https://github.com/ArneBinder/pie-datasets/pull/100

README.md

@@ -3,11 +3,32 @@
 This is a [PyTorch-IE](https://github.com/ChristophAlt/pytorch-ie) wrapper for the
 [SciDTB ArgMin Huggingface dataset loading script](https://huggingface.co/datasets/DFKI-SLT/scidtb_argmin).
 
+## Usage
+
+```python
+from pie_datasets import load_dataset
+from pytorch_ie.documents import TextDocumentWithLabeledSpansAndBinaryRelations
+
+# load English variant
+dataset = load_dataset("pie/scidtb_argmin")
+
+# if required, normalize the document type (see section Document Converters below)
+dataset_converted = dataset.to_document_type(TextDocumentWithLabeledSpansAndBinaryRelations)
+assert isinstance(dataset_converted["train"][0], TextDocumentWithLabeledSpansAndBinaryRelations)
+
+# get first relation in the first document
+doc = dataset_converted["train"][0]
+print(doc.binary_relations[0])
+# BinaryRelation(head=LabeledSpan(start=251, end=454, label='means', score=1.0), tail=LabeledSpan(start=455, end=712, label='proposal', score=1.0), label='detail', score=1.0)
+print(doc.binary_relations[0].resolve())
+# ('detail', (('means', 'We observe , identify , and detect naturally occurring signals of interestingness in click transitions on the Web between source and target documents , which we collect from commercial Web browser logs .'), ('proposal', 'The DSSM is trained on millions of Web transitions , and maps source-target document pairs to feature vectors in a latent space in such a way that the distance between source documents and their corresponding interesting targets in that space is minimized .')))
+```
+
 ## Data Schema
 
 The document type for this dataset is `SciDTBArgminDocument` which defines the following data fields:
 
-- `tokens` (Tuple of string)
+- `tokens` (tuple of string)
 - `id` (str, optional)
 - `metadata` (dictionary, optional)
 
@@ -23,6 +44,128 @@ See [here](https://github.com/ChristophAlt/pytorch-ie/blob/main/src/pytorch_ie/a
 The dataset provides document converters for the following target document types:
 
 - `pytorch_ie.documents.TextDocumentWithLabeledSpansAndBinaryRelations`
+  - `labeled_spans`: `LabeledSpan` annotations, converted from`SciDTBArgminDocument`'s `units`
+    - labels: `proposal`, `assertion`, `result`, `observation`, `means`, `description`
+    - tuples of `tokens` are joined with a whitespace to create `text` for `LabeledSpans`
+  - `binary_relations`: `BinaryRelation` annotations, converted from `SciDTBArgminDocument`'s `relations`
+    - labels: `support`, `attack`, `additional`, `detail`, `sequence`
 
