[][src]Crate orkhon

Orkhon

Orkhon: ML Inference Framework and Server Runtime

What is it?

Orkhon is Rust framework for Machine Learning to run/use inference/prediction code written in Python, frozen models and process unseen data. It is mainly focused on serving models and processing unseen data in a performant manner. Instead of using Python directly and having scalability problems for servers this framework tries to solve them with built-in async API.

Main features

  • Sync & Async API for models.
  • Easily embeddable engine for well-known Rust web frameworks.
  • API contract for interacting with Python code.
  • High processing throughput

Installation

You can include Orkhon into your project with;

[dependencies]
orkhon = "*"

Dependencies

You will need:

  • Rust Nightly needed (for now. until async support fully lands in)
  • Python dev dependencies installed and have proper python runtime to use Orkhon with your project.
  • Point out your PYTHONHOME environment variable to your Python installation.

Python API contract

Python API contract is hook based. If you want to call a method for prediction you should write Python code with args and **kwargs.

def model_hook(args, **kwargs):
    print("Doing prediction...")
    return args

Python Hook Input

Both args and kwargs are HashSets. args can take any acceptable hashset key and passes down to python level. But kwargs keys are restricted to [&str] for keeping it only for option passing. args can contain your data for making prediction. Input contract is opinionated for making interpreter work without unknown type conversions.

Python Hook Output

Python hook output is passed up without downcasting or casting. Python bindings are still exposed to make sure you get the type you wanted. By default; python passes PyObject to Rust interface. You can extract the type from the object that Python passed with

pyobj.extract()?

This api uses PyO3 bindings for Python <-> Rust. You can look for PyO3's documentation to make conversions. Auto conversion methods soon will be added.

Examples

Creating Orkhon

Orkhon::new()
   .config(OrkhonConfig::new())
   .pymodel("model_which_will_be_tested", // Unique identifier of the model
            "tests/pymodels",             // Python module directory
            "model_test",                 // Python module file name
       "model_hook"                       // Hook(Python method) that will be called by Orkhon
   )
   .build();

Requesting to Orkhon

// Args for the request hook
let mut request_args = HashMap::new();
request_args.insert("is", 10);
request_args.insert("are", 6);
request_args.insert("you", 5);

// Kwargs for the request hook
let mut request_kwargs = HashMap::<&str, &str>::new();

// Future handle (await over it... if you want)
let handle =
    o.pymodel_request_async(
        "model_which_will_be_tested",
            ORequest::with_body(
                PyModelRequest::new()
                    .with_args(request_args)
                    .with_kwargs(request_kwargs)
            )
    );

License

License is MIT

Discussion and Development

We use Gitter for development discussions. Also please don't hesitate to open issues on GitHub ask for features, report bugs, comment on design and more! More interaction and more ideas are better!

Contributing to Orkhon Open Source Helpers

All contributions, bug reports, bug fixes, documentation improvements, enhancements and ideas are welcome.

A detailed overview on how to contribute can be found in the CONTRIBUTING guide on GitHub.

Modules

config
errors
orkhon
pooled
reqrep
service
tensorflow

Macros

request_async_for
request_sync_for