vidformer 0.6.0__py3-none-any.whl → 0.6.2__py3-none-any.whl

vidformer/__init__.py

@@ -1,5 +1,5 @@
 """A Python library for creating and viewing videos with vidformer."""
 
-__version__ = "0.6.0"
+__version__ = "0.6.2"
 
 from .vf import *

vidformer/cv2/vf_cv2.py

@@ -1,5 +1,24 @@
+"""
+vidformer.cv2 is the cv2 frontend for [vidformer](https://github.com/ixlab/vidformer).
+
+> ⚠️ This module is a work in progress. See the [implemented functions list](https://ixlab.github.io/vidformer/opencv-filters.html).
+
+**Quick links:**
+* [📦 PyPI](https://pypi.org/project/vidformer/)
+* [📘 Documentation - vidformer-py](https://ixlab.github.io/vidformer/vidformer-py/)
+* [📘 Documentation - vidformer.cv2](https://ixlab.github.io/vidformer/vidformer-py-cv2/)
+* [🧑‍💻 Source Code](https://github.com/ixlab/vidformer/tree/main/vidformer-py/)
+"""
+
 from .. import vf
 
+try:
+    import cv2 as _opencv2
+except:
+    _opencv2 = None
+
+import numpy as np
+
 import uuid
 from fractions import Fraction
 from bisect import bisect_right
@@ -9,6 +28,7 @@ CAP_PROP_POS_FRAMES = 1
 CAP_PROP_FRAME_WIDTH = 3
 CAP_PROP_FRAME_HEIGHT = 4
 CAP_PROP_FPS = 5
+CAP_PROP_FRAME_COUNT = 7
 
 FONT_HERSHEY_SIMPLEX = 0
 FONT_HERSHEY_PLAIN = 1
@@ -28,6 +48,10 @@ LINE_AA = 16
 _filter_scale = vf.Filter("Scale")
 _filter_rectangle = vf.Filter("cv2.rectangle")
 _filter_putText = vf.Filter("cv2.putText")
+_filter_arrowedLine = vf.Filter("cv2.arrowedLine")
+_filter_line = vf.Filter("cv2.line")
+_filter_circle = vf.Filter("cv2.circle")
+_filter_addWeighted = vf.Filter("cv2.addWeighted")
 
 
 def _ts_to_fps(timestamps):
@@ -49,16 +73,18 @@ def _server():
     return _global_cv2_server
 
 
-def set_cv2_server(server):
+def set_cv2_server(server: vf.YrdenServer):
     """Set the server to use for the cv2 frontend."""
     global _global_cv2_server
     assert isinstance(server, vf.YrdenServer)
     _global_cv2_server = server
 
 
-class _Frame:
-    def __init__(self, f):
+class Frame:
+    def __init__(self, f, fmt):
         self._f = f
+        self._fmt = fmt
+        self.shape = (fmt["height"], fmt["width"], 3)
 
         # denotes that the frame has not yet been modified
         # when a frame is modified, it is converted to rgb24 first
@@ -67,6 +93,23 @@ class _Frame:
     def _mut(self):
         self._modified = True
         self._f = _filter_scale(self._f, pix_fmt="rgb24")
+        self._fmt["pix_fmt"] = "rgb24"
+
+    def numpy(self):
+        """
+        Return the frame as a numpy array.
+        """
+
+        self._mut()
+        spec = vf.Spec([Fraction(0, 1)], lambda t, i: self._f, self._fmt)
+        loader = spec.load(_server())
+
+        frame_raster_rgb24 = loader[0]
+        assert type(frame_raster_rgb24) == bytes
+        assert len(frame_raster_rgb24) == self.shape[0] * self.shape[1] * 3
+        raw_data_array = np.frombuffer(frame_raster_rgb24, dtype=np.uint8)
+        frame = raw_data_array.reshape(self.shape)
+        return frame
 
 
 class VideoCapture:
@@ -86,6 +129,10 @@ class VideoCapture:
             return self._source.fmt()["width"]
         elif prop == CAP_PROP_FRAME_HEIGHT:
             return self._source.fmt()["height"]
+        elif prop == CAP_PROP_FRAME_COUNT:
+            return len(self._source.ts())
+        elif prop == CAP_PROP_POS_FRAMES:
+            return self._next_frame_idx
 
         raise Exception(f"Unknown property {prop}")
 
