handle floating point values more accurate

README.md

@@ -131,3 +131,7 @@ A: If this error occurs while generating code using an ensemble model, try to re
 **Q: Generation fails with `ImportError: No module named <module_name_here>` error while transpiling model from a serialized model object.**
 
 A: This error indicates that pickle protocol cannot deserialize model object. For unpickling serialized model objects, it is required that their classes must be defined in the top level of an importable module in the unpickling environment. So installation of package which provided model's class definition should solve the problem.
+
+**Q: Generated by m2cgen code provides different results for some inputs compared to original Python model from which the code were obtained.**
+
+A: Some models force input data to be particular type during prediction phase in their native Python libraries. Currently, m2cgen works only with ``float64`` (``double``) data type. You can try to cast your input data to another type manually and check results again. Also, some small differences can happen due to specific implementation of floating-point arithmetic in a target language.

m2cgen/assemblers/boosting.py

@@ -151,7 +151,7 @@ def __init__(self, model):
 
     def _assemble_tree(self, tree):
         if "leaf" in tree:
-            return ast.NumVal(tree["leaf"])
+            return ast.NumVal(tree["leaf"], dtype=np.float32)
 
         threshold = ast.NumVal(tree["split_condition"], dtype=np.float32)
         split = tree["split"]

m2cgen/assemblers/tree.py

@@ -1,5 +1,3 @@
-import numpy as np
-
 from m2cgen import ast
 from m2cgen.assemblers import utils
 from m2cgen.assemblers.base import ModelAssembler
@@ -49,11 +47,5 @@ def _assemble_leaf(self, node_id):
 
     def _assemble_cond(self, node_id):
         feature_idx = self._tree.feature[node_id]
-        threshold = self._tree.threshold[node_id]
-
-        # sklearn's trees internally work with float32 numbers, so in order
-        # to have consistent results across all supported languages, we convert
-        # all thresholds into float32.
-        threshold_num_val = ast.NumVal(threshold, dtype=np.float32)
-
+        threshold_num_val = ast.NumVal(self._tree.threshold[node_id])
         return utils.lte(ast.FeatureRef(feature_idx), threshold_num_val)

m2cgen/interpreters/code_generator.py

@@ -1,6 +1,10 @@
 from io import StringIO
 from weakref import finalize
 
+import numpy as np
+
+from m2cgen.interpreters.utils import format_float
+
 
 class CodeTemplate:
 
@@ -11,12 +15,14 @@ def __str__(self):
         return self.str_template
 
     def __call__(self, *args, **kwargs):
-        # Force calling str() representation
-        # because without it numpy gives the same output
-        # for different float types
+
+        def _is_float(value):
+            return isinstance(value, (float, np.floating))
+
         return self.str_template.format(
-            *[str(i) for i in args],
-            **{k: str(v) for k, v in kwargs.items()})
+            *[format_float(i) if _is_float(i) else i for i in args],
+            **{k: format_float(v) if _is_float(v) else v
+               for k, v in kwargs.items()})
 
 
 class BaseCodeGenerator:

m2cgen/interpreters/utils.py

@@ -1,5 +1,7 @@
 import re
 
+import numpy as np
+
 from collections import namedtuple
 from functools import lru_cache
 from math import ceil, log
@@ -22,3 +24,7 @@ def _get_handler_name(expr_tpe):
 
 def _normalize_expr_name(name):
     return re.sub("(?!^)([A-Z]+)", r"_\1", name).lower()
+
+
+def format_float(value):
+    return np.format_float_positional(value, unique=True, trim="0")

tests/e2e/executors/c.py

@@ -54,7 +54,7 @@ def __init__(self, model):
     def predict(self, X):
 
         exec_args = [os.path.join(self._resource_tmp_dir, self.model_name)]
-        exec_args.extend(map(str, X))
+        exec_args.extend(map(interpreters.utils.format_float, X))
         return utils.predict_from_commandline(exec_args)
 
     def prepare(self):

tests/e2e/executors/c_sharp.py

@@ -49,7 +49,7 @@ def __init__(self, model):
 
     def predict(self, X):
         exec_args = [os.path.join(self.target_exec_dir, self.project_name)]
-        exec_args.extend(map(str, X))
+        exec_args.extend(map(interpreters.utils.format_float, X))
         return utils.predict_from_commandline(exec_args)
 
