indico_toolkit 7.2.2__tar.gz → 7.2.3__tar.gz

{indico_toolkit-7.2.2 → indico_toolkit-7.2.3}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.3
 Name: indico-toolkit
-Version: 7.2.2
+Version: 7.2.3
 Summary: Classes, functions, and abstractions for Indico IPA
 License: MIT
 Author: Indico Solutions Engineering
@@ -18,7 +18,7 @@ Provides-Extra: examples
 Provides-Extra: metrics
 Provides-Extra: predictions
 Provides-Extra: snapshots
-Requires-Dist: indico-client (>=6.14.0,<7.0.0)
+Requires-Dist: indico-client (>=6.14.0,<8.0.0)
 Requires-Dist: pandas (>=2.2.3,<3.0.0) ; extra == "all"
 Requires-Dist: pandas (>=2.2.3,<3.0.0) ; extra == "downloads"
 Requires-Dist: pandas (>=2.2.3,<3.0.0) ; extra == "examples"

{indico_toolkit-7.2.2 → indico_toolkit-7.2.3}/indico_toolkit/__init__.py

@@ -21,4 +21,4 @@ __all__ = (
     "ToolkitStaggeredLoopError",
     "ToolkitStatusError",
 )
-__version__ = "7.2.2"
+__version__ = "7.2.3"

{indico_toolkit-7.2.2 → indico_toolkit-7.2.3}/indico_toolkit/polling/queries.py

@@ -11,7 +11,13 @@ class SubmissionIdsPendingAutoReview(GraphQLRequest):  # type: ignore[misc, no-a
     query SubmissionIdsPendingAutoReview($workflowIds: [Int]) {
         submissions(
             desc: false
-            filters: { status: PENDING_AUTO_REVIEW }
+            filters: {
+                AND: [
+                    { status: PENDING_AUTO_REVIEW }
+                    { filesDeleted: false }
+                    { retrieved: false }
+                ]
+            }
             limit: 1000
             orderBy: ID
             workflowIds: $workflowIds
@@ -39,13 +45,16 @@ class SubmissionIdsPendingDownstream(GraphQLRequest):  # type: ignore[misc, no-a
         submissions(
             desc: false
             filters: {
-                AND: {
-                    retrieved: false
-                    OR: [
-                        { status: COMPLETE }
-                        { status: FAILED }
-                    ]
-                }
+                AND: [
+                    {
+                        OR: [
+                            { status: COMPLETE }
+                            { status: FAILED }
+                        ]
+                    }
+                    { filesDeleted: false }
+                    { retrieved: false }
+                ]
             }
             limit: 1000
             orderBy: ID

{indico_toolkit-7.2.2 → indico_toolkit-7.2.3}/indico_toolkit/results/normalization.py

@@ -54,6 +54,13 @@ def normalize_prediction_dict(task_type: TaskType, prediction: "Any") -> None:
     ) and not has(prediction, list, "spans"):
         prediction["spans"] = []
 
+    # Form Extraction bounding boxes may very rarely have a `.0` decimal place,
+    # which causes them to fail strict validation.
+    if task_type == TaskType.FORM_EXTRACTION:
+        for edge in ("top", "left", "right", "bottom"):
+            if has(prediction, float, edge):
+                prediction[edge] = int(prediction[edge])
+
     # Form Extractions added in review may lack bounding box information.
     # These values will match `NULL_BOX`.
     if task_type == TaskType.FORM_EXTRACTION and not has(prediction, int, "top"):

{indico_toolkit-7.2.2 → indico_toolkit-7.2.3}/indico_toolkit/results/predictions/__init__.py

@@ -1,3 +1,4 @@
+import sys
 from typing import TYPE_CHECKING
 
 from ..normalization import normalize_prediction_dict
@@ -55,3 +56,14 @@ def from_dict(
         return Unbundling.from_dict(document, task, review, prediction)
     else:
         raise ValueError(f"unsupported task type {task.type!r}")
+
+
+# `dataclass()` doesn't (yet) provide a way to configure the generated `__replace__`
+# method on Python 3.13+. Unshadow `Prediction.__replace__` in generated subclasses.
+if sys.version_info >= (3, 13):
+    del Classification.__replace__
+    del DocumentExtraction.__replace__
+    del Extraction.__replace__
+    del FormExtraction.__replace__
+    del Summarization.__replace__
+    del Unbundling.__replace__