@@ -106,7 +153,7 @@ class VideoCapture:
             return False, None
         frame = self._source.iloc[self._next_frame_idx]
         self._next_frame_idx += 1
-        frame = _Frame(frame)
+        frame = Frame(frame, self._source.fmt())
         return True, frame
 
     def release(self):
@@ -116,6 +163,8 @@ class VideoCapture:
 class VideoWriter:
     def __init__(self, path, fourcc, fps, size):
         assert isinstance(fourcc, VideoWriter_fourcc)
+        if path is not None and not isinstance(path, str):
+            raise Exception("path must be a string or None")
         self._path = path
         self._fourcc = fourcc
         self._fps = fps
@@ -125,8 +174,8 @@ class VideoWriter:
         self._pix_fmt = "yuv420p"
 
     def write(self, frame):
-        if not isinstance(frame, _Frame):
-            raise Exception("frame must be a _Frame object")
+        if not isinstance(frame, Frame):
+            raise Exception("frame must be a vidformer.cv2.Frame object")
         if frame._modified:
             f_obj = _filter_scale(frame._f, pix_fmt=self._pix_fmt)
             self._frames.append(f_obj)
@@ -134,11 +183,14 @@ class VideoWriter:
             self._frames.append(frame._f)
 
     def release(self):
-        spec = self.vf_spec()
+        if self._path is None:
+            return
+
+        spec = self.spec()
         server = _server()
         spec.save(server, self._path)
 
-    def vf_spec(self):
+    def spec(self) -> vf.Spec:
         fmt = {
             "width": self._size[0],
             "height": self._size[1],
@@ -154,12 +206,79 @@ class VideoWriter_fourcc:
         self._args = args
 
 
+def imread(path, *args):
+    if len(args) > 0:
+        raise NotImplementedError("imread does not support additional arguments")
+
+    assert path.lower().endswith((".jpg", ".jpeg", ".png"))
+    server = _server()
+    source = vf.Source(server, str(uuid.uuid4()), path, 0)
+    frame = Frame(source.iloc[0], source.fmt())
+    return frame
+
+
+def imwrite(path, img, *args):
+    if len(args) > 0:
+        raise NotImplementedError("imwrite does not support additional arguments")
+
+    if not isinstance(img, Frame):
+        raise Exception("img must be a vidformer.cv2.Frame object")
+
+    fmt = img._fmt.copy()
+    width = fmt["width"]
+    height = fmt["height"]
+    f = img._f
+
+    domain = [Fraction(0, 1)]
+
+    if path.lower().endswith(".png"):
+        img._mut()  # Make sure it's in rgb24
+        spec = vf.Spec(
+            domain,
+            lambda t, i: img._f,
+            {"width": width, "height": height, "pix_fmt": "rgb24"},
+        )
+        spec.save(_server(), path, encoder="png")
+    elif path.lower().endswith((".jpg", ".jpeg")):
+        if img._modified:
+            # it's rgb24, we need to convert to something jpeg can handle
+            f = _filter_scale(img._f, pix_fmt="yuv420p")
+            fmt["pix_fmt"] = "yuv420p"
+        else:
+            if fmt["pix_fmt"] not in ["yuvj420p", "yuvj422p", "yuvj444p"]:
+                f = _filter_scale(img._f, pix_fmt="yuvj420p")
+                fmt["pix_fmt"] = "yuvj420p"
+
+        spec = vf.Spec(domain, lambda t, i: f, fmt)
+        spec.save(_server(), path, encoder="mjpeg")
+    else:
+        raise Exception("Unsupported image format")
+
+
+def vidplay(video, *args, **kwargs):
+    """
+    Play a vidformer video specification.
+
+    Args:
+        video: one of [vidformer.Spec, vidformer.Source, vidformer.cv2.VideoWriter]
+    """
+
+    if isinstance(video, vf.Spec):
+        return video.play(_server(), *args, **kwargs)
+    elif isinstance(video, vf.Source):
+        return video.play(_server(), *args, **kwargs)
+    elif isinstance(video, VideoWriter):
+        return video.spec().play(_server(), *args, **kwargs)
+    else:
+        raise Exception("Unsupported video type to vidplay")
+
+
 def rectangle(img, pt1, pt2, color, thickness=None, lineType=None, shift=None):
     """
     cv.rectangle(	img, pt1, pt2, color[, thickness[, lineType[, shift]]]	)
     """
 
