binary-husky/gpt_academic · error · ValueError

Invalid image URL format.

Error message

Invalid image URL format.

What it means

gen_image_banana() supports only inline data URIs: it requires ';base64,' in image_url. If the provider returns an ordinary https URL or a malformed URL, this ValueError is raised. Like error 7, it is caught immediately and normally surfaces wrapped as error 9.

Source

Thrown at crazy_functions/Image_Generate.py:273

            yield from update_ui_latest_msg(lastmsg=f"Generate Failed\n\n{generated_content}\n\nStatus Code: {response.status_code}", chatbot=chatbot, history=history, delay=0)
            return

        if image_url is None:
            raise RuntimeError("No image URL found in the response.")

        logger.info(f'Generated image.')
        yield from update_ui_latest_msg(lastmsg=f"Downloading image", chatbot=chatbot, history=history, delay=0)

        if ';base64,' in image_url:
            base64_string = image_url.split('base64,')[-1]
            image_data = base64.b64decode(base64_string)
            file_path = f'{get_log_folder()}/image_gen/'
            os.makedirs(file_path, exist_ok=True)
            file_name = 'Image' + time.strftime("%Y-%m-%d-%H-%M-%S", time.localtime()) + '.png'
            fp = file_path+file_name
            with open(fp, 'wb+') as f: f.write(image_data)
        else:
            raise ValueError("Invalid image URL format.")

        return image_url, fp

    except Exception as e:
        yield from update_ui_latest_msg(lastmsg=f"Generate failed, please try again later.", chatbot=chatbot, history=history, delay=0)
        raise RuntimeError(f"Failed to generate image, please try again later: {str(e)}")









@CatchException
def 图片生成_NanoBanana(prompt, llm_kwargs, plugin_kwargs, chatbot, history, system_prompt, user_request):
    history = []    # 清空历史,以免输入溢出

View on GitHub (pinned to d6bde0fa54)

Solutions

1. Log image_url to see whether it is an https URL or malformed data.
2. Add an else branch that downloads the remote URL, as gen_image() already does.
3. Configure the provider/relay to return base64 data URIs if local download is undesirable.
4. Validate the URL scheme before decoding base64.
5. Preserve the original ValueError in the outer error message.

Example fix

# before
if ';base64,' in image_url:
    ...
else:
    raise ValueError("Invalid image URL format.")

# after
if image_url.startswith(("http://", "https://")):
    r = requests.get(image_url, proxies=proxies, timeout=120)
    r.raise_for_status()
    image_data = r.content
elif ";base64," in image_url:
    image_data = base64.b64decode(image_url.split("base64,", 1)[1])
else:
    raise ValueError(f"Invalid image URL format: {image_url[:100]}")

from urllib.parse import urlparse
parsed = urlparse(image_url or "")
is_data_uri = (image_url or "").startswith("data:") and ";base64," in image_url
is_http_url = parsed.scheme in {"http", "https"} and bool(parsed.netloc)
assert is_data_uri or is_http_url

def classify_image_url(url) -> str:
    if not isinstance(url, str) or not url:
        return "invalid"
    if url.startswith("data:") and ";base64," in url:
        return "base64"
    if urlparse(url).scheme in {"http", "https"} and urlparse(url).netloc:
        return "remote"
    return "invalid"

try:
    image_data = decode_or_download_image(image_url)
except ValueError as e:
    report_unsupported_image_url(image_url)
    raise

When it happens

Trigger: The image endpoint returns https://.../image.png instead of data:image/png;base64,..., or the nested image_url.url field has an unexpected format.

Common situations: Using an OpenAI-compatible relay that returns temporary download URLs; a NanoBanana proxy configured not to inline base64; partial URL truncation; provider schema changes.

Related errors

• response.content.decode()
• No image URL found in the response.
• Failed to generate image, please try again later: {str(e)}
• Invalid URL: {url}

AI-assisted analysis of binary-husky/gpt_academic@d6bde0fa54 (2026-08-14). Data as JSON: /api/errors/a6bc9f12361740d9. Report an issue: GitHub.