cleaned up ns

scicloj · Apr 14, 2023 · 4726d06 · 4726d06
1 parent 5fd5801
commit 4726d06
Show file tree

Hide file tree

Showing 8 changed files with 44 additions and 79 deletions.
diff --git a/src/scicloj/sklearn_clj.clj b/src/scicloj/sklearn_clj.clj
@@ -1,18 +1,18 @@
 (ns scicloj.sklearn-clj
   (:require
    [camel-snake-kebab.core :as csk]
-   [libpython-clj2.require :refer [require-python]]
+   [clojure.string :as str]
+   clojure.walk
    [libpython-clj2.python
     :as py
-    :refer [as-jvm as-python cfn path->py-obj python-type py. py.-]]
+    :refer [cfn path->py-obj python-type py.]]
+   [libpython-clj2.python.np-array]
    [tech.v3.dataset :as ds]
    [tech.v3.dataset.column-filters :as cf]
    [tech.v3.dataset.modelling :as ds-mod]
    [tech.v3.dataset.tensor :as dst]
-   [tech.v3.datatype.errors :as errors]
-   [tech.v3.tensor :as t]
-   [tech.v3.datatype :as dt]
-   [libpython-clj2.python.np-array]))
+   tech.v3.datatype.casting
+   [tech.v3.datatype.errors :as errors]))
 
 
 (defn assert-numeric-ds [ds]
@@ -38,6 +38,7 @@
      (prn (str ~label  " - " "elapsed time: " (/ (double (- (. System (nanoTime)) start#)) 1000000.0) " msecs"))
      ret#))
 
+(py/initialize!)
 (println  "'sklearn' version found: "
           (get
            (py/module-dict (py/import-module "sklearn"))
@@ -186,13 +187,11 @@
   "Calls `predict_proba` on the given sklearn estimator object, and returns the result as a tech.ml.dataset"
   ([ds estimator inference-target-column-names]
    (let
-       [_ (def ds ds)
-        first-target-column-name (first inference-target-column-names)
+       [first-target-column-name (first inference-target-column-names)
         _ (errors/when-not-error first-target-column-name "No inference target column name given")
         feature-ds (cf/feature ds)
-        X  (ds->X feature-ds)
-        prob-prediction (py. estimator predict_proba X)]
-
+        X  (ds->X feature-ds)]
+
      (->
       (py. estimator predict_proba X)
       dst/tensor->dataset
@@ -222,8 +221,8 @@
 (defn model-attribute-names [sklearn-model]
 
   (->> (py/dir sklearn-model)
-       (filter #(and  (clojure.string/ends-with? % "_")
-                      (not (clojure.string/starts-with? % "_"))))))
+       (filter #(and  (str/ends-with? % "_")
+                      (not (str/starts-with? % "_"))))))
 
 (defn save-py-get-attr [sklearn-model attr]
   ;; can fail in some cases
@@ -247,28 +246,9 @@
 
 
 (defn model-attributes [sklearn-model]
-  (def sklearn-model sklearn-model)
   (apply merge
          (map
           (fn [attr]
             (hash-map (keyword attr)
                       (save->jvm (save-py-get-attr sklearn-model attr))))
           (model-attribute-names sklearn-model))))
-
-(comment
-
-  (def big
-    (ds/->dataset
-     (apply merge
-            (map
-             #(hash-map % (seq (range 10000)))
-             (range 10000)))))
-
-  (println
-   (labeled-time
-     :numeric-ds<->ndarray-as
-    (->
-     big
-     numeric-ds->ndarray
-     ndarray->ds
-     (ds/shape)))))
diff --git a/src/scicloj/sklearn_clj/metamorph.clj b/src/scicloj/sklearn_clj/metamorph.clj
@@ -25,11 +25,6 @@
                              (sklearn/transform data estimator kw-args))))))))
 
 
-
-
-
-
-
 (defn fit-transform
   ([module-kw estimator-class-kw]
    (fit-transform module-kw estimator-class-kw {}))

diff --git a/src/scicloj/sklearn_clj/ml.clj b/src/scicloj/sklearn_clj/ml.clj
@@ -5,11 +5,10 @@
    [libpython-clj2.python :refer [->jvm as-jvm cfn path->py-obj py.- py.] :as py]
    [scicloj.metamorph.ml :as ml]
    [scicloj.sklearn-clj :as sklearn]
-   [tech.v3.datatype.functional :as dtf]
    [tech.v3.dataset :as ds]
-   [tech.v3.dataset.modelling :as ds-mod]
-   [tech.v3.dataset.categorical :as ds-cat]))
+   [tech.v3.dataset.modelling :as ds-mod]))
 
+(py/initialize!)
 (def pickle (py/import-module "pickle"))
 (def filter-map
   {"classifier" "classification"
@@ -102,9 +101,6 @@
 
      names)))
 
