prior to publishing this repository (and submitting the associated paper(s)), some reorganization is needed.

Data

The basic.ipynb notebook contains code for downloading the dataset, as .egg files (readable using quail). The files are too large to store directly in this repository. Some questions:

Should we keep the "raw" data files? E.g., are the participant database (.db) files still useful? Do we need the recall transcript (.wav.txt) files in /data/recall/room//? What's the difference between the "transcripts" stored in the etc folder vs. the data/recall folder? And what do the different subfolders in etc refer to (e.g., what is "old" and/or exp1, exp2, exp3, etc.?)
The current version of the basic.ipynb notebook (as of 11-18-22) also loads in the adaptive-FR data. I'm not sure yet if that experiment should be written up in the same paper, or whether it'd be cleaner to separate the studies into two papers. If everything is published together, it may not make sense to continue maintaining the adaptive-FR repository.

Code

The current codebase is split across several folders:

notebooks: this contains a bunch of now-outdated notebooks for carrying out analyses for conference posters/talks, compiling weekly reports, etc. most of these no longer work and don't load in the latest versions of the data (or only load in parts of the data). however, i'm also developing the main set of analyses in a notebook stored in this folder (basic.ipynb).
notebooks/etc seems to contain some materials related to compiling automated reports, but none of it seems functional
scripts: this contains (now outdated) code and materials for compiling weekly reports. however, it can likely be removed since it does not appear to be functional.

Stimuli

the stimulus folder contains a single .csv file that stores the word pool (and attributes/groupings for each word)
this is probably duplicated in the experiment code repository. a cleaner approach (it seems) would be to clean up the experiment repository and maintain it separately from the "paper/code/data" repository. it's possible that this already exists in the autoFR repository, but i'm not sure where the code is for organizing "early" lists according to a given feature, or where the "feature reduced" code lives-- maybe in this repo? the adaptiveFR experiment exists in its own repository, which (i think) makes sense...

Conda environment, etc.

After finishing up the analyses, I should create a conda environment w/ instructions for setting it up. Alternatively, I could use davos to ensure dependencies are met... 🤔...

Paper

text + figures + source files for the paper should go in a papers folder (and figures should go in papers/figs; source files should go in papers/figs/source).

ContextLab / FRFR-analyses

cleanup plan: revised #5

Data

Code

Stimuli

Conda environment, etc.

Paper