Craig Sanders
Craig Sanders
Summary: Refactoring models to accomodate continuous outcomes and pairwise inputs Differential Revision: D36935747
Summary: The Python client is now available in pypi, so it can now be included in the requirements. This will ensure tutorials can be run without any extra installation steps....
Summary: Adds some checks to make sure that x values lie within their bounds and y values are in [0,1] Reviewed By: mshvartsman Differential Revision: D36489189
Installing AEPsych on a new computer is hassle because you have to properly configure the Python environment. [Docker](https://www.docker.com/) can be used to make deployment easier, but the image needs to...
The database is currently poorly documented, and there are no tutorials that explain its design or usage. This documentation will be especially important for explaining the new features (#32, #33,...
The current database is unreliable (tables are not always correctly populated). To better identify and fix bugs, we need more, better unit tests and ensure that the db passes them....
The database should be able to store data from multiple experiments and participants so that users do not have to keep track of folders of multiple db files. Data from...
Currently experiment data is stored in the database as ask/tell messages in the replay table, and as blobs in the strat table, but extracting this data is cumbersome. A new...
Currently experiment configs are stored as blobs in the database, but these are not easily accessible. A new configuration table should be built that stores configs in a human-readable and...
The current database has fields for experiment metadata (name, description, and id), but that data is not actually recorded. Some sort of metadata intake flow needs to be built so...