Merge master into worksplit-multigpu (#13546)

* fix: pin SQLAlchemy>=2.0 in requirements.txt (fixes #13036) (#13316) * Refactor io to IO in nodes_ace.py (#13485) * Bump comfyui-frontend-package to 1.42.12 (#13489) * Make the ltx audio vae more native. (#13486) * feat(api-nodes): add automatic downscaling of videos for ByteDance 2 nodes (#13465) * Support standalone LTXV audio VAEs (#13499) * [Partner Nodes] added 4K resolution for Veo models; added Veo 3 Lite model (#13330) * feat(api nodes): added 4K resolution for Veo models; added Veo 3 Lite model Signed-off-by: bigcat88 <bigcat88@icloud.com> * increase poll_interval from 5 to 9 --------- Signed-off-by: bigcat88 <bigcat88@icloud.com> Co-authored-by: Jedrzej Kosinski <kosinkadink1@gmail.com> * Bump comfyui-frontend-package to 1.42.14 (#13493) * Add gpt-image-2 as version option (#13501) * Allow logging in comfy app files. (#13505) * chore: update workflow templates to v0.9.59 (#13507) * fix(veo): reject 4K resolution for veo-3.0 models in Veo3VideoGenerationNode (#13504) The tooltip on the resolution input states that 4K is not available for veo-3.1-lite or veo-3.0 models, but the execute guard only rejected the lite combination. Selecting 4K with veo-3.0-generate-001 or veo-3.0-fast-generate-001 would fall through and hit the upstream API with an invalid request. Broaden the guard to match the documented behavior and update the error message accordingly. Co-authored-by: Jedrzej Kosinski <kosinkadink1@gmail.com> * feat: RIFE and FILM frame interpolation model support (CORE-29) (#13258) * initial RIFE support * Also support FILM * Better RAM usage, reduce FILM VRAM peak * Add model folder placeholder * Fix oom fallback frame loss * Remove torch.compile for now * Rename model input * Shorter input type name --------- * fix: use Parameter assignment for Stable_Zero123 cc_projection weights (fixes #13492) (#13518) On Windows with aimdo enabled, disable_weight_init.Linear uses lazy initialization that sets weight and bias to None to avoid unnecessary memory allocation. This caused a crash when copy_() was called on the None weight attribute in Stable_Zero123.__init__. Replace copy_() with direct torch.nn.Parameter assignment, which works correctly on both Windows (aimdo enabled) and other platforms. * Derive InterruptProcessingException from BaseException (#13523) * bump manager version to 4.2.1 (#13516) * ModelPatcherDynamic: force cast stray weights on comfy layers (#13487) the mixed_precision ops can have input_scale parameters that are used in tensor math but arent a weight or bias so dont get proper VRAM management. Treat these as force-castable parameters like the non comfy weight, random params are buffers already are. * Update logging level for invalid version format (#13526) * [Partner Nodes] add SD2 real human support (#13509) * feat(api-nodes): add SD2 real human support Signed-off-by: bigcat88 <bigcat88@icloud.com> * fix: add validation before uploading Assets Signed-off-by: bigcat88 <bigcat88@icloud.com> * Add asset_id and group_id displaying on the node Signed-off-by: bigcat88 <bigcat88@icloud.com> * extend poll_op to use instead of custom async cycle Signed-off-by: bigcat88 <bigcat88@icloud.com> * added the polling for the "Active" status after asset creation Signed-off-by: bigcat88 <bigcat88@icloud.com> * updated tooltip for group_id * allow usage of real human in the ByteDance2FirstLastFrame node * add reference count limits * corrected price in status when input assets contain video Signed-off-by: bigcat88 <bigcat88@icloud.com> --------- Signed-off-by: bigcat88 <bigcat88@icloud.com> * feat: SAM (segment anything) 3.1 support (CORE-34) (#13408) * [Partner Nodes] GPTImage: fix price badges, add new resolutions (#13519) * fix(api-nodes): fixed price badges, add new resolutions Signed-off-by: bigcat88 <bigcat88@icloud.com> * proper calculate the total run cost when "n > 1" Signed-off-by: bigcat88 <bigcat88@icloud.com> --------- Signed-off-by: bigcat88 <bigcat88@icloud.com> * chore: update workflow templates to v0.9.61 (#13533) * chore: update embedded