-    assert isinstance(img, _Frame)
+    assert isinstance(img, Frame)
     img._mut()
 
     assert len(pt1) == 2
@@ -203,7 +322,7 @@ def putText(
     cv.putText(	img, text, org, fontFace, fontScale, color[, thickness[, lineType[, bottomLeftOrigin]]]	)
     """
 
-    assert isinstance(img, _Frame)
+    assert isinstance(img, Frame)
     img._mut()
 
     assert isinstance(text, str)
@@ -234,3 +353,186 @@ def putText(
         args.append(bottomLeftOrigin)
 
     img._f = _filter_putText(img._f, text, org, fontFace, fontScale, color, *args)
+
+
+def arrowedLine(
+    img, pt1, pt2, color, thickness=None, line_type=None, shift=None, tipLength=None
+):
+    """
+    cv.arrowedLine(	img, pt1, pt2, color[, thickness[, line_type[, shift[, tipLength]]]]	)
+    """
+    assert isinstance(img, Frame)
+    img._mut()
+
+    assert len(pt1) == 2
+    assert len(pt2) == 2
+    assert all(isinstance(x, int) for x in pt1)
+    assert all(isinstance(x, int) for x in pt2)
+
+    assert len(color) == 3 or len(color) == 4
+    color = [float(x) for x in color]
+    if len(color) == 3:
+        color.append(255.0)
+
+    args = []
+    if thickness is not None:
+        assert isinstance(thickness, int)
+        args.append(thickness)
+    if line_type is not None:
+        assert isinstance(line_type, int)
+        assert thickness is not None
+        args.append(line_type)
+    if shift is not None:
+        assert isinstance(shift, int)
+        assert shift is not None
+        args.append(shift)
+    if tipLength is not None:
+        assert isinstance(tipLength, float)
+        assert shift is not None
+        args.append(tipLength)
+
+    img._f = _filter_arrowedLine(img._f, pt1, pt2, color, *args)
+
+
+def line(img, pt1, pt2, color, thickness=None, lineType=None, shift=None):
+    assert isinstance(img, Frame)
+    img._mut()
+
+    assert len(pt1) == 2
+    assert len(pt2) == 2
+    assert all(isinstance(x, int) for x in pt1)
+    assert all(isinstance(x, int) for x in pt2)
+
+    assert len(color) == 3 or len(color) == 4
+    color = [float(x) for x in color]
+    if len(color) == 3:
+        color.append(255.0)
+
+    args = []
+    if thickness is not None:
+        assert isinstance(thickness, int)
+        args.append(thickness)
+    if lineType is not None:
+        assert isinstance(lineType, int)
+        assert thickness is not None
+        args.append(lineType)
+    if shift is not None:
+        assert isinstance(shift, int)
+        assert shift is not None
+        args.append(shift)
+
+    img._f = _filter_line(img._f, pt1, pt2, color, *args)
+
+
+def circle(img, center, radius, color, thickness=None, lineType=None, shift=None):
+    assert isinstance(img, Frame)
+    img._mut()
+
+    assert len(center) == 2
+    assert all(isinstance(x, int) for x in center)
+
+    assert isinstance(radius, int)
+
+    assert len(color) == 3 or len(color) == 4
+    color = [float(x) for x in color]
+    if len(color) == 3:
