Updates for CGGMP'24: Paillier/ring-Pedersen level

[fjarri]

Part of #157

• Split Paillier keys and ring-Pedersen secrets, with the underlying SecretPrimes/PublicModulus. Note that according to the new version of the paper Paillier only needs mod 3 = 4 primes, while ring-Pedersen needs safe primes.
• Use Exponentiable where approriate
• Normalize naming for "wire" and "expanded" structures in paillier. Namely: the wire ones have the "Wire" postfix, the expanded ones have no postfixes. The conversion methods are to_wire() and to_precomputed().
• Changed the type of various Uint bit size arguments to u32 to match crypto-bigint

The new Secret will be properly used throughout the codebase in a follow-up PR, this one is getting pretty big already.

[codecov]

Codecov Report

Attention: Patch coverage is 98.58233% with 13 lines in your changes missing coverage. Please review.

Project coverage is 92.36%. Comparing base (8ec70fd) to head (f6cd680).
Report is 11 commits behind head on master.

Files with missing lines	Patch %	Lines
synedrion/src/cggmp21/interactive_signing.rs	90.00%	4 Missing ⚠️
synedrion/src/tools/secret.rs	85.71%	3 Missing ⚠️
synedrion/src/paillier/encryption.rs	97.43%	2 Missing ⚠️
synedrion/src/paillier/rsa.rs	98.89%	2 Missing ⚠️
synedrion/src/cggmp21/key_refresh.rs	94.11%	1 Missing ⚠️
synedrion/src/paillier/keys.rs	99.47%	1 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##           master     #161      +/-   ##
==========================================
+ Coverage   92.21%   92.36%   +0.14%     
==========================================
  Files          33       35       +2     
  Lines        6422     6534     +112     
==========================================
+ Hits         5922     6035     +113     
+ Misses        500      499       -1     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[dvdplm]

LGTM.

Left a few comments and nitpicks. At the end of the day I think using "wire" is probably fine, but I don't love it and think that the previous names were mostly ok and quite readable.

[fjarri]

At the end of the day I think using "wire" is probably fine, but I don't love it and think that the previous names were mostly ok and quite readable.

The problem with the previous names was that the most used ones were also very long and had suffixes that had to be repeated everywhere. The goal was to make it so that one could use PublicKeyPaillier/SecretKeyPaillier in most places, and where they have to be serialized, they're converted to a "minimized" representation which has a longer name. Also sometimes it was Precomputed, and sometimes Mod, and the conversion method names were different while they were actually doing the same thing. Whether it's Wire or something else is secondary and can be changed.

[fjarri]

Merging this for now to free the way for #163, we can work on naming in a separate PR.