Refactor with richer data

Hendekagon / clique

Makes a graph of dependencies between functions

https://github.com/Hendekagon/clique/wiki

45 stars 6 forks source link

Refactor with richer data #7

Open metasoarous opened 9 years ago

metasoarous commented 9 years ago

This will help support #6.

[ ] Should be passing maps of data about functions/macros around instead of just qualified names. This gives us more flexibility.
[ ] Also in general, the processing passes here are not very clean. Would be better to clearly separate a full dependencies parse from all of the downstream filtering, writing, etc.

Hendekagon commented 9 years ago

preserving such information would allow for graphs to be coloured & styled to show more information, good idea.

btw I made a graph of your project's internal dependencies only, see the wiki - it's a planar graph! (scsv/spit-csv "sdeps.csv" (mapcat (fn [[k v]] (map (partial vector k "") (distinct (c/except (filter namespace (remove (partial = k) v)) (c/default-exclude))))) (c/all-deps "src")))

metasoarous commented 9 years ago

Awesome; that's great to see :-)

I've got a refactor going which simplifies this a little bit:

(-> (dependencies 'clojure.tools.cli)
    (deps-ns-filter ['clojure.tools.cli\/])
    ((fn [ds] (export-graphviz (nodes ds) (edges ds) "tools.cli"))))

I'll throw a pull request your way soon.