Main note for hypothesis protobuf testing project
Brainstorming Notes
- Containing type isn’t relevant for my purposes
- If it’s enum, check its enum type
- If it’s message, check message type
- Might have to do a special hardcode for the timestamp, I’m not sure, probably something to do with hypothesis’s datetime checking libraries
- If it’s stuff like “key”, “value”, then it’s just a normal python dictionary (map), about which we don’t really have any more information (will have to try to refer to the actual data in order to do that conversion)
- Some sort of fuzzy text model, where it can do some string comparisons? And figure out what is common to all of these strings. 202005150918
- maybe I go through the entire sample document, and for the leaves, I understand the distributions for all the leaf nodes? Not sure long that will take
This note should’ve been an outline 202006080125
Josh’s strategy for other components
Well so there are two components I think you could dig into, that will compose into a better larger tool:
- Hypothesis strategy for TU variables
- Given a pandas dataframe, create a hypothesis generator from it based on the values in the columns
References
Current Outline:
Current improvement areas
uid: 202005141716 tags: #theorem