This PR adds the ability to convert an Echidna corpus to the Medusa corpus format. It partially addresses issue #24

It includes the following changes:

• New CLI commands convert and generate. The former is used for corpus conversion, and the latter for generating unit tests from failed properties. 8 new flags were added to the convert command
• A new class EchidnaConverter was added that handles parsing and generation of the converted files. The class takes a lot of code from the Echidna class, so this should be refactored to reduce code duplication before the PR is merged.
• The way Echidna string and bytes are parsed was updated since the previous version was incorrect. More testing should be done on this part since I'm not 100% confident it's correctly decoded.
• web3.py and eth_tester were added as dependencies to be able to ABI encode the parsed arguments to the Medusa data field
• A --no-save flag was added to unit test generation to be able to more easily debug this feature without overriding the already generated files. When --no-save is passed in, the generated unit tests will not be saved to file.

Before merging this PR further work should be done to:

• Refactor the Echidna and EchidnaConverter classes to reduce code duplication
• Add tests to confirm the converted corpus can be used by Medusa
• Figure out how to classify the Echidna coverage call sequences into Medusa immutable and mutable sequences. Currently all sequences are added to both.
• Investigate if Medusa call sequences naming convention needs to be used. Currently we just use the same file names as Echidna does.
• Add Medusa to Echidna corpus conversion