 See [here](https://github.com/ChristophAlt/pytorch-ie/blob/main/src/pytorch_ie/documents.py) for the document type
 definitions.
+
+### Collected Statistics after Document Conversion
+
+We use the script `evaluate_documents.py` from [PyTorch-IE-Hydra-Template](https://github.com/ArneBinder/pytorch-ie-hydra-template-1) to generate these statistics.
+After checking out that code, the statistics and plots can be generated by the command:
+
+```commandline
+python src/evaluate_documents.py dataset=scidtb_argmin_base metric=METRIC
+```
+
+where a `METRIC` is called according to the available metric configs in `config/metric/METRIC` (see [metrics](https://github.com/ArneBinder/pytorch-ie-hydra-template-1/tree/main/configs/metric)).
+
+This also requires to have the following dataset config in `configs/dataset/scidtb_argmin_base.yaml` of this dataset within the repo directory:
+
+```commandline
+_target_: src.utils.execute_pipeline
+input:
+  _target_: pie_datasets.DatasetDict.load_dataset
+  path: pie/scidtb_argmin
+  revision: 335a8e6168919d7f204c6920eceb96745dbd161b
+```
+
+For token based metrics, this uses `bert-base-uncased` from `transformer.AutoTokenizer` (see [AutoTokenizer](https://huggingface.co/docs/transformers/v4.37.1/en/model_doc/auto#transformers.AutoTokenizer), and [bert-based-uncased](https://huggingface.co/bert-base-uncased) to tokenize `text` in `TextDocumentWithLabeledSpansAndBinaryRelations` (see [document type](https://github.com/ChristophAlt/pytorch-ie/blob/main/src/pytorch_ie/documents.py)).
+
+#### Relation argument (outer) token distance per label
+
+The distance is measured from the first token of the first argumentative unit to the last token of the last unit, a.k.a. outer distance.
+
+We collect the following statistics: number of documents in the split (*no. doc*), no. of relations (*len*), mean of token distance (*mean*), standard deviation of the distance (*std*), minimum outer distance (*min*), and maximum outer distance (*max*).
+We also present histograms in the collapsible, showing the distribution of these relation distances (x-axis; and unit-counts in y-axis), accordingly.
+
+<details>
+<summary>Command</summary>
+
+```
+python src/evaluate_documents.py dataset=scidtb_argmin_base metric=relation_argument_token_distances
+```
+
+</details>
+
+|            | len | max |   mean | min |    std |
+| :--------- | --: | --: | -----: | --: | -----: |
+| ALL        | 586 | 277 | 75.239 |  21 | 40.312 |
+| additional |  54 | 180 | 59.593 |  36 | 29.306 |
+| detail     | 258 | 163 |  65.62 |  22 |  29.21 |
+| sequence   |  22 |  93 | 59.727 |  38 | 17.205 |
+| support    | 252 | 277 | 89.794 |  21 | 48.118 |
+
+<details>
+  <summary>Histogram (split: train, 60 documents)</summary>
+
+![rtd-label_scitdb-argmin.png](img%2Frtd-label_scitdb-argmin.png)
+
+</details>
+
+#### Span lengths (tokens)
+
+The span length is measured from the first token of the first argumentative unit to the last token of the particular unit.
+
+We collect the following statistics: number of documents in the split (*no. doc*), no. of spans (*len*), mean of number of tokens in a span (*mean*), standard deviation of the number of tokens (*std*), minimum tokens in a span (*min*), and maximum tokens in a span (*max*).
+We also present histograms in the collapsible, showing the distribution of these token-numbers (x-axis; and unit-counts in y-axis), accordingly.
+
+<details>
+<summary>Command</summary>
+
+```
+python src/evaluate_documents.py dataset=scidtb_argmin_base metric=span_lengths_tokens
+```
+
+</details>
+
+| statistics |  train |
+| :--------- | -----: |
+| no. doc    |     60 |
+| len        |    353 |
+| mean       | 27.946 |
+| std        | 13.054 |
+| min        |      7 |
+| max        |    123 |
+
+<details>
+  <summary>Histogram (split: train, 60 documents)</summary>
+
+![slt_scitdb-argmin.png](img%2Fslt_scitdb-argmin.png)
+
+</details>
+
+#### Token length (tokens)
+
+The token length is measured from the first token of the document to the last one.
+
+We collect the following statistics: number of documents in the split (*no. doc*), mean of document token-length (*mean*), standard deviation of the length (*std*), minimum number of tokens in a document (*min*), and maximum number of tokens in a document (*max*).
+We also present histograms in the collapsible, showing the distribution of these token lengths (x-axis; and unit-counts in y-axis), accordingly.
+
+<details>
+<summary>Command</summary>
+
+```
+python src/evaluate_documents.py dataset=scidtb_argmin_base metric=count_text_tokens
+```
+
+</details>
+
+| statistics |   train |
+| :--------- | ------: |
+| no. doc    |      60 |
+| mean       | 164.417 |
+| std        |  64.572 |
+| min        |      80 |
+| max        |     532 |
+
+<details>
+  <summary>Histogram (split: train, 60 documents)</summary>
+
+![tl_scidtb-argmin.png](img%2Ftl_scidtb-argmin.png)
+
+</details>

