-
Notifications
You must be signed in to change notification settings - Fork 2
Future Development Todos
Tyler Yep edited this page Apr 4, 2021
·
7 revisions
-
Big Directional Goals:
- Host on AWS
- UI and secure move API
-
Unit Testing:
- Wolf_inds isn't used in stats.py
- Insomniac testing is very confusing to me.
- Test constructors more thoroughly.
- Make_prediction needs more thorough testing.
-
Design Considerations:
- Rethink Wolf characters and how the class mechanics work before writing their tests. This may be the next big step in the project, or I might go back to multi-statement.
- new_role should be a learnable vector or a dict that gets updated as the probability of change increases.
- Creating a directed statement index dependency graph to use in conjunction with the solver. Find the hubs and find statements that don't connect very much with the graph.
-
Small Stories:
- better eval function for minion / tanner / wolf
-
Wait for a second output change in order to redo the one_night output.
- (hopefully it also produces better results)
-
Add lru_cache tests - figure out if I can solve the problem of changing the const values.
- Call the function multiple times using different values and verify they give the same results. Verify that I’m not returning any dicts or lists (mutable types)
-
Combine basic_guesses_cache with a real lru_cache?
-
Figure out if I can use a fancier cache