docs to v0.4.4 (#13535) * add 4K resolution to Kling nodes (#13536) Signed-off-by: bigcat88 <bigcat88@icloud.com> * Fix LTXV Reference Audio node (#13531) * comfy-aimdo 0.2.14: Hotfix async allocator estimations (#13534) This was doing an over-estimate of VRAM used by the async allocator when lots of little small tensors were in play. Also change the versioning scheme to == so we can roll forward aimdo without worrying about stable regressions downstream in comfyUI core. * Disable sageattention for SAM3 (#13529) Causes Nans * execution: Add anti-cycle validation (#13169) Currently if the graph contains a cycle, the just inifitiate recursions, hits a catch all then throws a generic error against the output node that seeded the validation. Instead, fail the offending cycling mode chain and handlng it as an error in its own right. Co-authored-by: guill <jacob.e.segal@gmail.com> * chore: update workflow templates to v0.9.62 (#13539) --------- Signed-off-by: bigcat88 <bigcat88@icloud.com> Co-authored-by: Octopus <liyuan851277048@icloud.com> Co-authored-by: comfyanonymous <121283862+comfyanonymous@users.noreply.github.com> Co-authored-by: Comfy Org PR Bot <snomiao+comfy-pr@gmail.com> Co-authored-by: Alexander Piskun <13381981+bigcat88@users.noreply.github.com> Co-authored-by: Jukka Seppänen <40791699+kijai@users.noreply.github.com> Co-authored-by: AustinMroz <austin@comfy.org> Co-authored-by: Daxiong (Lin) <contact@comfyui-wiki.com> Co-authored-by: Matt Miller <matt@miller-media.com> Co-authored-by: blepping <157360029+blepping@users.noreply.github.com> Co-authored-by: Dr.Lt.Data <128333288+ltdrdata@users.noreply.github.com> Co-authored-by: rattus <46076784+rattus128@users.noreply.github.com> Co-authored-by: guill <jacob.e.segal@gmail.com>
2026-08-25 10:22:12 +08:00 · 2026-04-23 19:20:14 -07:00
parent aa464b36b3
commit 1b96430c60
34 changed files with 5123 additions and 250 deletions
--- a/comfy_api_nodes/util/init.py
+++ b/comfy_api_nodes/util/init.py
@@ -19,6 +19,7 @@ from .conversions import (
    image_tensor_pair_to_batch,
    pil_to_bytesio,
    resize_mask_to_image,
+    resize_video_to_pixel_budget,
    tensor_to_base64_string,
    tensor_to_bytesio,
    tensor_to_pil,
@@ -90,6 +91,7 @@ __all__ = [
    "image_tensor_pair_to_batch",
    "pil_to_bytesio",
    "resize_mask_to_image",
+    "resize_video_to_pixel_budget",
    "tensor_to_base64_string",
    "tensor_to_bytesio",
    "tensor_to_pil",
--- a/comfy_api_nodes/util/client.py
+++ b/comfy_api_nodes/util/client.py
@@ -156,6 +156,7 @@ async def poll_op(
    estimated_duration: int | None = None,
    cancel_endpoint: ApiEndpoint | None = None,
    cancel_timeout: float = 10.0,
+    extra_text: str | None = None,
 ) -> M:
    raw = await poll_op_raw(
        cls,
@@ -176,6 +177,7 @@ async def poll_op(
        estimated_duration=estimated_duration,
        cancel_endpoint=cancel_endpoint,
        cancel_timeout=cancel_timeout,
+        extra_text=extra_text,
    )
    if not isinstance(raw, dict):
        raise Exception("Expected JSON response to validate into a Pydantic model, got non-JSON (binary or text).")
@@ -260,6 +262,7 @@ async def poll_op_raw(
    estimated_duration: int | None = None,
    cancel_endpoint: ApiEndpoint | None = None,
    cancel_timeout: float = 10.0,
+    extra_text: str | None = None,
 ) -> dict[str, Any]:
    """
    Polls an endpoint until the task reaches a terminal state. Displays time while queued/processing,
@@ -299,6 +302,7 @@ async def poll_op_raw(
                    price=state.price,
                    is_queued=state.is_queued,
                    processing_elapsed_seconds=int(proc_elapsed),
+                    extra_text=extra_text,
                )
                await asyncio.sleep(1.0)
        except Exception as exc:
@@ -389,6 +393,7 @@ async def poll_op_raw(
                    price=state.price,
                    is_queued=False,
                    processing_elapsed_seconds=int(state.base_processing_elapsed),
+                    extra_text=extra_text,
                )
                return resp_json

