Create dependencies graph

[mtillman14]

With different analyses & specific functions indicated, etc. Each variable will be one row within its function, with vertical lines connecting an input to output variable.

[mtillman14]

This graph actually should be a centerpiece of interaction with the functions! In line with the "analysis threads" concept, there really is no clean way to slice/version these functions/variables. So, the only way to manage them all somewhat decently is graphically! The graph will be a linked set of nodes, with each node being one function. Arrows will run from one function to the next in the order they're run. If a function has different inputs but the same name as an existing one, it gets a parallel line to the original. New functions that branch off from existing ones just get a branch, that may or may not merge back in with the other(s).

[mtillman14]

Because it will get crowded, the graph should allow slicing via a bunch of different methods:

1. Project-specific (because by default it shows all of the projects)
2. Highlight the path from import to a specific variable
3. Highlight all of the functions that use a specific variable as an input.
4. Highlight all of the places where a function's outputs are used as inputs.
5. Specific processing group(s) functions and/or arguments flow. Others??

[mtillman14]

Graph pane should also have a search bar to search by functions and/or argument names (and whether they're inputs or outputs).

[mtillman14]

Because nodes are functions, the function names should label the nodes.

[mtillman14]

Where does the documentation of which Specify Trials were used come in?

[mtillman14]

Where does the documentation of which Specify Trials were used come in?

Still an open question :(

[mtillman14]

The processing map shows the general order of functions, but not which variables are used for which functions.

[mtillman14]

Am now using a digraph with a node for each type of object, even variables.