The nodes of a hypernode graph can be expanded into graphs.
Implemented in Graphviz using pygraphviz; produces hyperlinked SVG files.
why: hypernode graphs show up in a variety of knowledge management scenarios, so a visualization tool is useful.
make docker_live.make svg
The point of using docker is to capture the dependencies; Docker is not required. You could also run the Python3 script generate_graphviz.py bare-metal if you have pygraphviz available.
The underlying data for the nested graphs comes from browse_internet.py
I realized that the nested nature of a sequence of steps could be described using XML with nested <task> tags.
--> I found that with a high nested depth the navigation was burdensome, even with the ability for XML nodes to be collapsed in Firefox
--> I found that some chunks of tasks are used repeatedly. However, XML does not support reuse of encapsulated snippets (without a more complicated schema).
--> I found that XML's linear sequence did not reflect the branching decision tree flow
I moved to Graphviz (using dreampuf.github.io) and manually created nodes and edges, then converting the top-level nodes into subgraphs.
--> I found that every node eventually becomes a subgraph, which triggers relabeling of nodes and connections
--> Every node requires a unique name
--> reusing sequences meant copy-pasting snippets and manually renaming every variable
I started fresh with Graphviz and experimented with "everything is a subgraph".
--> problem: invisible nodes (necessary for edges) take up space
In a new Graphviz I manually split the hierarchy to limit the depth to 2. Each layer only contains subgraphs and nodes; no nested subgraphs.
--> Hyperlinking the nodes to other layers works well.
--> Automation is necessary to scale
Using Python with import pygraphviz is an obvious transition, but I don't have experience with which data structure
would be appropriate for a hypergraph.
--> Nested dictionaries would work for a tree, but do not for a graph
--> Although I can enter redundant keys in a dictionary, only the last key is actually used.
>>> my_dict = {4: 94, 5:99, 4: 4942}
>>> my_dict.keys()
dict_keys([4, 5])
>>> my_dict
{4: 4942, 5: 99}That led to the current data structure, a list of dictionaries. The key is the subgraph and the list of values is the sequential steps used to implement the key.
https://github.com/LoW12/FractalGraph
https://explained-from-first-principles.com/internet
"Statecharts: A Visual Formalism For Complex Systems" David Harel