René's URL Explorer Experiment

{"@context":"https://schema.org","@type":"DiscussionForumPosting","headline":"ImageSynthesis的extra_input参数如何设置本地图片路径","articleBody":"根据百炼平台的图像局部重绘的python demo [https://help.aliyun.com/zh/model-studio/user-guide/vary-region?spm=0.0.0.i1#3dc15c8345qbt](url) ，这里设置的base_image_url和mask_image_url服务器无法访问，因此我想从本地加载图片\n### 参考代码：\n`from http import HTTPStatus\nfrom urllib.parse import urlparse, unquote\nfrom pathlib import PurePosixPath\nimport requests\nfrom dashscope import ImageSynthesis\n\nprompt = \"a dog wearing red glasses\"\nmodel = \"wanx-x-painting\"\ntask = \"image2image\"\n\nimage_path = \"/root/llms/Bailian/source3.jpg\"\nmask_path = \"/root/llms/Bailian/glasses.png\"\n\nextra_input = {\n    \"base_image_url\": image_path,\n    \"mask_image_url\": mask_path\n}\n\n\nprint('----sync call, please wait a moment----')\nrsp = ImageSynthesis.call(model=model,\n                          prompt=prompt,\n                          n=1,\n                          size='1024*1024',\n                          task=task,\n                          extra_input=extra_input)\nif rsp.status_code == HTTPStatus.OK:\n    print(rsp.output)\n    # save file to current directory\n    for result in rsp.output.results:\n        file_name = PurePosixPath(unquote(urlparse(result.url).path)).parts[-1]\n        with open('./%s' % file_name, 'wb+') as f:\n            f.write(requests.get(result.url).content)\nelse:\n    print('sync_call Failed, status_code: %s, code: %s, message: %s' %\n          (rsp.status_code, rsp.code, rsp.message))`\n### 错误信息\n----sync call, please wait a moment----\nTraceback (most recent call last):\n  File \"/root/llms/Bailian/test_wanx.py\", line 25, in \u003cmodule\u003e\n    rsp = ImageSynthesis.call(model=model,\n  File \"/root/miniconda3/envs/wanx/lib/python3.10/site-packages/dashscope/aigc/image_synthesis.py\", line 67, in call\n    return super().call(model,\n  File \"/root/miniconda3/envs/wanx/lib/python3.10/site-packages/dashscope/client/base_api.py\", line 219, in call\n    response = cls.wait(task_response,\n  File \"/root/miniconda3/envs/wanx/lib/python3.10/site-packages/dashscope/aigc/image_synthesis.py\", line 198, in wait\n    response = super().wait(task, api_key, workspace=workspace)\n  File \"/root/miniconda3/envs/wanx/lib/python3.10/site-packages/dashscope/client/base_api.py\", line 379, in wait\n    task_id = cls._get_task_id(task)\n  File \"/root/miniconda3/envs/wanx/lib/python3.10/site-packages/dashscope/client/base_api.py\", line 232, in _get_task_id\n    raise InvalidTask('Invalid task, task create failed: %s' %\ndashscope.common.error.InvalidTask: Invalid task, task create failed: {\"status_code\": 400, \"request_id\": \"8af1fd62-8664-925e-9e37-5d71185c986c\", \"code\": \"InvalidParameter.DataInspection\", \"message\": \"The media format is not supported or incorrect for the data inspection.\", \"output\": null, \"usage\": null}\n我看ImageSynthesis的images参数还不支持，是否只支持extra_input参数来传递图片，请介绍一下具体用法，谢谢！","author":{"url":"https://github.com/LiYufengzz","@type":"Person","name":"LiYufengzz"},"datePublished":"2025-04-23T03:06:27.000Z","interactionStatistic":{"@type":"InteractionCounter","interactionType":"https://schema.org/CommentAction","userInteractionCount":2},"url":"https://github.com/5/dashscope-sdk-python/issues/5"}