{indico_toolkit-7.2.2 → indico_toolkit-7.2.3}/indico_toolkit/results/predictions/documentextraction.py

@@ -1,5 +1,5 @@
-from copy import copy, deepcopy
-from dataclasses import dataclass, field, replace
+from copy import copy
+from dataclasses import dataclass, field
 from typing import TYPE_CHECKING, Any
 
 from ...etloutput import (
@@ -131,6 +131,19 @@ class DocumentExtraction(Extraction):
                 self.tables.append(table)
                 self.cells.append(cell)
 
+    def __deepcopy__(self, memo: Any) -> "Self":
+        """
+        Supports `copy.deepcopy(prediction)` without copying immutable objects.
+        This provides a significant time and memory improvement when OCR is assigned.
+        """
+        new_instance = super().__deepcopy__(memo)
+        new_instance.groups = copy(self.groups)
+        new_instance.spans = copy(self.spans)
+        new_instance.tokens = copy(self.tokens)
+        new_instance.tables = copy(self.tables)
+        new_instance.cells = copy(self.cells)
+        return new_instance
+
     @staticmethod
     def from_dict(
         document: "Document",
@@ -190,15 +203,3 @@ class DocumentExtraction(Extraction):
             prediction["rejected"] = True
 
         return prediction
-
-    def copy(self) -> "Self":
-        return replace(
-            self,
-            groups=copy(self.groups),
-            spans=copy(self.spans),
-            tokens=copy(self.tokens),
-            tables=copy(self.tables),
-            cells=copy(self.cells),
-            confidences=copy(self.confidences),
-            extras=deepcopy(self.extras),
-        )

indico_toolkit-7.2.3/indico_toolkit/results/predictions/prediction.py

@@ -0,0 +1,71 @@
+from copy import copy, deepcopy
+from dataclasses import dataclass
+from typing import TYPE_CHECKING, Any
+
+if TYPE_CHECKING:
+    from typing_extensions import Self
+
+    from ..document import Document
+    from ..review import Review
+    from ..task import Task
+
+
+@dataclass
+class Prediction:
+    document: "Document"
+    task: "Task"
+    review: "Review | None"
+
+    label: str
+    confidences: "dict[str, float]"
+    extras: "dict[str, Any]"
+
+    @property
+    def confidence(self) -> float:
+        return self.confidences[self.label]
+
+    @confidence.setter
+    def confidence(self, value: float) -> None:
+        self.confidences[self.label] = value
+
+    def __deepcopy__(self, memo: Any) -> "Self":
+        """
+        Supports `copy.deepcopy(prediction)` without copying immutable objects.
+        """
+        new_instance = copy(self)
+        new_instance.confidences = copy(self.confidences)
+        new_instance.extras = deepcopy(self.extras, memo)
+        return new_instance
+
+    def __replace__override__(self, **attributes: Any) -> "Self":
+        """
+        Supports `copy.replace(prediction, **attrs)` on Python 3.13+
+
+        Unlike `dataclasses.replace(**attrs)` this performs a deep copy and allows
+        assigning properties in addition to attributes.
+
+        E.g.
+        >>> dataclasses.replace(prediction, confidence=1.0)
+        Shallow copy and raises TypeError(...)
+        >>> copy.replace(prediction, confidence=1.0)
+        Deep copy and returns Prediction(confidence=1.0, ...)
+        """
+        new_instance = deepcopy(self)
+
+        for attribute, value in attributes.items():
+            setattr(new_instance, attribute, value)
+
+        return new_instance
+
+    def to_dict(self) -> "dict[str, Any]":
+        """
+        Create a prediction dictionary for auto review changes.
+        """
+        raise NotImplementedError()
+
+
+# `dataclass()` doesn't (yet) provide a way to override the generated `__replace__`
+# method on Python 3.13+. It must be overridden after class generation and unshadowed
+# on all derived classes.
+Prediction.__replace__ = Prediction.__replace__override__  # type:ignore
+del Prediction.__replace__override__

{indico_toolkit-7.2.2 → indico_toolkit-7.2.3}/indico_toolkit/results/predictions/summarization.py