+        color.append(255.0)
+
+    args = []
+    if thickness is not None:
+        assert isinstance(thickness, int)
+        args.append(thickness)
+    if lineType is not None:
+        assert isinstance(lineType, int)
+        assert thickness is not None
+        args.append(lineType)
+    if shift is not None:
+        assert isinstance(shift, int)
+        assert shift is not None
+        args.append(shift)
+
+    img._f = _filter_circle(img._f, center, radius, color, *args)
+
+
+def getFontScaleFromHeight(*args, **kwargs):
+    """
+    cv.getFontScaleFromHeight(	fontFace, pixelHeight[, thickness]	)
+    """
+    if _opencv2 is None:
+        raise NotImplementedError("getFontScaleFromHeight requires the cv2 module")
+    return _opencv2.getFontScaleFromHeight(*args, **kwargs)
+
+
+def getTextSize(*args, **kwargs):
+    """
+    cv.getTextSize(	text, fontFace, fontScale, thickness	)
+    """
+    if _opencv2 is None:
+        raise NotImplementedError("getTextSize requires the cv2 module")
+    return _opencv2.getTextSize(*args, **kwargs)
+
+
+def addWeighted(src1, alpha, src2, beta, gamma, dst=None, dtype=-1):
+    """
+    cv.addWeighted(	src1, alpha, src2, beta, gamma[, dst[, dtype]]	) -> 	dst
+    """
+    assert isinstance(src1, Frame)
+    assert isinstance(src2, Frame)
+    src1._mut()
+    src2._mut()
+
+    if dst is None:
+        dst = Frame(src1._f, src1._fmt.copy())
+    else:
+        assert isinstance(dst, Frame)
+    dst._mut()
+
+    assert isinstance(alpha, float) or isinstance(alpha, int)
+    assert isinstance(beta, float) or isinstance(beta, int)
+    assert isinstance(gamma, float) or isinstance(gamma, int)
+    alpha = float(alpha)
+    beta = float(beta)
+    gamma = float(gamma)
+
+    if dtype != -1:
+        raise Exception("addWeighted does not support the dtype argument")
+
+    dst._f = _filter_addWeighted(src1._f, alpha, src2._f, beta, gamma)
+    return dst
+
+
+# Stubs for unimplemented functions
+
+
+def clipLine(*args, **kwargs):
+    raise NotImplementedError("clipLine is not yet implemented in the cv2 frontend")
+
+
+def drawContours(*args, **kwargs):
+    raise NotImplementedError("drawContours is not yet implemented in the cv2 frontend")
+
+
+def drawMarker(*args, **kwargs):
+    raise NotImplementedError("drawMarker is not yet implemented in the cv2 frontend")
+
+
+def ellipse(*args, **kwargs):
+    raise NotImplementedError("ellipse is not yet implemented in the cv2 frontend")
+
+
+def ellipse2Poly(*args, **kwargs):
+    raise NotImplementedError("ellipse2Poly is not yet implemented in the cv2 frontend")
+
+
+def fillConvexPoly(*args, **kwargs):
+    raise NotImplementedError(
+        "fillConvexPoly is not yet implemented in the cv2 frontend"
+    )
+
+
+def fillPoly(*args, **kwargs):
+    raise NotImplementedError("fillPoly is not yet implemented in the cv2 frontend")
+
+
+def polylines(*args, **kwargs):
+    raise NotImplementedError("polylines is not yet implemented in the cv2 frontend")

