Instance¶

Instance ¶

Instance(
    project_path: Optional[StrOrPathLike] = None,
    dist_path: Optional[StrOrPathLike] = None,
    port: Optional[int] = None,
    host: str = "127.0.0.1",
    *,
    already_running: bool = False,
    disable_code: bool = False,
    tutorial_only: bool = False,
    context_in_body: bool = False,
    wait_for_init_model: bool = False,
    ping: bool = True,
    gpu: int = 0
)

A class for starting or connecting to a PEKAT VISION project.

Parameters:

Name	Type	Description	Default
`project_path`	`Optional[StrOrPathLike]`	Path to an existing project. Ignored if `already_running` is `True`.	`None`
`dist_path`	`Optional[StrOrPathLike]`	Path to PEKAT VISION installation folder. Ignored if `already_running` is `True`.	`None`
`port`	`Optional[int]`	Port of the project.	`None`
`host`	`str`	IP address of the project. Set to `"0.0.0.0"` to listen on all interfaces.	`'127.0.0.1'`
`already_running`	`bool`	Whether the project is already running. If the project is not already running, project will be started.	`False`
`disable_code`	`bool`	Disable code module. Ignored if `already_running` is `True`.	`False`
`tutorial_only`	`bool`	Only allow Anomaly module and force tutorial. Ignored if `already_running` is `True`.	`False`
`context_in_body`	`bool`	Whether to send the `context` `dict` in the response body. Set to `True` if the resulting context is larger than 4KB.	`False`
`wait_for_init_model`	`bool`	Whether to wait for all models to be initialized before returning the Instance object. Ignored if `already_running` is `True`.	`False`
`ping`	`bool`	Whether to send a ping request to `HOST:PORT/ping` upon creating the instance.	`True`
`gpu`	`int`	Which GPU to start project on. Ignored if `already_running` is `True`.	`0`

Raises:

Type	Description
`DistNotExistsError`	If `dist_path` does not exist.
`DistNotFoundError`	If no PEKAT VISION was found in the default installation directory.
`PortIsAllocatedError`	If the specified port is already used by another process.
`ValueError`	If not `already_running` and project path is `None`.
`RuntimeError`	If the created process has no `stdout`.
`ProjectNotFoundError`	If the project was not found in the specified location.
`PekatNotStartedError`	If PEKAT VISION couldn't start a project.

analyze ¶

analyze(
    image: Union[NDArray[np.uint8], bytes, StrOrPathLike],
    response_type: ResponseType = "context",
    data: Optional[str] = None,
    timeout: float = 20,
) -> Result

Send an image to the running project and get the results.

response_type will affect the image of the returned Result:

"context": context only, Result.image_bytes will be None.
"image": Result.image_bytes will contain the processed image.
"annotated_image": Result.image_bytes will contain the processed image with drawn detected rectangles.
"heatmap": Result.image_bytes will contain the layered heatmaps of the processed image.

In order to decode the image in Result.image_bytes, use get_decoded_image.

Parameters:

Name	Type	Description	Default
`image`	`Union[NDArray[uint8], bytes, StrOrPathLike]`	Path to the image, encoded image bytes or numpy image.	required
`response_type`	`ResponseType`	Type of response data.	`'context'`
`data`	`Optional[str]`	Data to be added to the query. Project will be able to access this under the `"data"` key in `context`.	`None`
`timeout`	`float`	Timeout in seconds for the analyze request.	`20`

Raises:

Type	Description
`InvalidResponseTypeError`	If `response_type` is not any of the allowed response types.
`InvalidDataTypeError`	If `image` is not any of the allowed data types.
`Timeout`	Upon timeout.

Returns:

Type	Description
`Result`	A tuple of image and `context`. If `response_type` is `"context"`, then the returned image is `None`.

send_random ¶

send_random(
    shape: Tuple[int, ...] = (512, 512, 3),
    response_type: ResponseType = "context",
    data: Optional[str] = None,
    timeout: float = 20,
) -> Result

Send random data for analysis.

Parameters:

Name	Type	Description	Default
`shape`	`Tuple[int, ...]`	Shape of the image to be sent.	`(512, 512, 3)`
`response_type`	`ResponseType`	Type of response data.	`'context'`
`data`	`Optional[str]`	Data to be added to the query. Project will be able to access this under the `"data"` key in `context`.	`None`
`timeout`	`float`	Timeout in seconds for the analyze request.	`20`

Raises:

Type	Description
`Exception`	Same as `Instance.analyze`

Returns:

Type	Description
`Result`	A tuple of image and `context`. If `response_type` is `"context"`, then the returned image is `None`.

stop ¶

stop(timeout: float = 5) -> None

Stop the project if it's running and not stopping already.

It can only stop the project if already_running was set to True.

Parameters:

Name	Type	Description	Default
`timeout`	`float`	Timeout in seconds to kill process.	`5`

Raises:

Type	Description
`NoConnectionError`	Upon timeout.

ping ¶

ping(timeout: float = 5) -> requests.Response

Ping the project to check if it's running.

Parameters:

Name	Type	Description	Default
`timeout`	`float`	Timeout in seconds to ping.	`5`

Raises:

Type	Description
`NoConnectionError`	Upon timeout.

Returns:

Type	Description
`Response`	Ping response.