Replace batch DAG system with streaming architecture

- Remove legacy_tasks.py, hybrid_state.py, render.py
- Remove old task modules (analyze, execute, execute_sexp, orchestrate)
- Add streaming interpreter from test repo
- Add sexp_effects with primitives and video effects
- Add streaming Celery task with CID-based asset resolution
- Support both CID and friendly name references for assets
- Add .dockerignore to prevent local clones from conflicting

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>

sexp_effects/primitive_libs/image.py

@@ -0,0 +1,144 @@
+"""
+Image Primitives Library
+
+Basic image operations: dimensions, pixels, resize, crop, paste.
+"""
+import numpy as np
+import cv2
+
+
+def prim_width(img):
+    return img.shape[1]
+
+
+def prim_height(img):
+    return img.shape[0]
+
+
+def prim_make_image(w, h, color=None):
+    """Create a new image filled with color (default black)."""
+    if color is None:
+        color = [0, 0, 0]
+    img = np.zeros((h, w, 3), dtype=np.uint8)
+    img[:] = color
+    return img
+
+
+def prim_copy(img):
+    return img.copy()
+
+
+def prim_pixel(img, x, y):
+    """Get pixel color at (x, y) as [r, g, b]."""
+    h, w = img.shape[:2]
+    if 0 <= x < w and 0 <= y < h:
+        return list(img[int(y), int(x)])
+    return [0, 0, 0]
+
+
+def prim_set_pixel(img, x, y, color):
+    """Set pixel at (x, y) to color, returns modified image."""
+    result = img.copy()
+    h, w = result.shape[:2]
+    if 0 <= x < w and 0 <= y < h:
+        result[int(y), int(x)] = color
+    return result
+
+
+def prim_sample(img, x, y):
+    """Bilinear sample at float coordinates, returns [r, g, b] as floats."""
+    h, w = img.shape[:2]
+    x = max(0, min(w - 1.001, x))
+    y = max(0, min(h - 1.001, y))
+
+    x0, y0 = int(x), int(y)
+    x1, y1 = min(x0 + 1, w - 1), min(y0 + 1, h - 1)
+    fx, fy = x - x0, y - y0
+
+    c00 = img[y0, x0].astype(float)
+    c10 = img[y0, x1].astype(float)
+    c01 = img[y1, x0].astype(float)
+    c11 = img[y1, x1].astype(float)
+
+    top = c00 * (1 - fx) + c10 * fx
+    bottom = c01 * (1 - fx) + c11 * fx
+    return list(top * (1 - fy) + bottom * fy)
+
+
+def prim_channel(img, c):
+    """Extract single channel (0=R, 1=G, 2=B)."""
+    return img[:, :, c]
+
+
+def prim_merge_channels(r, g, b):
+    """Merge three single-channel arrays into RGB image."""
+    return np.stack([r, g, b], axis=2).astype(np.uint8)
+
+
+def prim_resize(img, w, h, mode="linear"):
+    """Resize image to w x h."""
+    interp = cv2.INTER_LINEAR
+    if mode == "nearest":
+        interp = cv2.INTER_NEAREST
+    elif mode == "cubic":
+        interp = cv2.INTER_CUBIC
+    elif mode == "area":
+        interp = cv2.INTER_AREA
+    return cv2.resize(img, (int(w), int(h)), interpolation=interp)
+
+
+def prim_crop(img, x, y, w, h):
+    """Crop rectangle from image."""
+    x, y, w, h = int(x), int(y), int(w), int(h)
+    ih, iw = img.shape[:2]
+    x = max(0, min(x, iw - 1))
+    y = max(0, min(y, ih - 1))
+    w = min(w, iw - x)
+    h = min(h, ih - y)
+    return img[y:y+h, x:x+w].copy()
+
+
+def prim_paste(dst, src, x, y):
+    """Paste src onto dst at position (x, y)."""
+    result = dst.copy()
+    x, y = int(x), int(y)
+    sh, sw = src.shape[:2]
+    dh, dw = dst.shape[:2]
+
+    # Clip to bounds
+    sx1 = max(0, -x)
+    sy1 = max(0, -y)
+    dx1 = max(0, x)
+    dy1 = max(0, y)
+    sx2 = min(sw, dw - x)
+    sy2 = min(sh, dh - y)
+
+    if sx2 > sx1 and sy2 > sy1:
+        result[dy1:dy1+(sy2-sy1), dx1:dx1+(sx2-sx1)] = src[sy1:sy2, sx1:sx2]
+
+    return result
+
+
+PRIMITIVES = {
+    # Dimensions
+    'width': prim_width,
+    'height': prim_height,
+
+    # Creation
+    'make-image': prim_make_image,
+    'copy': prim_copy,
+
+    # Pixel access
+    'pixel': prim_pixel,
+    'set-pixel': prim_set_pixel,
+    'sample': prim_sample,
+
+    # Channels
+    'channel': prim_channel,
+    'merge-channels': prim_merge_channels,
+
+    # Geometry
+    'resize': prim_resize,
+    'crop': prim_crop,
+    'paste': prim_paste,
+}