vidformer/vf.py

@@ -1,3 +1,13 @@
+"""
+vidformer-py is a Python 🐍 interface for [vidformer](https://github.com/ixlab/vidformer).
+
+**Quick links:**
+* [📦 PyPI](https://pypi.org/project/vidformer/)
+* [📘 Documentation - vidformer-py](https://ixlab.github.io/vidformer/vidformer-py/)
+* [📘 Documentation - vidformer.cv2](https://ixlab.github.io/vidformer/vidformer-py-cv2/)
+* [🧑‍💻 Source Code](https://github.com/ixlab/vidformer/tree/main/vidformer-py/)
+"""
+
 import subprocess
 from fractions import Fraction
 import random
@@ -43,19 +53,40 @@ def _check_hls_link_exists(url, max_attempts=150, delay=0.1):
 
 
 class Spec:
+    """
+    A video transformation specification.
+
+    See https://ixlab.github.io/vidformer/concepts.html for more information.
+    """
+
     def __init__(self, domain: list[Fraction], render, fmt: dict):
         self._domain = domain
         self._render = render
         self._fmt = fmt
 
     def __repr__(self):
-        lines = []
-        for i, t in enumerate(self._domain):
-            frame_expr = self._render(t, i)
-            lines.append(
-                f"{t.numerator}/{t.denominator} => {frame_expr}",
-            )
-        return "\n".join(lines)
+        if len(self._domain) <= 20:
+            lines = []
+            for i, t in enumerate(self._domain):
+                frame_expr = self._render(t, i)
+                lines.append(
+                    f"{t.numerator}/{t.denominator} => {frame_expr}",
+                )
+            return "\n".join(lines)
+        else:
+            lines = []
+            for i, t in enumerate(self._domain[:10]):
+                frame_expr = self._render(t, i)
+                lines.append(
+                    f"{t.numerator}/{t.denominator} => {frame_expr}",
+                )
+            lines.append("...")
+            for i, t in enumerate(self._domain[-10:]):
+                frame_expr = self._render(t, i)
+                lines.append(
+                    f"{t.numerator}/{t.denominator} => {frame_expr}",
+                )
+            return "\n".join(lines)
 
     def _sources(self):
         s = set()
@@ -351,12 +382,18 @@ class Loader:
 
 
 class YrdenServer:
-    """A connection to a Yrden server"""
+    """
+    A connection to a Yrden server.
+
+    A yrden server is the main API for local use of vidformer.
+    """
 
     def __init__(self, domain=None, port=None, bin=None):
-        """Connect to a Yrden server
+        """
+        Connect to a Yrden server
 
-        Can either connect to an existing server, if domain and port are provided, or start a new server using the provided binary
+        Can either connect to an existing server, if domain and port are provided, or start a new server using the provided binary.
+        If no domain or binary is provided, the `VIDFORMER_BIN` environment variable is used.
         """
 
         self._domain = domain
@@ -503,11 +540,13 @@ class SourceILoc:
 
     def __getitem__(self, idx):
         if type(idx) != int:
-            raise Exception("Source iloc index must be an integer")
+            raise Exception(f"Source iloc index must be an integer, got a {type(idx)}")
         return SourceExpr(self._source, idx, True)
 
 
 class Source:
+    """A video source."""
+
     def __init__(
         self, server: YrdenServer, name: str, path: str, stream: int, service=None
     ):
@@ -610,6 +649,8 @@ def _json_arg(arg, skip_data_anot=False):
 
 
 class Filter:
+    """A video filter."""
+
     def __init__(self, name: str, tl_func=None, **kwargs):
         self._name = name
 
@@ -854,6 +895,10 @@ class UDF:
 
 
 class UDFFrameType:
+    """
+    Frame type for use in UDFs.
+    """
+
     def __init__(self, width: int, height: int, pix_fmt: str):
         assert type(width) == int
         assert type(height) == int
@@ -886,6 +931,8 @@ class UDFFrameType:
 
 
 class UDFFrame:
+    """A symbolic reference to a frame for use in UDFs."""
+
     def __init__(self, data: np.ndarray, f_type: UDFFrameType):
         assert type(data) == np.ndarray
         assert type(f_type) == UDFFrameType

vidformer-0.6.2.dist-info/METADATA

@@ -0,0 +1,36 @@
+Metadata-Version: 2.1
+Name: vidformer
+Version: 0.6.2
+Summary: A Python library for creating and viewing videos with vidformer.
+Author-email: Dominik Winecki <dominikwinecki@gmail.com>
+Requires-Python: >=3.8
+Description-Content-Type: text/markdown
+Classifier: Programming Language :: Python :: 3
+Classifier: Operating System :: OS Independent
+Requires-Dist: requests
+Requires-Dist: msgpack
+Requires-Dist: numpy
+Project-URL: Documentation, https://ixlab.github.io/vidformer/vidformer-py/
+Project-URL: Homepage, https://ixlab.github.io/vidformer/
+Project-URL: Issues, https://github.com/ixlab/vidformer/issues
+
+# vidformer-py
+
+[![PyPI version](https://img.shields.io/pypi/v/vidformer.svg)](https://pypi.org/project/vidformer/)
+[![License](https://img.shields.io/badge/License-Apache%202.0-blue.svg)](https://github.com/ixlab/vidformer/blob/main/LICENSE)
+
+vidformer-py is a Python 🐍 interface for [vidformer](https://github.com/ixlab/vidformer).
+Our [getting started guide](https://ixlab.github.io/vidformer/getting-started.html) explains how to use it.
+
+**Quick links:**
+* [📦 PyPI](https://pypi.org/project/vidformer/)
+* [📘 Documentation - vidformer-py](https://ixlab.github.io/vidformer/vidformer-py/)
+* [📘 Documentation - vidformer.cv2](https://ixlab.github.io/vidformer/vidformer-py-cv2/)
+* [🧑‍💻 Source Code](https://github.com/ixlab/vidformer/tree/main/vidformer-py/)
+
+**Publish:**
+```bash
+ export FLIT_USERNAME='__token__' FLIT_PASSWORD='<token>'
+flit publish
+```
+

