deepdoctection 0.42.0__py3-none-any.whl → 0.43__py3-none-any.whl

deepdoctection/eval/tedsmetric.py

@@ -13,7 +13,9 @@
 # Apache 2.0 License for more details.
 
 """
-Tree distance similarity metric taken from <https://github.com/ibm-aur-nlp/PubTabNet/blob/master/src/metric.py>
+Tree distance similarity (TEDS) metric
+
+Taken from <https://github.com/ibm-aur-nlp/PubTabNet/blob/master/src/metric.py>
 """
 
 import statistics
@@ -81,7 +83,8 @@ class TableTree(Tree):
 
 class CustomConfig(Config):
     """
-    CustomConfig for calculating APTED tree edit distance. Check APTED docs for more information
+    `CustomConfig` for calculating `APTED` tree edit distance.
+    Check APTED docs for more information
     """
 
     @staticmethod
@@ -90,7 +93,7 @@ class CustomConfig(Config):
         return max(map(len, sequences))
 
     def normalized_distance(self, *sequences: Any) -> float:
-        """Get distance from 0 to 1"""
+        """Get distance from `0` to `1`"""
         return float(distance.levenshtein(*sequences)) / self.maximum(*sequences)
 
     def rename(self, node1: Any, node2: Any) -> float:
@@ -104,7 +107,7 @@ class CustomConfig(Config):
 
 
 class TEDS:
-    """Tree Edit Distance based Similarity"""
+    """Tree Edit Distance similarity"""
 
     def __init__(self, structure_only: bool = False):
         self.structure_only = structure_only
@@ -123,7 +126,7 @@ class TEDS:
             self.__tokens__ += list(node.tail)
 
     def load_html_tree(self, node: TableTree, parent: Optional[TableTree] = None) -> Optional[TableTree]:
-        """Converts HTML tree to the format required by apted"""
+        """Converts `HTML` tree to the format required by APTED"""
         global __tokens__  # pylint: disable = W0602
         if node.tag == "td":
             if self.structure_only:
@@ -151,8 +154,15 @@ class TEDS:
         return None
 
     def evaluate(self, inputs: tuple[str, str]) -> float:
-        """Computes TEDS score between the prediction and the ground truth of a
+        """
+        Computes TEDS score between the prediction and the ground truth of a
         given sample
+
+        Args:
+            inputs: A tuple of ground truth and prediction in xml format
+
+        Returns:
+            A float value between 0.0 and 1.0, where 1.0 means perfect match
         """
 
         ground_truth, pred = inputs[0], inputs[1]
@@ -192,8 +202,13 @@ class TEDS:
 def teds_metric(gt_list: list[str], predict_list: list[str], structure_only: bool) -> tuple[float, int]:
     """
     Computes tree edit distance score (TEDS) between the prediction and the ground truth of a batch of samples. The
-    approach to measure similarity of tables by means of their html representation has been adovacated in
-    <https://arxiv.org/abs/1911.10683> .
+    approach to measure similarity of tables by means of their html representation has been advocated in
+    <https://arxiv.org/abs/1911.10683>
+
+    Args:
+        gt_list: A list of ground truth samples in `xml` format
+        predict_list: A list of predictions in `xml` format
+        structure_only: If `True`, only the structure of the table is considered, but no text
 
     """
     teds = TEDS(structure_only=structure_only)
@@ -218,7 +233,7 @@ def teds_metric(gt_list: list[str], predict_list: list[str], structure_only: boo
 @metric_registry.register("teds")
 class TedsMetric(MetricBase):
     """
-    Metric induced by `teds`
+    Metric induced by `TEDS`
     """
 
     metric = teds_metric  # type: ignore

deepdoctection/eval/tp_eval_callback.py

@@ -16,7 +16,7 @@
 # limitations under the License.
 
 """
-Module for EvalCallback in Tensorpack
+`EvalCallback` in Tensorpack
 """
 
 from __future__ import annotations
@@ -53,7 +53,7 @@ __all__ = ["EvalCallback"]
 
 class EvalCallback(Callback):  # pylint: disable=R0903
     """
-    A callback that runs evaluation once a while. It supports evaluation on any pipeline component.
+    A callback that runs evaluation once in a while. It supports evaluation on any pipeline component.
     """
 
     _chief_only = False
@@ -71,17 +71,18 @@ class EvalCallback(Callback):  # pylint: disable=R0903
         **build_eval_kwargs: str,
     ) -> None:
         """
-        :param dataset: dataset
-        :param category_names: String or list of category names
-        :param sub_categories: Dict of categories/sub-categories or categories/list of sub-categories. See also
+        Args:
+            dataset: dataset
+            category_names: String or list of category names
+            sub_categories: Dict of categories/sub-categories or categories/list of sub-categories. See also
                                `eval.Evaluator`
-        :param metric: metric
-        :param pipeline_component: Pipeline component with a detector.
-        :param in_names: Specify tensor input names.
-                         E.g. `extern.tp.tpfrcnn.GeneralizedRCNN.get_inference_tensor_names`
-        :param out_names: Specify tensor output names.
-        :param build_eval_kwargs: Pass the necessary arguments in order to build the dataflow, e.g. "split",
-                                  "build_mode", "max_datapoints" etc.
+            metric: metric
+            pipeline_component: Pipeline component with a detector.
+            in_names: Specify tensor input names.
+                      E.g. `extern.tp.tpfrcnn.GeneralizedRCNN.get_inference_tensor_names`
+            out_names: Specify tensor output names.
+            build_eval_kwargs: Pass the necessary arguments in order to build the dataflow, e.g. `split`,
+                               `build_mode`, `max_datapoints` etc.
         """
         self.dataset_name = dataset.dataset_info.name
         self.build_eval_kwargs = build_eval_kwargs

deepdoctection/extern/__init__.py

@@ -16,7 +16,7 @@
 # limitations under the License.
 
 """
-Wrappers for models of external libraries as well as implementation of the Cascade-RCNN model of Tensorpack.
+# Wrappers for models of external libraries
 """
 
 from .base import *