跳到内容

Openai

OpenAIImageGenerationToolSpec #

基类：BaseToolSpec

OpenAI 图像生成工具 Spec。

源代码位置：llama-index-integrations/tools/llama-index-tools-openai/llama_index/tools/openai/image_generation/base.py 
 29
 30
 31
 32
 33
 34
 35
 36
 37
 38
 39
 40
 41
 42
 43
 44
 45
 46
 47
 48
 49
 50
 51
 52
 53
 54
 55
 56
 57
 58
 59
 60
 61
 62
 63
 64
 65
 66
 67
 68
 69
 70
 71
 72
 73
 74
 75
 76
 77
 78
 79
 80
 81
 82
 83
 84
 85
 86
 87
 88
 89
 90
 91
 92
 93
 94
 95
 96
 97
 98
 99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
class OpenAIImageGenerationToolSpec(BaseToolSpec):
    """OpenAI Image Generation tool spec."""

    spec_functions = ["image_generation"]

    def __init__(
        self, api_key: Optional[str] = None, cache_dir: Optional[str] = None
    ) -> None:
        try:
            from openai import OpenAI
        except ImportError:
            raise ImportError(
                "Please install openai with `pip install openai` to use this tool"
            )

        """Initialize with parameters."""
        self.client = OpenAI(api_key=api_key)
        self.cache_dir = cache_dir or DEFAULT_CACHE_DIR

    def get_cache_dir(self):
        return self.cache_dir

    def save_base64_image(self, base64_str, image_name):
        try:
            from io import BytesIO

            from PIL import Image
        except ImportError:
            raise ImportError(
                "Please install Pillow with `pip install Pillow` to use this tool"
            )
        cache_dir = self.cache_dir

        # Create cache directory if it doesn't exist
        if not os.path.exists(cache_dir):
            os.makedirs(cache_dir)

        # Decode the base64 string
        image_data = base64.b64decode(base64_str)

        # Create an image from the decoded bytes and save it
        image_path = os.path.join(cache_dir, image_name)
        with Image.open(BytesIO(image_data)) as img:
            img.save(image_path)

        return image_path

    def image_generation(
        self,
        text: str,
        model: Optional[str] = "dall-e-3",
        quality: Optional[str] = "standard",
        num_images: Optional[int] = 1,
        size: Optional[str] = DEFAULT_SIZE,
        style: Optional[str] = "vivid",
        timeout: Optional[int] = None,
        download: Optional[bool] = False,
    ) -> str:
        """
        This tool accepts a natural language string and will use OpenAI's DALL-E model to generate an image.

        Args:
            text: The text to generate an image from.

            model: The model to use for image generation. Defaults to `dall-e-3`.
                Must be one of `dall-e-2` or `dall-e-3`.

            num_images: The number of images to generate. Defaults to 1.
                Must be between 1 and 10. For `dall-e-3`, only `n=1` is supported.

            quality: The quality of the image that will be generated. Defaults to `standard`.
                Must be one of `standard` or `hd`. `hd` creates images with finer
                details and greater consistency across the image. This param is only supported
                for `dall-e-3`.

            size: The size of the generated images. Defaults to `1024x1024`.
                Must be one of `256x256`, `512x512`, or `1024x1024` for `dall-e-2`.
                Must be one of `1024x1024`, `1792x1024`, or `1024x1792` for `dall-e-3` models.

            style: The style of the generated images. Defaults to `vivid`.
                Must be one of `vivid` or `natural`.
                Vivid causes the model to lean towards generating hyper-real and dramatic images.
                Natural causes the model to produce more natural, less hyper-real looking images.
                This param is only supported for `dall-e-3`.

            timeout: Override the client-level default timeout for this request, in seconds. Defaults to `None`.

            download: If `True`, the image will be downloaded to the cache directory. Defaults to `True`.