@@ -462,6 +467,7 @@ def _display_time_progress(
    price: float | None = None,
    is_queued: bool | None = None,
    processing_elapsed_seconds: int | None = None,
+    extra_text: str | None = None,
 ) -> None:
    if estimated_total is not None and estimated_total > 0 and is_queued is False:
        pe = processing_elapsed_seconds if processing_elapsed_seconds is not None else elapsed_seconds
@@ -469,7 +475,8 @@ def _display_time_progress(
        time_line = f"Time elapsed: {int(elapsed_seconds)}s (~{remaining}s remaining)"
    else:
        time_line = f"Time elapsed: {int(elapsed_seconds)}s"
-    _display_text(node_cls, time_line, status=status, price=price)
+    text = f"{time_line}\n\n{extra_text}" if extra_text else time_line
+    _display_text(node_cls, text, status=status, price=price)


 async def _diagnose_connectivity() -> dict[str, bool]:
--- a/comfy_api_nodes/util/conversions.py
+++ b/comfy_api_nodes/util/conversions.py
@@ -129,22 +129,38 @@ def pil_to_bytesio(img: Image.Image, mime_type: str = "image/png") -> BytesIO:
    return img_byte_arr


+def _compute_downscale_dims(src_w: int, src_h: int, total_pixels: int) -> tuple[int, int] | None:
+    """Return downscaled (w, h) with even dims fitting ``total_pixels``, or None if already fits.
+
+    Source aspect ratio is preserved; output may drift by a fraction of a percent because both dimensions
+    are rounded down to even values (many  codecs require divisible-by-2).
+    """
+    pixels = src_w * src_h
+    if pixels <= total_pixels:
+        return None
+    scale = math.sqrt(total_pixels / pixels)
+    new_w = max(2, int(src_w * scale))
+    new_h = max(2, int(src_h * scale))
+    new_w -= new_w % 2
+    new_h -= new_h % 2
+    return new_w, new_h
+
+
 def downscale_image_tensor(image: torch.Tensor, total_pixels: int = 1536 * 1024) -> torch.Tensor:
-    """Downscale input image tensor to roughly the specified total pixels."""
+    """Downscale input image tensor to roughly the specified total pixels.
+
+    Output dimensions are rounded down to even values so that the result is guaranteed to fit within ``total_pixels``
+    and is compatible with codecs that require even dimensions (e.g. yuv420p).
+    """
    samples = image.movedim(-1, 1)
-    total = int(total_pixels)
-    scale_by = math.sqrt(total / (samples.shape[3] * samples.shape[2]))
-    if scale_by >= 1:
+    dims = _compute_downscale_dims(samples.shape[3], samples.shape[2], int(total_pixels))
+    if dims is None:
        return image
-    width = round(samples.shape[3] * scale_by)
-    height = round(samples.shape[2] * scale_by)
-
-    s = common_upscale(samples, width, height, "lanczos", "disabled")
-    s = s.movedim(1, -1)
-    return s
+    new_w, new_h = dims
+    return common_upscale(samples, new_w, new_h, "lanczos", "disabled").movedim(1, -1)


-def downscale_image_tensor_by_max_side(image: torch.Tensor, *,  max_side: int) -> torch.Tensor:
+def downscale_image_tensor_by_max_side(image: torch.Tensor, *, max_side: int) -> torch.Tensor:
    """Downscale input image tensor so the largest dimension is at most max_side pixels."""
    samples = image.movedim(-1, 1)
    height, width = samples.shape[2], samples.shape[3]
@@ -399,6 +415,72 @@ def trim_video(video: Input.Video, duration_sec: float) -> Input.Video:
        raise RuntimeError(f"Failed to trim video: {str(e)}") from e


+def resize_video_to_pixel_budget(video: Input.Video, total_pixels: int) -> Input.Video:
+    """Downscale a video to fit within ``total_pixels`` (w * h), preserving aspect ratio.
+
+    Returns the original video object untouched when it already fits. Preserves frame rate, duration, and audio.
+    Aspect ratio is preserved up to a fraction of a percent (even-dim rounding).
+    """
+    src_w, src_h = video.get_dimensions()
+    scale_dims = _compute_downscale_dims(src_w, src_h, total_pixels)
+    if scale_dims is None:
+        return video
+    return _apply_video_scale(video, scale_dims)
+
+
+def _apply_video_scale(video: Input.Video, scale_dims: tuple[int, int]) -> Input.Video:
+    """Re-encode ``video`` scaled to ``scale_dims`` with a single decode/encode pass."""
+    out_w, out_h = scale_dims
+    output_buffer = BytesIO()
+    input_container = None
+    output_container = None
+
+    try:
+        input_source = video.get_stream_source()
+        input_container = av.open(input_source, mode="r")
+        output_container = av.open(output_buffer, mode="w", format="mp4")
+
+        video_stream = output_container.add_stream("h264", rate=video.get_frame_rate())
+        video_stream.width = out_w
+        video_stream.height = out_h
+        video_stream.pix_fmt = "yuv420p"
+
+        audio_stream = None
+        for stream in input_container.streams:
+            if isinstance(stream, av.AudioStream):
+                audio_stream = output_container.add_stream("aac", rate=stream.sample_rate)
+                audio_stream.sample_rate = stream.sample_rate
+                audio_stream.layout = stream.layout
+                break
+
+        for frame in input_container.decode(video=0):
+            frame = frame.reformat(width=out_w, height=out_h, format="yuv420p")
+            for packet in video_stream.encode(frame):
+                output_container.mux(packet)
+        for packet in video_stream.encode():
+            output_container.mux(packet)
+
+        if audio_stream is not None:
+            input_container.seek(0)
+            for audio_frame in input_container.decode(audio=0):
+                for packet in audio_stream.encode(audio_frame):
+                    output_container.mux(packet)
+            for packet in audio_stream.encode():
+                output_container.mux(packet)
+
+        output_container.close()
+        input_container.close()
+        output_buffer.seek(0)
+        return InputImpl.VideoFromFile(output_buffer)
+
+    except Exception as e:
+        if input_container is not None:
+            input_container.close()
+        if output_container is not None:
+            output_container.close()
+        raise RuntimeError(f"Failed to resize video: {str(e)}") from e
+
+
 def _f32_pcm(wav: torch.Tensor) -> torch.Tensor:
    """Convert audio to float 32 bits PCM format. Copy-paste from nodes_audio.py file."""
    if wav.dtype.is_floating_point: