code-loader 1.0.206.dev1__py3-none-any.whl → 1.0.206.dev3__py3-none-any.whl

code_loader/contract/datasetclasses.py

@@ -143,6 +143,10 @@ class ElementInstance:
     name: str
     mask: npt.NDArray[np.float32]
     instance_metadata: Optional[Dict[str, Union[Optional[str], int, bool, Optional[float]]]] = None
+    # Per-key "this value is absent", the same pair shape DatasetSample carries. Populated at the
+    # masks funnel (utils.to_numpy_return_masks_wrapper) so the engine copies flags rather than
+    # deriving them, exactly as it does for sample-level metadata.
+    instance_metadata_is_none: Optional[Dict[str, bool]] = None
 
 SectionCallableInterface = Callable[[Union[int, str], PreprocessResponse], npt.NDArray[np.float32]]
 InstanceCallableInterface = Callable[[Union[int, str], PreprocessResponse, int], Optional[ElementInstance]]

code_loader/inner_leap_binder/leapbinder_decorators.py

@@ -15,7 +15,7 @@ from typing import Optional, Union, Callable, List, Dict, get_args, get_origin,
 import numpy as np
 import numpy.typing as npt
 
-from code_loader.utils import get_metadata_type_from_variable, map_dict_to_metadata_types, \
+from code_loader.utils import get_metadata_type_from_variable, is_absent_metadata_value, \
     validate_autoregressive_state_types, autoregressive_nests_equal, \
     simulate_engine_float16_downcast_on_call_args, ENGINE_STORAGE_DTYPE, \
     TL_DISABLE_ENGINE_FLOAT16_SIMULATION_ENV_VAR
@@ -27,7 +27,8 @@ from code_loader.contract.datasetclasses import CustomCallableInterfaceMultiArgs
     VisualizerCallableInterface, MetadataSectionCallableInterface, PreprocessResponse, SectionCallableInterface, \
     ConfusionMatrixElement, SamplePreprocessResponse, PredictionTypeHandler, InstanceCallableInterface, ElementInstance, \
     InstanceLengthCallableInterface, InstanceSectionCallableInterface, AutoregressiveStepCallableInterface
-from code_loader.contract.enums import MetricDirection, LeapDataType, DatasetMetadataType, DataStateType
+from code_loader.contract.enums import MetricDirection, LeapDataType, DatasetMetadataType, DataStateType, \
+    DataStateEnum
 from code_loader import leap_binder, LeapLoader
 from code_loader.contract.mapping import NodeMapping, NodeMappingType, NodeConnection
 from code_loader.contract.visualizer_classes import LeapImage, LeapImageMask, LeapTextMask, LeapText, LeapGraph, \
@@ -3003,28 +3004,54 @@ def tensorleap_element_instance_preprocess(
             leap_binder.set_metadata(builtin_instance_extra_metadata,
                                      "builtin_instance_extra_metadata", metadata_types)
 
