Testing/dump load for each module (#192)

* add tests for decision modules

* fix codestyle

* adjust how transformers and tokenizers are saved

* add test for generic dumper

* add tests for embedding modules

* add tests for scoring modules

* add test for regex

* make `load` a classmethod

* try to fix file exists error

* try to fix pydantic error

* try to fix windows cleanup error

Author: voorhs

autointent/_dump_tools.py

@@ -20,6 +20,7 @@
 
 from autointent import Embedder, Ranker, VectorIndex
 from autointent.configs import CrossEncoderConfig, EmbedderConfig
+from autointent.context.optimization_info import Artifact
 from autointent.schemas import TagsList
 
 ModuleSimpleAttributes = None | str | int | float | bool | list  # type: ignore[type-arg]
@@ -83,7 +84,7 @@ def dump(obj: Any, path: Path, exists_ok: bool = False, exclude: list[type[Any]]
         Dumper.make_subdirectories(path, exists_ok)
 
         for key, val in attrs.items():
-            if exclude and isinstance(val, tuple(exclude)):
+            if isinstance(val, Artifact) or (exclude and isinstance(val, tuple(exclude))):
                 continue
             if isinstance(val, TagsList):
                 val.dump(path / Dumper.tags / key)

autointent/modules/base/_base.py

@@ -8,7 +8,7 @@
 
 import numpy as np
 import numpy.typing as npt
-from typing_extensions import assert_never
+from typing_extensions import Self, assert_never
 
 from autointent._dump_tools import Dumper
 from autointent.configs import CrossEncoderConfig, EmbedderConfig
@@ -83,20 +83,23 @@ def dump(self, path: str) -> None:
         """
         Dumper.dump(self, Path(path))
 
+    @classmethod
     def load(
-        self,
+        cls,
         path: str,
         embedder_config: EmbedderConfig | None = None,
         cross_encoder_config: CrossEncoderConfig | None = None,
-    ) -> None:
+    ) -> Self:
         """Load data from file system.
 
         Args:
             path: Path to load
             embedder_config: one can override presaved settings
             cross_encoder_config: one can override presaved settings
         """
-        Dumper.load(self, Path(path), embedder_config=embedder_config, cross_encoder_config=cross_encoder_config)
+        instance = cls()
+        Dumper.load(instance, Path(path), embedder_config=embedder_config, cross_encoder_config=cross_encoder_config)
+        return instance
 
     @abstractmethod
     def predict(

autointent/modules/embedding/_logreg.py

@@ -50,7 +50,7 @@ class LogregAimedEmbedding(BaseEmbedding):
 
     def __init__(
         self,
-        embedder_config: EmbedderConfig | str | dict[str, Any],
+        embedder_config: EmbedderConfig | str | dict[str, Any] | None = None,
         cv: PositiveInt = 3,
     ) -> None:
         self.embedder_config = EmbedderConfig.from_search_config(embedder_config)
@@ -64,7 +64,7 @@ def __init__(
     def from_context(
         cls,
         context: Context,
-        embedder_config: EmbedderConfig | str,
+        embedder_config: EmbedderConfig | str | None = None,
         cv: PositiveInt = 3,
     ) -> "LogregAimedEmbedding":
         """Create a LogregAimedEmbedding instance using a Context object.

autointent/modules/embedding/_retrieval.py

@@ -46,7 +46,7 @@ class RetrievalAimedEmbedding(BaseEmbedding):
 
     def __init__(
         self,
-        embedder_config: EmbedderConfig | str | dict[str, Any],
+        embedder_config: EmbedderConfig | str | dict[str, Any] | None = None,
         k: PositiveInt = 10,
     ) -> None:
         self.k = k
@@ -61,7 +61,7 @@ def __init__(
     def from_context(
         cls,
         context: Context,
-        embedder_config: EmbedderConfig | str,
+        embedder_config: EmbedderConfig | str | None = None,
         k: PositiveInt = 10,
     ) -> "RetrievalAimedEmbedding":
         """Create an instance using a Context object.

autointent/modules/regex/_simple.py

@@ -253,13 +253,17 @@ def dump(self, path: str) -> None:
         with (dump_dir / "regex_patterns.json").open("w") as file:
             json.dump(serialized, file, indent=4, ensure_ascii=False)
 
