imandra 0.1.12

Python interface to Imandra (cloud-native automated reasoning engine for analysis of algorithms)

imandra is a Python client for Imandra - a cloud-native automated reasoning engine built on the latest advances in formal verification and symbolic reasoning.

Usage

The client idiomatically exposes the Imandra Iterative Decomposition Framework (IDF) functionality to decompose state-transition algorithms. The client accepts a string containing the Python code to decompose - see the documentation to learn how to write accepted model code:

code = """
class State:
     def __init__(self):
          self.x : int = 0
     def receive_Update(self, v : int):
          self.x : int = v

scenario = [ 'Update' , 'Update' ]
"""

You can launch an IDF decomposition job in the cloud with imandra.idf.decompose, passing it the code string:

from imandra import idf
decomposition = idf.decompose(code)

Using the decomposition object as a handle, you can check the status of the job which can be in queued, processing, done or error status:

print(decomposition.status())

When the decomposition job is finished, you can obtain the generated Python code for the regions:

# Create a separate Python function per region predicate 
print(decomposition.dumps(kind='flat'))
# Create a single optimized function that returns a region number
print(decomposition.dumps(kind='tree'))

Links

• Imandra Python Documentation
• Imandra Project Homepage
• Describing Algorithms: Introduction