     @classmethod

tests/e2e/executors/dart.py

@@ -42,7 +42,7 @@ def predict(self, X):
                                  f"{self.executor_name}.dart")
         exec_args = [self._dart,
                      file_name,
-                     *map(str, X)]
+                     *map(interpreters.utils.format_float, X)]
         return utils.predict_from_commandline(exec_args)
 
     def prepare(self):

tests/e2e/executors/f_sharp.py

@@ -35,7 +35,7 @@ def __init__(self, model):
 
     def predict(self, X):
         exec_args = [os.path.join(self.target_exec_dir, self.project_name)]
-        exec_args.extend(map(str, X))
+        exec_args.extend(map(interpreters.utils.format_float, X))
         return utils.predict_from_commandline(exec_args)
 
     @classmethod

tests/e2e/executors/go.py

@@ -55,7 +55,7 @@ def __init__(self, model):
     def predict(self, X):
 
         exec_args = [os.path.join(self._resource_tmp_dir, self.model_name)]
-        exec_args.extend(map(str, X))
+        exec_args.extend(map(interpreters.utils.format_float, X))
         return utils.predict_from_commandline(exec_args)
 
     def prepare(self):

tests/e2e/executors/haskell.py

@@ -39,7 +39,8 @@ def __init__(self, model):
     def predict(self, X):
         app_name = os.path.join(self._resource_tmp_dir,
                                 self.executor_name)
-        exec_args = [app_name, *map(str, X)]
+        exec_args = [app_name,
+                     *map(interpreters.utils.format_float, X)]
         return utils.predict_from_commandline(exec_args)
 
     def prepare(self):

tests/e2e/executors/java.py

@@ -24,7 +24,7 @@ def predict(self, X):
             self._java_bin, "-cp", self._resource_tmp_dir,
             "Executor", "Model", "score"
         ]
-        exec_args.extend(map(str, X))
+        exec_args.extend(map(m2c.interpreters.utils.format_float, X))
         return utils.predict_from_commandline(exec_args)
 
     def prepare(self):

tests/e2e/executors/javascript.py

@@ -1,4 +1,5 @@
 import os
+
 from py_mini_racer import py_mini_racer
 
 import m2cgen as m2c
@@ -16,10 +17,11 @@ def predict(self, X):
         with open(file_name, 'r') as myfile:
             code = myfile.read()
 
-        caller = f"score([{','.join(map(str, X))}]);\n"
+        args = ",".join(map(m2c.interpreters.utils.format_float, X))
+        caller = f"score([{args}]);\n"
 
         ctx = py_mini_racer.MiniRacer()
-        result = ctx.eval(caller + code)
+        result = ctx.eval(f"{caller}{code}")
 
         return result
 

tests/e2e/executors/php.py

@@ -47,7 +47,7 @@ def predict(self, X):
         exec_args = [self._php,
                      "-f",
                      file_name,
-                     *map(str, X)]
+                     *map(interpreters.utils.format_float, X)]
         return utils.predict_from_commandline(exec_args)
 
     def prepare(self):

tests/e2e/executors/powershell.py

@@ -40,7 +40,7 @@ def predict(self, X):
                      "-File",
                      file_name,
                      "-InputArray",
-                     ",".join(map(str, X))]
+                     ",".join(map(interpreters.utils.format_float, X))]
         return utils.predict_from_commandline(exec_args)
 
     def prepare(self):

tests/e2e/executors/r.py

@@ -34,7 +34,7 @@ def predict(self, X):
         exec_args = [self._r,
                      "--vanilla",
                      file_name,
-                     *map(str, X)]
+                     *map(interpreters.utils.format_float, X)]
         return utils.predict_from_commandline(exec_args)
 
     def prepare(self):

