all: Fix Typing and JSON Conversion

[marioevz]

Changes Included

• Homogenizes the typing of all fields in common/types.py, with the goal of making JSON conversion more streamlined:
  • Every field in each type now has JSON metadata that instructs the JSON encoder how to convert the type before inserting it into the output, with the following possible types:
    • Number: int, str or bytes that converts into a string when inserted into the JSON output
    • HexNumber: int, str or bytes that converts into a hex string when inserted into the JSON output
    • ZeroPaddedHexNumber: int, str or bytes that converts into a zero padded hex string when inserted into the JSON output, where zeroes are added to make the string of an even length.
    • Bytes: str or bytes that converts into an hex representation of the bytes described by the input (e.g. in tx data, extraData in the block header)
    • FixedSizeBytes: base class that is used to represent bytes of a fixed length (e.g. addresses, hashes, or header nonces)
    • Address: Fixed size bytes of length 20
    • Hash: Fixed size bytes of length 32
    • Bloom: Fixed size bytes of length 256
    • HeaderNonce: Fixed size bytes of length 8
  • common/json.py contains all JSON encoder logic.
  • Appropriate tests have been added to verify that the output of the new JSON encoding logic matches the previous output
• Code base class and derivatives do no longer contain assemble method, and now implement SupportsBytes and Sized types.
• Most methods in common/conversions.py have been removed, in favor of the standardized JSON encoder routines.
• Adds Alloc type, which is a Mapping subclass that automatically converts all keys to appropriate address strings, while accepting strings, integers or bytes as key inputs, which would allow us to remove to_address usage across most tests.

[danceratopz]

Added a few commits after comparing fixture output.

This really adds consistency to our json formatting! Thanks for this!