        """
        if size not in valid_sizes[model]:
            raise Exception(f"Invalid size for {model}: {size}")

        response = self.client.images.generate(
            prompt=text,
            n=num_images,
            model=model,
            quality=quality,
            size=size,
            response_format="b64_json" if download else "url",
            style=style,
            timeout=timeout,
        )
        if download:
            image_bytes = response.data[0].b64_json
            ext = get_extension(image_bytes)
            filename = f"{time.time()}.{ext}"

            return (self.save_base64_image(image_bytes, filename),)

        return response.data[0].url

image_generation #

image_generation(text: str, model: Optional[str] = 'dall-e-3', quality: Optional[str] = 'standard', num_images: Optional[int] = 1, size: Optional[str] = DEFAULT_SIZE, style: Optional[str] = 'vivid', timeout: Optional[int] = None, download: Optional[bool] = False) -> str

此工具接受自然语言字符串，并将使用 OpenAI 的 DALL-E 模型生成图像。

参数

名称 类型 描述 默认值
text Optional[str]

用于生成图像的文本。

 必需
model Optional[str]

用于图像生成的模型。默认为 dall-e-3。必须是 dall-e-2 或 dall-e-3 之一。

 'dall-e-3'
num_images Optional[int]

要生成的图像数量。默认为 1。必须介于 1 和 10 之间。对于 dall-e-3，仅支持 n=1。

 1
quality Optional[str]

将生成的图像质量。默认为 standard。必须是 standard 或 hd 之一。hd 会生成具有更精细细节和更高图像一致性的图像。此参数仅支持 dall-e-3。

 'standard'
size Optional[str]

生成图像的尺寸。默认为 1024x1024。对于 dall-e-2 模型，必须是 256x256、512x512 或 1024x1024 之一。对于 dall-e-3 模型，必须是 1024x1024、1792x1024 或 1024x1792 之一。

 DEFAULT_SIZE
style Optional[str]

生成图像的风格。默认为 vivid。必须是 vivid 或 natural 之一。vivid 风格使模型倾向于生成超现实和戏剧性的图像。natural 风格使模型生成更自然、看起来不那么超现实的图像。此参数仅支持 dall-e-3 模型。

 'vivid'
timeout Optional[int]

覆盖此请求的客户端级别默认超时时间，以秒为单位。默认为 None。

 无
download Optional[bool]

如果为 True，图像将被下载到缓存目录。默认为 True。

 False
源代码位置：llama-index-integrations/tools/llama-index-tools-openai/llama_index/tools/openai/image_generation/base.py 
 76
 77
 78
 79
 80
 81
 82
 83
 84
 85
 86
 87
 88
 89
 90
 91
 92
 93
 94
 95
 96
 97
 98
 99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
def image_generation(
    self,
    text: str,
    model: Optional[str] = "dall-e-3",
    quality: Optional[str] = "standard",
    num_images: Optional[int] = 1,
    size: Optional[str] = DEFAULT_SIZE,
    style: Optional[str] = "vivid",
    timeout: Optional[int] = None,
    download: Optional[bool] = False,
) -> str:
    """
    This tool accepts a natural language string and will use OpenAI's DALL-E model to generate an image.

    Args:
        text: The text to generate an image from.

        model: The model to use for image generation. Defaults to `dall-e-3`.
            Must be one of `dall-e-2` or `dall-e-3`.

        num_images: The number of images to generate. Defaults to 1.
            Must be between 1 and 10. For `dall-e-3`, only `n=1` is supported.

        quality: The quality of the image that will be generated. Defaults to `standard`.
            Must be one of `standard` or `hd`. `hd` creates images with finer
            details and greater consistency across the image. This param is only supported
            for `dall-e-3`.

        size: The size of the generated images. Defaults to `1024x1024`.
            Must be one of `256x256`, `512x512`, or `1024x1024` for `dall-e-2`.
            Must be one of `1024x1024`, `1792x1024`, or `1024x1792` for `dall-e-3` models.

        style: The style of the generated images. Defaults to `vivid`.
            Must be one of `vivid` or `natural`.
            Vivid causes the model to lean towards generating hyper-real and dramatic images.
            Natural causes the model to produce more natural, less hyper-real looking images.
            This param is only supported for `dall-e-3`.

        timeout: Override the client-level default timeout for this request, in seconds. Defaults to `None`.

        download: If `True`, the image will be downloaded to the cache directory. Defaults to `True`.

    """
    if size not in valid_sizes[model]:
        raise Exception(f"Invalid size for {model}: {size}")

    response = self.client.images.generate(
        prompt=text,
        n=num_images,
        model=model,
        quality=quality,
        size=size,
        response_format="b64_json" if download else "url",
        style=style,
        timeout=timeout,
    )
    if download:
        image_bytes = response.data[0].b64_json
        ext = get_extension(image_bytes)
        filename = f"{time.time()}.{ext}"

        return (self.save_base64_image(image_bytes, filename),)

    return response.data[0].url