tests/e2e/executors/ruby.py

@@ -38,7 +38,9 @@ def __init__(self, model):
     def predict(self, X):
         file_name = os.path.join(self._resource_tmp_dir,
                                  f"{self.model_name}.rb")
-        exec_args = [self._ruby, file_name, *map(str, X)]
+        exec_args = [self._ruby,
+                     file_name,
+                     *map(interpreters.utils.format_float, X)]
         return utils.predict_from_commandline(exec_args)
 
     def prepare(self):

tests/e2e/executors/visual_basic.py

@@ -51,7 +51,7 @@ def __init__(self, model):
 
     def predict(self, X):
         exec_args = [os.path.join(self.target_exec_dir, self.project_name)]
-        exec_args.extend(map(str, X))
+        exec_args.extend(map(interpreters.utils.format_float, X))
         return utils.predict_from_commandline(exec_args)
 
     @classmethod

tests/e2e/test_e2e.py

@@ -568,6 +568,8 @@ def test_e2e(estimator, executor_cls, model_trainer,
     with executor.prepare_then_cleanup():
         for idx in idxs_to_test:
             y_pred_executed = executor.predict(X_test[idx])
+            y_pred_executed = np.array(
+                y_pred_executed, dtype=y_pred_true.dtype, copy=False)
             print(f"expected={y_pred_true[idx]}, actual={y_pred_executed}")
             res = np.isclose(y_pred_true[idx], y_pred_executed, atol=ATOL)
             assert res if isinstance(res, bool) else res.all()

tests/utils.py

@@ -13,11 +13,13 @@
 from lightning.impl.base import BaseClassifier as LightBaseClassifier
 from sklearn import datasets
 from sklearn.base import BaseEstimator, RegressorMixin, clone
-from sklearn.ensemble._forest import ForestClassifier
+from sklearn.ensemble._forest import ForestClassifier, BaseForest
 from sklearn.utils import shuffle
 from sklearn.linear_model._base import LinearClassifierMixin
 from sklearn.tree import DecisionTreeClassifier
+from sklearn.tree._classes import BaseDecisionTree
 from sklearn.svm import SVC, NuSVC
+from sklearn.svm._base import BaseLibSVM
 from xgboost import XGBClassifier
 
 from m2cgen import ast
@@ -125,15 +127,22 @@ def __call__(self, estimator):
         if isinstance(estimator, (LinearClassifierMixin, SVC, NuSVC,
                                   LightBaseClassifier)):
             y_pred = estimator.decision_function(self.X_test)
-        elif isinstance(estimator, DecisionTreeClassifier):
-            y_pred = estimator.predict_proba(self.X_test.astype(np.float32))
         elif isinstance(
                 estimator,
-                (ForestClassifier, XGBClassifier, LGBMClassifier)):
+                (ForestClassifier, DecisionTreeClassifier,
+                 XGBClassifier, LGBMClassifier)):
             y_pred = estimator.predict_proba(self.X_test)
         else:
             y_pred = estimator.predict(self.X_test)
 
+        # Some models force input data to be particular type
+        # during prediction phase in their native Python libraries.
+        # For correct comparison of testing results we mimic the same behavior
+        if isinstance(estimator, (BaseDecisionTree, BaseForest)):
+            self.X_test = self.X_test.astype(np.float32, copy=False)
+        elif isinstance(estimator, BaseLibSVM):
+            self.X_test = self.X_test.astype(np.float64, copy=False)
+
         return self.X_test, y_pred, fitted_estimator
 
 
@@ -238,9 +247,9 @@ def predict_from_commandline(exec_args):
     items = stdout.decode("utf-8").strip().split(" ")
 
     if len(items) == 1:
-        return float(items[0])
+        return np.float64(items[0])
     else:
-        return [float(i) for i in items]
+        return [np.float64(i) for i in items]
 
 
 def cartesian_e2e_params(executors_with_marks, models_with_trainers_with_marks,
@@ -284,4 +293,4 @@ def inner(*args, **kwarg):
 
 
 def _is_float(value):
-    return isinstance(value, (float, np.float16, np.float32, np.float64))
+    return isinstance(value, (float, np.floating))