Skip to content

Image query task

ImageQueryTask

Bases: BaseTask

A task that executes a natural language query on one or more input images. Accepts a text prompt and a list of images as input in one of the following formats: - tuple of (template string, list[ImageArtifact]) - tuple of (TextArtifact, list[ImageArtifact]) - Callable that returns a tuple of (TextArtifact, list[ImageArtifact])

Attributes:

Name Type Description
image_query_engine ImageQueryEngine

The engine used to execute the query.
Source code in griptape/tasks/image_query_task.py 
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
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
@define
class ImageQueryTask(BaseTask):
    """A task that executes a natural language query on one or more input images. Accepts a text prompt and a list of
    images as input in one of the following formats:
    - tuple of (template string, list[ImageArtifact])
    - tuple of (TextArtifact, list[ImageArtifact])
    - Callable that returns a tuple of (TextArtifact, list[ImageArtifact])

    Attributes:
        image_query_engine: The engine used to execute the query.
    """

    _image_query_engine: ImageQueryEngine = field(default=None, kw_only=True, alias="image_query_engine")
    _input: tuple[str, list[ImageArtifact]] | tuple[TextArtifact, list[ImageArtifact]] | Callable[
        [BaseTask], tuple[TextArtifact, list[ImageArtifact]]
    ] = field(default=None, alias="input")

    @property
    def input(self) -> tuple[TextArtifact, list[ImageArtifact]]:
        if isinstance(self._input, tuple):
            if isinstance(self._input[0], TextArtifact):
                query_text = self._input[0]
            else:
                query_text = TextArtifact(J2().render_from_string(self._input[0], **self.full_context))

            return query_text, self._input[1]
        elif isinstance(self._input, Callable):
            return self._input(self)
        else:
            raise ValueError(
                "Input must be a tuple of a TextArtifact and a list of ImageArtifacts or a callable that "
                "returns a tuple of a TextArtifact and a list of ImageArtifacts."
            )

    @input.setter
    def input(
        self,
        value: tuple[TextArtifact, list[ImageArtifact]]
        | Callable[[BaseTask], tuple[TextArtifact, list[ImageArtifact]]],
    ) -> None:
        self._input = value

    @property
    def image_query_engine(self) -> ImageQueryEngine:
        if self._image_query_engine is None:
            if self.structure is not None:
                self._image_query_engine = ImageQueryEngine(
                    image_query_driver=self.structure.config.global_drivers.image_query_driver
                )
            else:
                raise ValueError("Image Query Engine is not set.")
        return self._image_query_engine

    @image_query_engine.setter
    def image_query_engine(self, value: ImageQueryEngine) -> None:
        self._image_query_engine = value

    def run(self) -> TextArtifact:
        query, image_artifacts = self.input

        response = self.image_query_engine.run(query.value, image_artifacts)

        return response

image_query_engine: ImageQueryEngine property writable

input: tuple[TextArtifact, list[ImageArtifact]] property writable

run()

Source code in griptape/tasks/image_query_task.py 
70
71
72
73
74
75
def run(self) -> TextArtifact:
    query, image_artifacts = self.input

    response = self.image_query_engine.run(query.value, image_artifacts)

    return response