@@ -1,4 +1,4 @@
-from copy import copy, deepcopy
+from copy import copy
 from dataclasses import dataclass, replace
 from typing import TYPE_CHECKING, Any
 
@@ -70,6 +70,14 @@ class Summarization(Extraction):
         """
         self.citation = replace(self.citation, span=span)
 
+    def __deepcopy__(self, memo: Any) -> "Self":
+        """
+        Supports `copy.deepcopy(prediction)` without copying immutable objects.
+        """
+        new_instance = super().__deepcopy__(memo)
+        new_instance.citations = copy(self.citations)
+        return new_instance
+
     @staticmethod
     def from_dict(
         document: "Document",
@@ -125,11 +133,3 @@ class Summarization(Extraction):
             prediction["rejected"] = True
 
         return prediction
-
-    def copy(self) -> "Self":
-        return replace(
-            self,
-            citations=copy(self.citations),
-            confidences=copy(self.confidences),
-            extras=deepcopy(self.extras),
-        )

{indico_toolkit-7.2.2 → indico_toolkit-7.2.3}/indico_toolkit/results/predictions/unbundling.py

@@ -1,5 +1,5 @@
-from copy import copy, deepcopy
-from dataclasses import dataclass, replace
+from copy import copy
+from dataclasses import dataclass
 from typing import TYPE_CHECKING, Any
 
 from ...etloutput import Span
@@ -25,6 +25,14 @@ class Unbundling(Prediction):
         """
         return tuple(span.page for span in self.spans)
 
+    def __deepcopy__(self, memo: Any) -> "Self":
+        """
+        Supports `copy.deepcopy(prediction)` without copying immutable objects.
+        """
+        new_instance = super().__deepcopy__(memo)
+        new_instance.spans = copy(self.spans)
+        return new_instance
+
     @staticmethod
     def from_dict(
         document: "Document",
@@ -55,11 +63,3 @@ class Unbundling(Prediction):
             "confidence": self.confidences,
             "spans": [span.to_dict() for span in self.spans],
         }
-
-    def copy(self) -> "Self":
-        return replace(
-            self,
-            spans=copy(self.spans),
-            confidences=copy(self.confidences),
-            extras=deepcopy(self.extras),
-        )

{indico_toolkit-7.2.2 → indico_toolkit-7.2.3}/indico_toolkit/results/result.py

@@ -1,6 +1,8 @@
-from dataclasses import dataclass
+from copy import deepcopy
+from dataclasses import dataclass, replace
 from functools import partial
 from itertools import chain
+from typing import TYPE_CHECKING, Any
 
 from . import predictions as prediction
 from .document import Document
@@ -11,6 +13,9 @@ from .review import Review, ReviewType
 from .task import Task
 from .utils import get
 
+if TYPE_CHECKING:
+    from typing_extensions import Self
+
 
 @dataclass(frozen=True, order=True)
 class Result:
@@ -44,6 +49,12 @@ class Result:
     def final(self) -> "PredictionList[Prediction]":
         return self.predictions.where(review=self.reviews[-1] if self.reviews else None)
 
+    def __deepcopy__(self, memo: Any) -> "Self":
+        """
+        Supports `copy.deepcopy(result)` without copying immutable objects.
+        """
+        return replace(self, predictions=deepcopy(self.predictions, memo))
+
     @staticmethod
     def from_dict(result: object) -> "Result":
         """

{indico_toolkit-7.2.2 → indico_toolkit-7.2.3}/indico_toolkit/retry.py

@@ -10,8 +10,7 @@ if TYPE_CHECKING:
     from typing import ParamSpec, TypeVar
 
     ArgumentsType = ParamSpec("ArgumentsType")
-    OuterReturnType = TypeVar("OuterReturnType")
-    InnerReturnType = TypeVar("InnerReturnType")
+    ReturnType = TypeVar("ReturnType")
 
 
 def retry(
@@ -20,7 +19,7 @@ def retry(
     wait: float = 1,
     backoff: float = 4,
     jitter: float = 0.5,
-) -> "Callable[[Callable[ArgumentsType, OuterReturnType]], Callable[ArgumentsType, OuterReturnType]]":  # noqa: E501
+) -> "Callable[[Callable[ArgumentsType, ReturnType]], Callable[ArgumentsType, ReturnType]]":  # noqa: E501
     """
     Decorate a function or coroutine to retry when it raises specified errors,
     apply exponential backoff and jitter to the wait time,
@@ -46,15 +45,15 @@ def retry(
 
