For the Physics Derivation Graph, I started out entering expressions manually into a plaintext file. I then graduated to manually entering expressions into a XML, then CSV file. Rendering was done in GraphViz and then d3js. In the past few months, I created an interactive terminal prompt for a more automated entry using Python, still rendering with d3js. I experimented with a web-based interface using Flask, but recognized I was slowly evolving towards the EquationMap interface for both display and entry.
Instead of re-inventing EquationMap, I realized I could use EquationMap as a front end. I would then export the graph .tex file and parse it into the Physics Derivation Graph. This benefits both EquationMap and the Physics Derivation Graph: EquationMap gets more traffic and more use cases, and PDG doesn't need to create a front-end.
Here's two different examples of using EquationMap with Inference Rules:
http://equationmap.com/#sSCdqOvUii
http://equationmap.com/#6NXnawarhb
I like the second instance better.
Here's an example of completing the square without inference rules
http://equationmap.com/#OZbmPo15Ha
and with inference rules
http://equationmap.com/#7xXumkNaTH
and generalized symbolically
http://equationmap.com/#xq8KZ0bnkT
I like the fact that I can easily refer to different iterations of the development
The ease of entering Latex and connecting nodes is very convenient!
The export function produces a combination of JSON and Latex.
I've written a parser for the .tex output from EquationMap
https://github.com/allofphysicsgraph/proofofconcept/tree/gh-pages/sandbox/equationMap_tex_parser
No comments:
Post a Comment