-        def probe_instance_metadata_types(sample_id: Union[int, str], preprocess_response: PreprocessResponse,
-                                          idx: int) -> Optional[Dict[str, DatasetMetadataType]]:
-            # Fallback when the types are not declared: infer them from one probed instance. A
-            # None there carries no type, so name the fields and point at the declaration.
+        def resolve_instance_metadata_types(sample_id: Union[int, str], preprocess_response: PreprocessResponse,
+                                            idx: int) -> Optional[Dict[str, DatasetMetadataType]]:
+            # The probe enumerates the KEYS; the declaration, where given, supplies the TYPE. A key
+            # that is absent on the probed instance and undeclared has no type available from either
+            # side -- fail with mechanism (a)'s own message so both paths read the same.
             element_instance = instance_mask_encoder(sample_id, preprocess_response, idx)
             element_instance_metadata = getattr(element_instance, 'instance_metadata', None)
             if element_instance_metadata is None:
-                return None
-            not_inferable = [name for name, value in element_instance_metadata.items() if value is None]
-            if not_inferable:
-                raise Exception(
-                    f"tensorleap_element_instance_preprocess validation failed: cannot infer the "
-                    f"type of instance metadata {not_inferable} from instance {idx} of sample "
-                    f"{sample_id!r}, because it is None on that instance. The type cannot come from "
-                    f"the value, so declare it: tensorleap_element_instance_preprocess(..., "
-                    f"instance_metadata_types={{'{not_inferable[0]}': DatasetMetadataType.float}}).")
-            return map_dict_to_metadata_types(element_instance_metadata)
+                return dict(instance_metadata_types) if instance_metadata_types else None
+            declared = instance_metadata_types or {}
+            resolved = {}
+            for name, value in element_instance_metadata.items():
+                if name in declared:
+                    # Still type-check the VALUE. Master ran every key through
+                    # map_dict_to_metadata_types, so declaring a key must not become a way to
+                    # smuggle a list or an ndarray past validation and have it fail at eval time.
+                    if not is_absent_metadata_value(value):
+                        get_metadata_type_from_variable(value)
+                    continue
+                if is_absent_metadata_value(value):
+                    raise Exception(
+                        f"Metadata {name} is None and no metadata type is provided. It is absent on "
+                        f"instance {idx} of sample {sample_id!r}, so the type cannot come from the "
+                        f"value -- declare it: tensorleap_element_instance_preprocess(..., "
+                        f"instance_metadata_types={{'{name}': DatasetMetadataType.float}}).")
+                resolved[name] = get_metadata_type_from_variable(value)
+            # Declared wins, and a declared key gets a column even when the probed instance lacks it.
+            resolved.update(declared)
+            return resolved
 
         def user_function_instance() -> List[PreprocessResponse]:
             result = user_function()
-            found_instance_metadata = instance_metadata_types is not None
-            for preprocess_response in result:
+            found_instance_metadata = False
+            # Resolve types on training, like mechanism (a) does (leapbinder.get_preprocess_result):
+            # only training is allowed to fail on an unresolvable absence. `state` is still None here
+            # whenever the integration did not set it, so fall back to list position -- which is the
+            # state get_preprocess_result would assign, and position 0 is training. Rank training
+            # against everything else rather than sorting on the state value itself: mixing enum
+            # values with list positions in one key lets a validation response tie with position 1
+            # and keep its place ahead of training.
+            def _is_training(i: int) -> bool:
+                response = result[i]
+                if response.state is not None:
+                    return response.state == DataStateType.training
+                return i == 0
+
+            probe_order = sorted(range(len(result)), key=lambda i: (0 if _is_training(i) else 1, i))
+            for preprocess_response in (result[i] for i in probe_order):
                 if preprocess_response.is_grouped:
                     raise Exception(
                         "tensorleap_element_instance_preprocess validation failed: a grouped "
@@ -3047,10 +3074,10 @@ def tensorleap_element_instance_preprocess(
                     # "Index <id> with sample_id: <id> cannot be found!".
                     for idx, instance_id in enumerate(instances_ids):
                         if not found_instance_metadata:
-                            probed_types = probe_instance_metadata_types(
+                            resolved_types = resolve_instance_metadata_types(
                                 sample_id, preprocess_response, idx)
-                            if probed_types is not None:
-                                register_instance_extra_metadata(probed_types)
+                            if resolved_types:
+                                register_instance_extra_metadata(resolved_types)
                             found_instance_metadata = True
 
