengl-0.0.13.dist-info/METADATA,sha256=DvAvfeMY-o4WMVKdk5d8TiphrseGwMFadDrinfE0wh8,714
+yta_video_opengl-0.0.13.dist-info/WHEEL,sha256=XbeZDeTWKc1w7CSIyre5aMDU_-PohRwTQceYnisIYYY,88
+yta_video_opengl-0.0.13.dist-info/RECORD,,

yta_video_opengl-0.0.11.dist-info/RECORD

@@ -1,20 +0,0 @@
-yta_video_opengl/__init__.py,sha256=ycAx_XYMVDfkuObSvtW6irQ0Wo-fgxEz3fjIRMe8PpY,205
-yta_video_opengl/classes.py,sha256=t5-Tfc7ecvHl8JlVBp_FVzZT6ole6Ly5-FeBBH7wcxo,37742
-yta_video_opengl/complete/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-yta_video_opengl/complete/timeline.py,sha256=o6nAtpkdr6RzYlIEB9kLfNAZiu9eRXyA6QE3TXuNL_I,5465
-yta_video_opengl/complete/track.py,sha256=gPushEuLYIoMBjvS3wTDbOKSV8VEtVTqrEm666ABFUc,3937
-yta_video_opengl/complete/video_on_track.py,sha256=sTSd8gqNwK1wT1uzJEGkJqTbu5qM3wWqk361K7OAbl0,2851
-yta_video_opengl/nodes/__init__.py,sha256=TZ-ZO05PZ0_ABq675E22_PngLWOe-_w5s1cLlV3NbWM,3469
-yta_video_opengl/nodes/audio/__init__.py,sha256=4nKkC70k1UgLcCSPqFWm3cKdaJM0KUmQTwGWv1xFarQ,2926
-yta_video_opengl/nodes/video/__init__.py,sha256=gSoaoEmjdQmyRwH18mf5z3NAhap3S0RgbeBbfBXi4jc,132
-yta_video_opengl/nodes/video/opengl.py,sha256=K2pyCJEd9z4gnZqJetKyGPbtHuBzFsx74ZYyzhSqYPo,8510
-yta_video_opengl/reader/__init__.py,sha256=zqVMMUjWHCsOaWFrnkYdnbO4YEAx3zLIBTuMnzQJRug,17180
-yta_video_opengl/reader/cache.py,sha256=Gtwl2kRLSJn3JJOVI-45Wc2vMPYj5-_g3AR5wPbw8OY,9152
-yta_video_opengl/tests.py,sha256=p2Pq4o2H0DMZkV7HNNNAlebSjrDMHKTKk0d_weiiPHQ,26221
-yta_video_opengl/utils.py,sha256=_89-IrTDFVbY86qLlJzw42MxYAsNAnux_DOO89mul64,10710
-yta_video_opengl/video.py,sha256=3n7jgZab7PUSOpODoaH4iNg0sy7NMRo_OaJ4Zj8u0NM,5855
-yta_video_opengl/writer.py,sha256=rzOfxPtlClLRIVLFyIOoBVcEgse3ISmpKRWHg8OwFHQ,8355
-yta_video_opengl-0.0.11.dist-info/LICENSE,sha256=6kbiFSfobTZ7beWiKnHpN902HgBx-Jzgcme0SvKqhKY,1091
-yta_video_opengl-0.0.11.dist-info/METADATA,sha256=8JNEZ05Elie6lMQREjRjGPQQ-kLW-GCW6uZpePTiLgY,671
-yta_video_opengl-0.0.11.dist-info/WHEEL,sha256=XbeZDeTWKc1w7CSIyre5aMDU_-PohRwTQceYnisIYYY,88
-yta_video_opengl-0.0.11.dist-info/RECORD,,