Mnemo

Muses

The following Muses are under development to work with Mnemosyne.
For information on developing new Muses see Architecture#Muses.
For examples of multiple muses working together see multi-muse flows.

Argument Predictor

Probabilistic code completion of callsite arguments using a statistical database.

DIG

Execution trace driven Dynamic Invariant Generation (DIG).

Function Generator

Creation of a function body based on the prototype and documentation string using OpenAI's Codex model.

Galois Autocomplete

Fork of the Galois project for machine learning based code completion.

GenPatcher

Automated program repair technique.

Compiler Explorer

Integration of the Compiler Explorer interactive compiler framework.

Herbie

Integration of the Herbie system for improving the accuracy of floating-point expressions.

Hypothesis Tests

Type-driven automated tests using the hypothesis library

Snippet Mining

Implementation of the algorithm described in "Exempla Gratis (E.G.) - Code Examples for Free" to provide developers with example API usage snippets mined from a corpus of programs

SSR

AST-level software search and replacement. Enables automated structured library replacement using pre-defined rules.

Trinity

Integration of the Trinity An Extensible Synthesis Framework for Data Science technique for program synthesis.

TypeWriter

Integration of the TypeWriter technique for the inference of types in Python.