Allow user-provided ratchet seed and inumber

[appcypher]

Summary

In some cases, we want to be able to deterministically create a directory with seeds generated at userland, for example, from private keys. There is a hacky solution to this by passing a carefully designed rng, but that is likely to break in future versions.

This PR adds the API for accepting user-specified ratchet seed and inumber that solves the problem outlined above.

This PR implements the following features

• Support user-specified ratchet seed and inumber
• Expose corresponding wasm/js API
• Bump version

Test plan (required)

• Testing

  rs-wnfs test --all

Closing issues

Fixes #86

[codecov]

Codecov Report

Merging #91 (4f346ec) into main (9d3b5c2) will increase coverage by 1.35%.
The diff coverage is 81.03%.

@@            Coverage Diff             @@
##             main      #91      +/-   ##
==========================================
+ Coverage   67.89%   69.25%   +1.35%     
==========================================
  Files          26       27       +1     
  Lines        2003     2010       +7     
  Branches      482      470      -12     
==========================================
+ Hits         1360     1392      +32     
+ Misses        252      239      -13     
+ Partials      391      379      -12     

Impacted Files	Coverage Δ
wnfs/examples/private.rs	0.00% <0.00%> (ø)
wnfs/src/common/utils.rs	70.00% <ø> (ø)
wnfs/src/private/forest.rs	59.15% <ø> (ø)
wnfs/src/private/hamt/node.rs	64.45% <ø> (+0.38%)	⬆️
wnfs/src/private/privateref.rs	68.96% <68.96%> (ø)
wnfs/src/private/node.rs	77.90% <93.33%> (+8.99%)	⬆️
wnfs/src/private/directory.rs	77.34% <100.00%> (+1.92%)	⬆️
wnfs/src/private/file.rs	80.67% <100.00%> (+1.50%)	⬆️
wnfs/src/private/previous.rs	72.98% <100.00%> (+0.57%)	⬆️
wnfs/src/public/directory.rs	71.17% <0.00%> (-0.21%)	⬇️
... and 4 more

[matheus23]

Nice! Should we also add a matching function PrivateRef::from_seed(parent_bare_name: Namefilter, ratchet_seed: HashOutput, inumber: HashOutput) -> PrivateRef?

[matheus23]

What's the reasoning on using with_* naming?
I actually find it slightly confusing and think from_* makes more sense, because the parameters to the from_seed and from_ratchet_key functions deterministically define the PrivateRef. Every property in the PrivateRef depends on the arguments in these functions.
with seems to indicate there's 'something else' to it.

That said, I may just not know the normal rust conventions yet! 😅

[appcypher]

I don't mind either really but generally with_* seems to be preferred if you constructing an object out of other thing(s), and from if you are going to implement the From<T> trait. In this case that means implementing a From<(Namefilter, HashOutput, ...)> which doesn't feel nice. So I went with with_*. But I agree with you from_* sounds better.

Maybe this is sth @zeeshanlakhani can help out with.

[matheus23]

Ah and @appcypher can you add a unit test that creates a PrivateDirectory::with_seed, stores it in a PrivateForest, then loads it again via a PrivateRef::with_seed with the same parameters? 🙏

[matheus23]

Awesome!