refactor(api): move fit into top-most init (#84)

finetuner/__init__.py

@@ -1,8 +1,65 @@
-__default_tag_key__ = 'finetuner'
-
-from .fit import fit
-
 # do not change this line manually
 # this is managed by git tag and updated on every release
 # NOTE: this represents the NEXT release version
 __version__ = '0.0.3'
+
+__default_tag_key__ = 'finetuner'
+
+# define the high-level API: fit()
+from typing import Optional, overload, TYPE_CHECKING
+
+
+from .helper import AnyDNN, DocumentArrayLike
+from .tuner.fit import TunerReturnType
+
+# fit interface generated from Labeler + Tuner
+# overload_inject_fit_tailor_tuner_start
+
+# overload_inject_fit_tailor_tuner_end
+
+
+# fit interface generated from Labeler + Tuner
+# overload_inject_fit_labeler_tuner_start
+@overload
+def fit(
+    embed_model: AnyDNN,
+    train_data: DocumentArrayLike,
+    clear_labels_on_start: bool = False,
+    port_expose: Optional[int] = None,
+    runtime_backend: str = 'thread',
+    interactive: bool = True,
+    head_layer: str = 'CosineLayer',
+) -> None:
+    ...
+
+
+# overload_inject_fit_labeler_tuner_end
+
+
+# fit interface generated from Tuner
+# overload_inject_fit_tuner_start
+@overload
+def fit(
+    embed_model: AnyDNN,
+    train_data: DocumentArrayLike,
+    eval_data: Optional[DocumentArrayLike] = None,
+    epochs: int = 10,
+    batch_size: int = 256,
+    head_layer: str = 'CosineLayer',
+) -> TunerReturnType:
+    ...
+
+
+# overload_inject_fit_tuner_end
+
+
+def fit(*args, **kwargs) -> Optional[TunerReturnType]:
+    if kwargs.get('interactive', False):
+        kwargs.pop('interactive')
+        from .labeler.fit import fit
+
+        return fit(*args, **kwargs)
+    else:
+        from .tuner.fit import fit
+
+        return fit(*args, **kwargs)

finetuner/labeler/fit.py

@@ -26,22 +26,25 @@ def get_embed_model(self):
             return embed_model
 
     f = (
-        Flow(protocol='http', port_expose=port_expose, prefetch=1)
+        Flow(
+            protocol='http',
+            port_expose=port_expose,
+            prefetch=1,
+            runtime_backend=runtime_backend,
+        )
         .add(
             uses=DataIterator,
             uses_with={
                 'dam_path': dam_path,
                 'clear_labels_on_start': clear_labels_on_start,
             },
-            runtime_backend=runtime_backend,
         )
         .add(
             uses=MyExecutor,
             uses_with={
                 'dam_path': dam_path,
                 'head_layer': head_layer,
             },
-            runtime_backend=runtime_backend,  # eager-mode tf2 (M1-compiled) can not be run under `process` mode
         )
     )
 

tests/integration/fit/test_fit_lstm.py

@@ -4,7 +4,7 @@
 import tensorflow as tf
 import torch
 
-import finetuner as jft
+from finetuner import fit
 from finetuner.toydata import generate_qa_match
 
 
@@ -45,7 +45,7 @@ def test_fit_all(tmpdir):
 
     for kb, b in embed_models.items():
         for h in ['CosineLayer', 'TripletLayer']:
-            result = jft.fit(
+            result = fit(
                 b(),
                 head_layer=h,
                 train_data=lambda: generate_qa_match(