agentscope-ai/agentscope · error · ValueError

Unsupported video source type: {type(source)}

Error message

Unsupported video source type: {type(source)}

What it means

Video counterpart of the image guard: while formatting a video DataBlock for DashScope, the source is neither a supported inline/local type (converted to a data: URL) nor a URL source, so formatting aborts with ValueError.

Source

Thrown at src/agentscope/formatter/_dashscope_formatter.py:164

        """Convert a video source to DashScope's ``video_url`` format
        (OpenAI-compatible extension).

        Local ``file://`` URLs are read from disk and converted to base64
        data URIs. Remote URLs are passed through unchanged.
        """
        if isinstance(source, Base64Source):
            url = f"data:{source.media_type};base64,{source.data}"
        elif isinstance(source, URLSource):
            url_str = str(source.url)
            if url_str.startswith("file://"):
                local_path = url_str.removeprefix("file://")
                with open(local_path, "rb") as f:
                    encoded = base64.b64encode(f.read()).decode("utf-8")
                url = f"data:{source.media_type};base64,{encoded}"
            else:
                url = url_str
        else:
            raise ValueError(f"Unsupported video source type: {type(source)}")

        return {
            "type": "video_url",
            "video_url": {"url": url},
        }

    @staticmethod
    def _format_audio_source(
        source: URLSource | Base64Source,
    ) -> dict[str, Any]:
        """Convert an audio source to DashScope ``input_audio`` format.

        DashScope's compatible API accepts URLs directly in the ``data``
        field. Base64-encoded audio must be wrapped in a data URL. Local
        files are read, base64-encoded, and wrapped in the same form.
        """
        fmt = source.media_type.split("/")[-1]
        if fmt == "mpeg":

View on GitHub (pinned to e90f1c7592)

Solutions

1. Re-encode path as Base64Source/local file source or reference the hosted file with URLSource
2. Check that video loading actually succeeded (source is not None) before building the message
3. Keep a single agentscope version in the environment

Example fix

# before
block = DataBlock(source=None, media_type="video/mp4")

# after
from agentscope.message import DataBlock, URLSource
block = DataBlock(source=URLSource(url="https://cdn/x.mp4", media_type="video/mp4"))

from agentscope.message import Base64Source, URLSource
for block in msg_video_blocks:
    if not isinstance(getattr(block, "source", None), (Base64Source, URLSource)):
        raise TypeError("video block source unsupported")

from agentscope.message import DataBlock, Base64Source, URLSource

def video_block_ok(block: DataBlock) -> bool:
    return isinstance(getattr(block, "source", None), (Base64Source, URLSource))

try:
    resp = await model(msg)
except ValueError as e:
    if "Unsupported video source type" in str(e):
        msg = normalize_video_blocks(msg)
        resp = await model(msg)
    else:
        raise

When it happens

Trigger: Video DataBlock with source=None, a plain dict, or a foreign/custom source class passed to a DashScope model; _format_video_source reaches its final else.

Common situations: Large video files where source loading failed silently and left None; mixed-version imports; converting from OpenAI video formats that use dicts.

Related errors

• Unsupported image source type: {type(source)}
• Multimodal embedding API only supports URL input for video d
• Unsupported source type: {type(source)}
• Unsupported audio source type: {type(source)}
• Text embedding model {self.model!r} only accepts str inputs,

AI-assisted analysis of agentscope-ai/agentscope@e90f1c7592 (2026-08-28). Data as JSON: /api/errors/27bb4a7e28c5f6f2. Report an issue: GitHub.