                         instance_to_sample_ids_mappings[instance_id] = sample_id
@@ -3066,8 +3093,6 @@ def tensorleap_element_instance_preprocess(
 
         leap_binder.set_preprocess(user_function_instance)
         leap_binder.set_metadata(builtin_instance_metadata, "builtin_instance_metadata")
-        if instance_metadata_types is not None:
-            register_instance_extra_metadata(instance_metadata_types)
 
         def _validate_input_args(*args, **kwargs):
             assert len(args) == 0 and len(kwargs) == 0, \

code_loader/leaploader.py

@@ -28,7 +28,7 @@ from code_loader.inner_leap_binder import global_leap_binder
 from code_loader.inner_leap_binder.leapbinder import mapping_runtime_mode_env_var_mame
 from code_loader.leaploaderbase import LeapLoaderBase
 from code_loader.utils import get_root_exception_file_and_line_number, get_metadata_type_from_variable, \
-    validate_autoregressive_state_types, autoregressive_nests_equal
+    validate_autoregressive_state_types, autoregressive_nests_equal, is_absent_metadata_value
 
 
 def _serialize_sim_bounds(bounds) -> dict:
@@ -1228,7 +1228,7 @@ class LeapLoader(LeapLoaderBase):
 
         try:
             is_none = False
-            if value is None or (isinstance(value, (int, float, np.integer, np.floating)) and not np.isfinite(value)):
+            if is_absent_metadata_value(value):
                 raise ValueError()
             converted_value = metadata_type_to_python_type[metadata_name_to_type[metadata_name]](value)
         except ValueError:

code_loader/utils.py

@@ -15,6 +15,12 @@ from code_loader.contract.datasetclasses import SectionCallableInterface, Prepro
 from code_loader.contract.enums import DatasetMetadataType
 
 
+def is_absent_metadata_value(value: Any) -> bool:
+    """The one definition of an absent metadata value, shared by every mechanism."""
+    return value is None or (isinstance(value, (int, float, np.integer, np.floating))
+                             and not np.isfinite(value))
+
+
 def to_numpy_return_wrapper(encoder_function: SectionCallableInterface) -> SectionCallableInterface:
     def numpy_encoder_function(idx: Union[int, str], samples: PreprocessResponse) -> npt.NDArray[np.float32]:
         result = encoder_function(idx, samples)
@@ -39,6 +45,9 @@ def to_numpy_return_masks_wrapper(encoder_function: InstanceCallableInterface) -
         if result is None:
             return None
         result.mask = np.array(result.mask)
+        if result.instance_metadata is not None and result.instance_metadata_is_none is None:
+            result.instance_metadata_is_none = {
+                k: is_absent_metadata_value(v) for k, v in result.instance_metadata.items()}
         return result
     return numpy_encoder_function
 

{code_loader-1.0.206.dev1.dist-info → code_loader-1.0.206.dev3.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: code-loader
-Version: 1.0.206.dev1
+Version: 1.0.206.dev3
 Summary: 
 Home-page: https://github.com/tensorleap/code-loader
 License: MIT

