Darcy AI Engine

AI Engine Pipeline

Darcy AI Engine is a Python library that makes building AI apps as easy as building any other type of app. AI Engine exposes high-level constructs (InputStream, Perceptor, Callback, OutputStream) that you assemble in a Pipeline with a few lines of Python.

To get started, see the Build Guide, look at the examples, and consult the Python reference docs.

Example

This code (from basic.py) shows how easy it is to create a minimal AI Engine pipeline:

#Add the DarcyAI components that we need, particularly the InputStream, OutputStream, Pipeline, and PerceptorMock
from darcyai.tests.perceptor_mock import PerceptorMock
from darcyai.pipeline import Pipeline
from sample_input_stream import SampleInputStream
from sample_output_stream import SampleOutputStream

#Define a class to hold all of our operations
class SingleStreamDemo():
    def __init__(self):
        #Create an input stream and an output stream that we can use in our demo
        ping = SampleInputStream()
        output_stream = SampleOutputStream()

        #Give our class a pipeline property and instantiate it with a Darcy AI pipeline
        self.__pipeline = Pipeline(input_stream=ping,
                                   input_stream_error_handler_callback=self.__input_stream_error_handler_callback,
                                   universal_rest_api=True,
                                   rest_api_base_path="/pipeline",
                                   rest_api_host="0.0.0.0",
                                   rest_api_port=8080)

        #Add our output stream to the pipeline
        self.__pipeline.add_output_stream("output", self.__output_stream_callback, output_stream)

        #Create a mock perceptor and add it to the pipeline
        p1 = PerceptorMock()
        self.__pipeline.add_perceptor("p1", p1, accelerator_idx=0, input_callback=self.__perceptor_input_callback)

    #Define a "run" method that just calls "run" on the pipeline
    def run(self):
        self.__pipeline.run()

    #Define an input callback for the mock perceptor that just sends the data onward with no manipulation
    def __perceptor_input_callback(self, input_data, pom, config):
        return input_data

    #Define an output stream callback that does not manipulate the data
    def __output_stream_callback(self, pom, input_data):
        pass

    #Define an input stream error handler callback that just continues onward
    def __input_stream_error_handler_callback(self, exception):
        pass

#In the main thread, start the application by instantiating our demo class and calling "run"
if __name__ == "__main__":
    single_stream = SingleStreamDemo()
    single_stream.run()

Issues, Contributing, Discussion

If you discover issues with AI Engine, view the issues, or create a new one. You can also submit a Pull Request, or join the discussions.

Metadata

Release files for darcyai 2.2.0.dev20230113

For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.

Built distribution (wheel)

Table of built distributions (wheels) for darcyai 2.2.0.dev20230113
File	Interpreter	ABI	Platform	Reset
darcyai-2.2.0.dev20230113-py3-none-any.whl 13.9 MB Jan 13, 2023	Python 3	none	any	Details

Release files / darcyai-2.2.0.dev20230113-py3-none-any.whl

Download URL	darcyai-2.2.0.dev20230113-py3-none-any.whl
Size	13.9 MB
Tags	Python 3
SHA-256 checksum How to use checksums	`2ac588235d30dc40c7458e714927d3c030133ea7c3d04e9942ff30ff045a6e23`
BLAKE2b-256 checksum How to use checksums	`ab4c49df8fe61d69ef57cade9829e6638906a2f4ead22537b222b85c6ee9c75d`
Upload date	Jan 13, 2023
Uploaded using Trusted Publishing? What is trusted publishing?	No
Uploaded via	`twine/4.0.2 CPython/3.8.13`