-
-
-
 (define-estimators! "regressor")
 (define-estimators! "classifier")
 
diff --git a/src/scicloj/sklearn_clj/sklearn_wrapper.clj b/src/scicloj/sklearn_clj/sklearn_wrapper.clj
@@ -4,9 +4,7 @@
    [scicloj.metamorph.core :as mm]
    [scicloj.sklearn-clj]
    [tablecloth.api :as tc]
-   [tech.v3.dataset :as ds]
-   [tech.v3.dataset.categorical :as ds-cat]
-   [tech.v3.dataset.column-filters :as cf]))
+   [tech.v3.dataset :as ds]))
 
 
 (py/initialize!)

diff --git a/test/scicloj/cat_test.clj b/test/scicloj/cat_test.clj
@@ -15,7 +15,7 @@
    [scicloj.sklearn-clj.ml]
    [clojure.test :refer [deftest is]]))
 
-
+(py/initialize!)
 (def builtins (py/import-module "builtins"))
 
 

diff --git a/test/scicloj/metamorph_test.clj b/test/scicloj/metamorph_test.clj
@@ -1,22 +1,20 @@
 (ns scicloj.metamorph-test
   (:require
-   [scicloj.sklearn-clj.ml]
-   [tech.v3.dataset.modelling :as ds-mod]
-   [tech.v3.dataset :as ds]
-   [tech.v3.dataset.tensor :as dst]
-   [tech.v3.tensor :as dtt]
-   [tablecloth.api :as tc]
-   [libpython-clj2.python :refer [->jvm py.- py.
-                                  python-type]]
-   [libpython-clj2.require :refer [require-python]]
-   [scicloj.sklearn-clj.metamorph :as sklearn-mm]
    [clojure.test :refer [deftest is]]
+   [libpython-clj2.python :refer [->jvm]
+    :as py]
+   [libpython-clj2.require :refer [require-python]]
+   [scicloj.metamorph.core :as morph]
    [scicloj.metamorph.ml :as mm-ml]
+   [scicloj.sklearn-clj.metamorph :as sklearn-mm]
+   [scicloj.sklearn-clj.ml]
+   [tablecloth.api :as tc]
+   [tech.v3.dataset :as ds]
    [tech.v3.dataset.column-filters :as ds-cf]
-   [scicloj.metamorph.core :as morph]))
+   [tech.v3.dataset.modelling :as ds-mod]))
 
 
-
+(py/initialize!)
 (deftest test-evaluate
   (let [XY (->
             (tc/dataset [ [-1, -1], [-2, -1], [1, 1], [2, 1]]

diff --git a/test/scicloj/perf_test.clj b/test/scicloj/perf_test.clj
@@ -5,7 +5,7 @@
              [libpython-clj2.python.np-array]
              [scicloj.sklearn-clj :refer [fit]]
              [libpython-clj2.python  :as py]))
-
+(py/initialize!)
 (defmacro labeled-time
   "Evaluates expr and prints the time it took.  Returns the value of
  expr."
@@ -44,5 +44,5 @@
          _ (fit train :sklearn.ensemble :GradientBoostingClassifier)
          end (System/currentTimeMillis)
          clojure-time (/ (- end start) 1000)]
-    (t/is (< python-time 20))
-    (t/is (< clojure-time 20))))
+    (t/is (< python-time 21))
+    (t/is (< clojure-time 21))))
diff --git a/test/scicloj/sklearn_clj_test.clj b/test/scicloj/sklearn_clj_test.clj
@@ -1,23 +1,21 @@
 (ns scicloj.sklearn-clj-test
-  (:require [clojure.test :refer :all]
-            [tech.v3.dataset :as ds]
-            [tech.v3.dataset.modelling :as ds-mod]
-            [tablecloth.api :as tc]
-            [tech.v3.dataset.tensor :as dst]
-            [tech.v3.tensor :as tensor]
-            [libpython-clj2.python :refer [python-type] :as py]
-            [tech.v3.datatype.functional :as f]
-            [scicloj.metamorph.ml.toydata :as toydata]
-            [scicloj.metamorph.ml :as ml]
-            [scicloj.metamorph.core :as mm]
-            [scicloj.sklearn-clj :refer :all]
-            [tablecloth.pipeline :as tcm]
-            [tech.v3.dataset.metamorph :as ds-mm]
-            [tech.v3.dataset.column-filters :as cf]
-            [scicloj.sklearn-clj.ml]))
+  (:require
+   [clojure.test :refer :all]
+   [libpython-clj2.python :refer [python-type] :as py]
+   [scicloj.metamorph.core :as mm]
+   [scicloj.metamorph.ml :as ml]
+   [scicloj.metamorph.ml.toydata :as toydata]
+   [scicloj.sklearn-clj :refer :all]
+   [scicloj.sklearn-clj.ml]
+   [tablecloth.api :as tc]
+   [tech.v3.dataset :as ds]
+   [tech.v3.dataset.modelling :as ds-mod]
+   [tech.v3.dataset.tensor :as dst]
+   [tech.v3.datatype.functional :as f]
+   [tech.v3.tensor :as tensor]))
 
 
-
+(py/initialize!)
 (def texts ["This is the first document."
                "This document is the second document."
                "And this is the third one."