+    @classmethod
     def load(
-        self,
+        cls,
         path: str,
         embedder_config: EmbedderConfig | None = None,
         cross_encoder_config: CrossEncoderConfig | None = None,
-    ) -> None:
+    ) -> "SimpleRegex":
+        instance = cls()
+
         with (Path(path) / "regex_patterns.json").open() as file:
             serialized: list[dict[str, Any]] = json.load(file)
 
-        self._compile_regex_patterns(serialized)
+        instance._compile_regex_patterns(serialized)  # noqa: SLF001
+        return instance

autointent/modules/scoring/_dnnc/dnnc.py

@@ -61,7 +61,7 @@ class DNNCScorer(BaseScorer):
 
     def __init__(
         self,
-        k: PositiveInt,
+        k: PositiveInt = 5,
         cross_encoder_config: CrossEncoderConfig | str | dict[str, Any] | None = None,
         embedder_config: EmbedderConfig | str | dict[str, Any] | None = None,
     ) -> None:
@@ -77,7 +77,7 @@ def __init__(
     def from_context(
         cls,
         context: Context,
-        k: PositiveInt,
+        k: PositiveInt = 5,
         cross_encoder_config: CrossEncoderConfig | str | None = None,
         embedder_config: EmbedderConfig | str | None = None,
     ) -> "DNNCScorer":

autointent/modules/scoring/_knn/knn.py

@@ -56,7 +56,7 @@ class KNNScorer(BaseScorer):
 
     def __init__(
         self,
-        k: PositiveInt,
+        k: PositiveInt = 5,
         embedder_config: EmbedderConfig | str | dict[str, Any] | None = None,
         weights: WeightType = "distance",
     ) -> None:
@@ -76,7 +76,7 @@ def __init__(
     def from_context(
         cls,
         context: Context,
-        k: PositiveInt,
+        k: PositiveInt = 5,
         weights: WeightType = "distance",
         embedder_config: EmbedderConfig | str | None = None,
     ) -> "KNNScorer":

autointent/modules/scoring/_knn/rerank_scorer.py

@@ -36,10 +36,10 @@ class RerankScorer(KNNScorer):
 
     def __init__(
         self,
-        k: int,
+        k: PositiveInt = 5,
         weights: WeightType = "distance",
         use_crosencoder_scores: bool = False,
-        m: int | None = None,
+        m: PositiveInt | None = None,
         cross_encoder_config: CrossEncoderConfig | str | dict[str, Any] | None = None,
         embedder_config: EmbedderConfig | str | dict[str, Any] | None = None,
     ) -> None:
@@ -62,7 +62,7 @@ def __init__(
     def from_context(
         cls,
         context: Context,
-        k: int,
+        k: PositiveInt = 5,
         weights: WeightType = "distance",
         m: PositiveInt | None = None,
         cross_encoder_config: CrossEncoderConfig | str | None = None,

autointent/modules/scoring/_mlknn/mlknn.py

@@ -63,7 +63,7 @@ class MLKnnScorer(BaseScorer):
 
     def __init__(
         self,
-        k: PositiveInt,
+        k: PositiveInt = 5,
         embedder_config: EmbedderConfig | str | dict[str, Any] | None = None,
         s: float = 1.0,
         ignore_first_neighbours: int = 0,
@@ -84,7 +84,7 @@ def __init__(
     def from_context(
         cls,
         context: Context,
-        k: PositiveInt,
+        k: PositiveInt = 5,
         s: PositiveFloat = 1.0,
         ignore_first_neighbours: NonNegativeInt = 0,
         embedder_config: EmbedderConfig | str | None = None,

autointent/modules/scoring/_sklearn/sklearn_scorer.py

@@ -59,7 +59,7 @@ class SklearnScorer(BaseScorer):
 
     def __init__(
         self,
-        clf_name: str,
+        clf_name: str = "LogisticRegression",
         embedder_config: EmbedderConfig | str | dict[str, Any] | None = None,
         **clf_args: Any,  # noqa: ANN401
     ) -> None:
@@ -83,7 +83,7 @@ def __init__(
     def from_context(
         cls,
         context: Context,
-        clf_name: str,
+        clf_name: str = "LogisticRegression",
         embedder_config: EmbedderConfig | str | None = None,
         **clf_args: float | str | bool,
     ) -> Self: