move literal-function => abstract, switch fn inheritance order #171
Conversation
sritchie
changed the title
sritchie
changed the title
Codecov Report
@@ Coverage Diff @@
## sritchie/abstract_updates #171 +/- ##
=============================================================
+ Coverage 84.41% 84.50% +0.08%
=============================================================
Files 73 73
Lines 6858 6859 +1
Branches 394 393 -1
=============================================================
+ Hits 5789 5796 +7
+ Misses 675 670 -5
+ Partials 394 393 -1
Continue to review full report at Codecov.
|
sritchie
changed the title
| ;; along with this code; if not, see <http://www.gnu.org/licenses/>. | ||
| ;; | ||
|
|
||
| (ns sicmutils.abstract.function |
There was a problem hiding this comment.
just a move, nothing new.
| @@ -110,6 +110,13 @@ | |||
| (x/expression-of expr) | |||
| expr)) | |||
|
|
|||
| (defn- defunary [generic-op op-sym] | |||
There was a problem hiding this comment.
I moved this up to keep unary => binary order...
| @@ -431,14 +432,11 @@ | |||
| ((d f) (first xs)) | |||
| ((d #(apply f %)) (matrix/seq-> xs))))))) | |||
|
|
|||
| (derive ::x/numeric ::codiff) | |||
There was a problem hiding this comment.
I had introduced this and forgotten that ::scalar was all we really need, no need now for a new ::codiff.
| @@ -493,11 +491,14 @@ | |||
| (defmethod g/partial-derivative [::matrix/matrix v/seqtype] [f selectors] | |||
| (multivariate-derivative f selectors)) | |||
|
|
|||
| (def derivative-symbol 'D) | |||
There was a problem hiding this comment.
I moved this in here so it lives in this one spot.
| @@ -31,6 +32,8 @@ | |||
|
|
|||
| (declare generate) | |||
|
|
|||
| (derive ::matrix ::f/cofunction) | |||
There was a problem hiding this comment.
@littleredcomputer , curious what you think here. I realized that if we were building this up from scratch, we wouldn't have matrices at the point where we added functions... so I moved this derivation over to matrix, so that function.cljc could stay cleaner.
What do you think?
If someone ever wants to use a matrix, they have to require this ns anyway, which will trigger the derive call. And if they don't, no sweat.
| @@ -563,6 +563,7 @@ | |||
|
|
|||
| (derive ::x/numeric ::coseries) | |||
| (derive ::v/number ::coseries) | |||
| (derive ::v/function ::coseries) | |||
There was a problem hiding this comment.
same thought here as in matrix; series should register themselves.
| (derive ::up ::structure) | ||
| (derive ::down ::structure) | ||
| (derive PersistentVector ::up) | ||
|
|
||
| ;; Structures can interact with functions. | ||
| (derive ::structure ::f/cofunction) |
| ;; along with this code; if not, see <http://www.gnu.org/licenses/>. | ||
| ;; | ||
|
|
||
| (ns sicmutils.abstract.function-test |
There was a problem hiding this comment.
these are all the tests that explicitly use literal-function.
| [sicmutils.generic :as g] | ||
| [sicmutils.generators :as sg] | ||
| [sicmutils.matrix :as m] |
There was a problem hiding this comment.
The only tests left here are the ones that did NOT use literal-function. Now we can add more, only about Clojure functions.
This PR
::v/functionand::sicmutils.functionliteral-functionmachinery and tests into their own namespace,sicmutils.abstract.functionsicmutils.functionquite clean. This gets the library closer to a "built from Clojure up" style.::v/function
is now the main thing, and::sicmutils.function/function` extends from this.