     @overload
     def retry_decorator(
-        decorated: "Callable[ArgumentsType, Awaitable[InnerReturnType]]",
-    ) -> "Callable[ArgumentsType, Awaitable[InnerReturnType]]": ...
+        decorated: "Callable[ArgumentsType, Awaitable[ReturnType]]",
+    ) -> "Callable[ArgumentsType, Awaitable[ReturnType]]": ...
     @overload
     def retry_decorator(
-        decorated: "Callable[ArgumentsType, InnerReturnType]",
-    ) -> "Callable[ArgumentsType, InnerReturnType]": ...
+        decorated: "Callable[ArgumentsType, ReturnType]",
+    ) -> "Callable[ArgumentsType, ReturnType]": ...
     def retry_decorator(
-        decorated: "Callable[ArgumentsType, InnerReturnType]",
-    ) -> "Callable[ArgumentsType, Awaitable[InnerReturnType]] | Callable[ArgumentsType, InnerReturnType]":  # noqa: E501
+        decorated: "Callable[ArgumentsType, ReturnType]",
+    ) -> "Callable[ArgumentsType, Awaitable[ReturnType]] | Callable[ArgumentsType, ReturnType]":  # noqa: E501
         """
         Decorate either a function or coroutine as appropriate.
         """
@@ -63,7 +62,7 @@ def retry(
             @wraps(decorated)
             async def retrying_coroutine(  # type: ignore[return]
                 *args: "ArgumentsType.args", **kwargs: "ArgumentsType.kwargs"
-            ) -> "InnerReturnType":
+            ) -> "ReturnType":
                 for times_retried in range(count + 1):
                     try:
                         return await decorated(*args, **kwargs)  # type: ignore[no-any-return]
@@ -79,7 +78,7 @@ def retry(
             @wraps(decorated)
             def retrying_function(  # type: ignore[return]
                 *args: "ArgumentsType.args", **kwargs: "ArgumentsType.kwargs"
-            ) -> "InnerReturnType":
+            ) -> "ReturnType":
                 for times_retried in range(count + 1):
                     try:
                         return decorated(*args, **kwargs)

{indico_toolkit-7.2.2 → indico_toolkit-7.2.3}/pyproject.toml

@@ -12,8 +12,8 @@ authors = [
 readme = "README.md"
 urls = { source = "https://github.com/IndicoDataSolutions/Indico-Solutions-Toolkit" }
 requires-python = ">=3.10"
-version = "7.2.2"
-dependencies = ["indico-client (>=6.14.0,<7.0.0)"]
+version = "7.2.3"
+dependencies = ["indico-client (>=6.14.0,<8.0.0)"]
 
 [project.optional-dependencies]
 all = ["pandas (>=2.2.3,<3.0.0)", "plotly (>=5.2.1,<6.0.0)", "tqdm (>=4.50.0,<5.0.0)"]

indico_toolkit-7.2.2/indico_toolkit/results/predictions/prediction.py

@@ -1,42 +0,0 @@
-from copy import copy, deepcopy
-from dataclasses import dataclass, replace
-from typing import TYPE_CHECKING, Any
-
-if TYPE_CHECKING:
-    from typing_extensions import Self
-
-    from ..document import Document
-    from ..review import Review
-    from ..task import Task
-
-
-@dataclass
-class Prediction:
-    document: "Document"
-    task: "Task"
-    review: "Review | None"
-
-    label: str
-    confidences: "dict[str, float]"
-    extras: "dict[str, Any]"
-
-    @property
-    def confidence(self) -> float:
-        return self.confidences[self.label]
-
-    @confidence.setter
-    def confidence(self, value: float) -> None:
-        self.confidences[self.label] = value
-
-    def to_dict(self) -> "dict[str, Any]":
-        """
-        Create a prediction dictionary for auto review changes.
-        """
-        raise NotImplementedError()
-
-    def copy(self) -> "Self":
-        return replace(
-            self,
-            confidences=copy(self.confidences),
-            extras=deepcopy(self.extras),
-        )