fix(particle-data-store): use signature in particle data store path name [NET-632]

[kmd-fl]

Description

Include signature into the names of the directories for particles in the Particle Data Store

The path format is particle_{particle-id}-peer_{peer-id}_sig-{hex signature}.

The resulting name example (signature is encoded in base58):

particle_16d3c69c-31f1-4506-a16d-f9924d6cf8f4-peer_12D3KooWN4XNKgu76nwB7iKUXmE4FKCA5Ycak6SbSqLTaWo2nFsQ-sig_3XDge4jyajeovwx2av4sarPc3vfEPMpAS1AgibHN1DwujnhM8J4myyWXkHJYCr5xe7E96V2eqtWE8oYjttNbed3y

Motivation

Particle directories must be:

• unique for each particle (the user can assign any particle id, so it's not unique enough)
• protected from malicious particles (a wrong-doer can try to forge a particle to access the data store)

The usage of signatures in the path will provide both by definition.

Why we still use peer id and particle id:

• a peer id is needed since we have several workers on the same host, and need to create a separate data store for each of them;
• a particle ID isn't particularly necessary, but it can help debug, and the path size is still in the range of 255 symbols, which is the minimal path length restriction among modern FS I found.

Additional Notes

I thought maybe we should add a separate structure for (String, PeerId, Vec<u8>) but decided that it doesn't make much sense; however, it would be good to provide a set of aliases for the particle id and the signature to get the (ParticleId, PeerId, Signature) tuple. I'm not sure if I should do it in the same PR, though.

[linear]

NET-632 Use particle.signature instead of particle.id in data store

[justprosh]

Resolve conflicts with upstream and use base58 instead of hex for sig serialization. Everything else is good