foldit_tree_visualization

Requirements:

• Python3 and the following packages
• graphviz
• matplotlib (only for getting the Tableau color palette in the output_tree function in make_tree_viz.py)
• editdistance (only for scaling edge width by sequence difference in the output_tree_helper function in make_tree_viz.py)
• numpy
• scipy

Usage: python3 make_tree_viz.py [--viz] [--evolver] [--uid UID] RANKS_FILE PIDS

--viz the script will generate tree visualizations. Output directory specified in the dot.render call in the output_tree function in make_tree_viz.py
--evolver if --uid is given, the trees will include nodes from evolvers
--uid UID generate trees for user with uid of UID, otherwise trees will be generated for the top 15 soloists, and soloists ranked 36 through 50
RANKS_FILE the path to a csv dump of the rprp_puzzle_ranks table in the Foldit database PIDS one or more puzzle ids to generate trees for

In addition to the ranks file, you will need csv files for the solution data and the history data for each puzzle in whatever subdirectory the datapath function in make_tree_viz.py specifies. Running the script will generate pickled versions of these files in order to save time on future runs.

File	Description
RawNode.py	Code for generating the raw solution trees as well as a number of useful operations related to that
Node.py	Class representing a solution tree, takes a RawNode as input. This class includes the tree summarization functions.
tree_analysis.py	Collection of functions for detecting and quantifiying various solver behaviors
make_tree_viz.py	Main script, loads the csv data and genereates visualizations. Code for labeling nodes with macro ids or as containing a manual-only action has been commented out.