Make proto output deterministic

alan-turing-institute / uatk-spc

Synthetic Population Catalyst

MIT License

20 stars 12 forks source link

Closed sgreenbury closed 1 year ago

sgreenbury commented 1 year ago

Closes #52.

This PR ensures protobuf outputs are deterministic for a given seed by:

Using BTreeMap over HashMap as venues_per_sic is looped over non-determininistically with HashMap (see comment).
Configuring prost to serialize with BTreeMap over HashMap (see comment).
Adding integration test for equality and inequality of the SHA256 file hashes upon population generation.

A script is also added for comparing any given region with diff on JSON printed when outputs differ.

sgreenbury commented 1 year ago

Awesome PR! Thank you for fixing the problem and adding 2 thorough tests. Few tiny comments, then feel free to merge

Thanks @dabreegster, no problem! I've made some small amendments given your comments and will now merge.