On 7/21/12 7:33 AM, ChrisR wrote:
I am wondering if there is a simpler idiom to deal with functions on
primitives that only differ by type hints (see example), this is
the current idiom I am using which is starting to appear in my
codebase, however I feel like whenever this much replication
occurs I am missing a key point somewhere in my learning of Clojure.
Any advice appreciated.
(:import (cern.colt.matrix.tfloat.impl
DenseFloatMatrix2D SelectedDenseFloatMatrix2D))
(defmulti aggregate-rows class)
(defmethod aggregate-rows DenseFloatMatrix2D [matrix]
(let [n-rows (.rows ^DenseFloatMatrix2D matrix)
n-columns (.columns ^DenseFloatMatrix2D matrix)
aggregate (float-array n-rows 0)]
(dotimes [i n-rows]
(dotimes [j n-columns]
(aset aggregate i
(unchecked-add (aget aggregate i)
(.getQuick ^DenseFloatMatrix2D matrix i j)))))
aggregate))
(defmethod aggregate-rows SelectedDenseFloatMatrix2D [matrix]
(let [n-rows (.rows ^SelectedDenseFloatMatrix2D matrix)
n-columns (.columns ^SelectedDenseFloatMatrix2D matrix)
aggregate (float-array n-rows 0)]
(dotimes [i n-rows]
(dotimes [j n-columns]
(aset aggregate i
(unchecked-add (aget aggregate i)
(.getQuick ^SelectedDenseFloatMatrix2D matrix i j)))))
aggregate))
I've noticed that when wrapping java libs like this you sometimes have
to live with a certain amount of duplication if you want/need type
hints. However, if the java lib is designed well with interfaces you
can get away with type hinting with the shared interface (or superclass
as Dennis already mentioned).
Kinda off topic, but I have some colt wrappers on github if you want to
take a look at the approach I took:
https://github.com/bmabey/claw/blob/master/src/claw/core.clj
You'll notice that I had to have some duplication as well for type
hinting in my protocols. Once I wrapped each matrix type in the main
protocols all my main API functions can call them uniformly without
having to worry about type hints.
BTW, the best place to see how to use the wrappers are the tests (midje
facts), which includes an example on how I would my #'reduce-rows
function to do what your #'aggregate-rows is doing:
https://github.com/bmabey/claw/blob/master/test/claw/test/core.clj#L110
-Ben
--
You received this message because you are subscribed to the Google
Groups "Clojure" group.
To post to this group, send email to clojure@googlegroups.com
Note that posts from new members are moderated - please be patient with your
first post.
To unsubscribe from this group, send email to
clojure+unsubscr...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/clojure?hl=en