{code_loader-1.0.206.dev1.dist-info → code_loader-1.0.206.dev3.dist-info}/RECORD

@@ -1,7 +1,7 @@
 LICENSE,sha256=qIwWjdspQeSMTtnFZBC8MuT-95L02FPvzRUdWFxrwJY,1067
 code_loader/__init__.py,sha256=outxRQ0M-zMfV0QGVJmAed5qWfRmyD0TV6-goEGAzBw,406
 code_loader/contract/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-code_loader/contract/datasetclasses.py,sha256=ySmO5Ld4fyfUWOJxUdnH2qCMota4dtwtgMHJS7Y8IvA,17070
+code_loader/contract/datasetclasses.py,sha256=6Ozy9_C7HCP6YHqM27teBLe503MRvn7jg7h8uXHPuUY,17395
 code_loader/contract/enums.py,sha256=__GkPkwAXi2agmDGtEQgbMucPm4-n80maG6n_MmObPA,1691
 code_loader/contract/exceptions.py,sha256=jWqu5i7t-0IG0jGRsKF4DjJdrsdpJjIYpUkN1F4RiyQ,51
 code_loader/contract/mapping.py,sha256=sWJhpng-IkOzQnWQdMT5w2ZZ3X1Z_OOzSwCLXIS7oxE,1446
@@ -22,17 +22,17 @@ code_loader/experiment_api/utils.py,sha256=XZHtxge12TS4H4-8PjV3sKuhp8Ud6ojAiIzTZ
 code_loader/experiment_api/workingspace_config_utils.py,sha256=DLzXQCg4dgTV_YgaSbeTVzq-2ja_SQw4zi7LXwKL9cY,990
 code_loader/inner_leap_binder/__init__.py,sha256=koOlJyMNYzGbEsoIbXathSmQ-L38N_pEXH_HvL7beXU,99
 code_loader/inner_leap_binder/leapbinder.py,sha256=mWcjx31fnrtRfV5agveW79lcMGJhwhDfLz7brpexBv0,60408
-code_loader/inner_leap_binder/leapbinder_decorators.py,sha256=lM2wFqzxdTVoe9lwC5YFdgnY8UwvY5r9foDnKJpQ7Mc,195252
-code_loader/leaploader.py,sha256=JjC5ljuofrtUP4qMh2KzRQLrIWvSTpR8ep6om92c3qc,91308
+code_loader/inner_leap_binder/leapbinder_decorators.py,sha256=GQu5bGZ7dlZbDNy9LcNkw82CMBqEdNogHtO1D1x84dc,196885
+code_loader/leaploader.py,sha256=CA2Ctpw9EN_x-wM6SmAWrBv4pijpA74ckJcqAtLJtv8,91263
 code_loader/leaploaderbase.py,sha256=Aa8wCcxojf8EBn_14W5I1gKI4OMBdpdEcNAZ_6E2eV4,10940
 code_loader/mixpanel_tracker.py,sha256=rNwRmFifNbdUoqLQvvhhgpKczWpWiEmd8MfyJe27sxw,9131
 code_loader/plot_functions/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 code_loader/plot_functions/plot_functions.py,sha256=2DC-zlVaN13P4VNx5d8csgs80C6SisaeP1-Kq2LW7iM,16075
 code_loader/plot_functions/visualize.py,sha256=gsBAYYkwMh7jIpJeDMPS8G4CW-pxwx6LznoQIvi4vpo,657
-code_loader/utils.py,sha256=mlwr-4ZeeEMMIRdMAL7JF0b-5ko7QSRhbKZ64DqKbTY,11452
+code_loader/utils.py,sha256=Auw39lKpEQ7Ld2q1nj6S18GIDKWcjCxU0TDrur2UEko,11968
 code_loader/visualizers/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 code_loader/visualizers/default_visualizers.py,sha256=grTPin_lCE9aci8i8CqA7DqQwAyXRB7_EamA3na_pls,5438
-code_loader-1.0.206.dev1.dist-info/LICENSE,sha256=qIwWjdspQeSMTtnFZBC8MuT-95L02FPvzRUdWFxrwJY,1067
-code_loader-1.0.206.dev1.dist-info/METADATA,sha256=v7DbX7F0l3-_L0R5gieoiNSAff1YAA9dqnJ9ncCD2yE,1095
-code_loader-1.0.206.dev1.dist-info/WHEEL,sha256=Nq82e9rUAnEjt98J6MlVmMCZb-t9cYE2Ir1kpBmnWfs,88
-code_loader-1.0.206.dev1.dist-info/RECORD,,
+code_loader-1.0.206.dev3.dist-info/LICENSE,sha256=qIwWjdspQeSMTtnFZBC8MuT-95L02FPvzRUdWFxrwJY,1067
+code_loader-1.0.206.dev3.dist-info/METADATA,sha256=_kkkAHCgoFUWxh8XzxcoE-pM8oDFSB3Pf32_Aqmq-Qw,1095
+code_loader-1.0.206.dev3.dist-info/WHEEL,sha256=Nq82e9rUAnEjt98J6MlVmMCZb-t9cYE2Ir1kpBmnWfs,88
+code_loader-1.0.206.dev3.dist-info/RECORD,,