vidformer-0.6.2.dist-info/RECORD

@@ -0,0 +1,7 @@
+vidformer/__init__.py,sha256=JnJ5T-Afn8dZXAkwmq38cFzrnK4aZZZSFhHdUqVLdno,113
+vidformer/vf.py,sha256=0idKHOkWXC5bki1thrgihKQwJ91gDCzf96tl3Gm5vaA,31059
+vidformer/cv2/__init__.py,sha256=wOjDsYyUKlP_Hye8-tyz-msu9xwaPMpN2sGMu3Lh3-w,22
+vidformer/cv2/vf_cv2.py,sha256=CO9LZ7FwOhrzNJ1eFjd9JddZm0PjY-XtOlKIYGB3iQQ,15486
+vidformer-0.6.2.dist-info/WHEEL,sha256=EZbGkh7Ie4PoZfRQ8I0ZuP9VklN_TvcZ6DSE5Uar4z4,81
+vidformer-0.6.2.dist-info/METADATA,sha256=_jQnVbcd_OHAy57G7HXIijyXUOoR9Q8AzURvjn3ljDA,1487
+vidformer-0.6.2.dist-info/RECORD,,

vidformer-0.6.0.dist-info/METADATA

@@ -1,24 +0,0 @@
-Metadata-Version: 2.1
-Name: vidformer
-Version: 0.6.0
-Summary: A Python library for creating and viewing videos with vidformer.
-Author-email: Dominik Winecki <dominikwinecki@gmail.com>
-Requires-Python: >=3.8
-Description-Content-Type: text/markdown
-Classifier: Programming Language :: Python :: 3
-Classifier: Operating System :: OS Independent
-Requires-Dist: requests
-Requires-Dist: msgpack
-Requires-Dist: numpy
-Project-URL: Homepage, https://ixlab.github.io/vidformer/
-Project-URL: Issues, https://ixlab.github.io/vidformer/issues
-
-# vidformer-py
-
-## Publish
-
-```bash
- export FLIT_USERNAME='__token__' FLIT_PASSWORD='<token>'
-flit publish
-```
-

vidformer-0.6.0.dist-info/RECORD

@@ -1,7 +0,0 @@
-vidformer/__init__.py,sha256=MgZOCVL7wOxSseM5LArXHtW3FzPKLcmGzUaosUhSn6A,113
-vidformer/vf.py,sha256=gexrp0PQ8cbkixCPLY9BCquHeHWfD6iUcA_wbSxGmFQ,29511
-vidformer/cv2/__init__.py,sha256=wOjDsYyUKlP_Hye8-tyz-msu9xwaPMpN2sGMu3Lh3-w,22
-vidformer/cv2/vf_cv2.py,sha256=C3b8OEGSHh7AFbtddybP_DwR7rvy34lCewt9JWcLTaM,6196
-vidformer-0.6.0.dist-info/WHEEL,sha256=EZbGkh7Ie4PoZfRQ8I0ZuP9VklN_TvcZ6DSE5Uar4z4,81
-vidformer-0.6.0.dist-info/METADATA,sha256=hK3hz1DM6NQfkmMz2L5Uc098wgCvPJzaHbpkZy-DDt8,643
-vidformer-0.6.0.dist-info/RECORD,,