requirements.txt

@@ -1,2 +1,2 @@
-pie-datasets>=0.6.0,<0.9.0
-pie-modules>=0.8.0,<0.9.0
+pie-datasets>=0.6.0,<0.11.0
+pie-modules>=0.8.0,<0.12.0

scidtb_argmin.py

@@ -1,173 +1,173 @@
-import dataclasses
-import logging
-from typing import Any, Dict, List, Tuple
-
-import datasets
-from pie_modules.document.processing import token_based_document_to_text_based
-from pytorch_ie.annotations import BinaryRelation, LabeledSpan
-from pytorch_ie.core import AnnotationList, annotation_field
-from pytorch_ie.documents import (
-    TextDocumentWithLabeledSpansAndBinaryRelations,
-    TokenBasedDocument,
-)
-from pytorch_ie.utils.span import bio_tags_to_spans
-
-from pie_datasets import GeneratorBasedBuilder
-
-log = logging.getLogger(__name__)
-
-
-def labels_and_spans_to_bio_tags(
-    labels: List[str], spans: List[Tuple[int, int]], sequence_length: int
-) -> List[str]:
-    bio_tags = ["O"] * sequence_length
-    for label, (start, end) in zip(labels, spans):
-        bio_tags[start] = f"B-{label}"
-        for i in range(start + 1, end):
-            bio_tags[i] = f"I-{label}"
-    return bio_tags
-
-
-@dataclasses.dataclass
-class SciDTBArgminDocument(TokenBasedDocument):
-    units: AnnotationList[LabeledSpan] = annotation_field(target="tokens")
-    relations: AnnotationList[BinaryRelation] = annotation_field(target="units")
-
-
-@dataclasses.dataclass
-class SimplifiedSciDTBArgminDocument(TokenBasedDocument):
-    labeled_spans: AnnotationList[LabeledSpan] = annotation_field(target="tokens")
-    binary_relations: AnnotationList[BinaryRelation] = annotation_field(target="labeled_spans")
-
-
-def example_to_document(
-    example: Dict[str, Any],
-    unit_bio: datasets.ClassLabel,
-    unit_label: datasets.ClassLabel,
-    relation: datasets.ClassLabel,
-):
-    document = SciDTBArgminDocument(id=example["id"], tokens=tuple(example["data"]["token"]))
-    bio_tags = unit_bio.int2str(example["data"]["unit-bio"])
-    unit_labels = unit_label.int2str(example["data"]["unit-label"])
-    roles = relation.int2str(example["data"]["role"])
-    tag_sequence = [
-        f"{bio}-{label}|{role}|{parent_offset}"
-        for bio, label, role, parent_offset in zip(
-            bio_tags, unit_labels, roles, example["data"]["parent-offset"]
-        )
-    ]
-    spans_with_label = sorted(
-        bio_tags_to_spans(tag_sequence), key=lambda label_and_span: label_and_span[1][0]
-    )
-    labels, spans = zip(*spans_with_label)
-    span_unit_labels, span_roles, span_parent_offsets = zip(
-        *[label.split("|") for label in labels]
-    )
-
-    units = [
-        LabeledSpan(start=start, end=end + 1, label=label)
-        for (start, end), label in zip(spans, span_unit_labels)
-    ]
-    document.units.extend(units)
-
-    # The relation direction is as in "f{head} {relation_label} {tail}"
-    relations = []
-    for idx, parent_offset in enumerate(span_parent_offsets):
-        if span_roles[idx] != "none":
-            relations.append(
-                BinaryRelation(
-                    head=units[idx], tail=units[idx + int(parent_offset)], label=span_roles[idx]
-                )
-            )
-
-    document.relations.extend(relations)
-
-    return document
-
-
-def document_to_example(
-    document: SciDTBArgminDocument,
-    unit_bio: datasets.ClassLabel,
-    unit_label: datasets.ClassLabel,
-    relation: datasets.ClassLabel,
-) -> Dict[str, Any]:
-    unit2idx = {unit: idx for idx, unit in enumerate(document.units)}
-    unit2parent_relation = {relation.head: relation for relation in document.relations}
-
-    unit_labels = [unit.label for unit in document.units]
-    roles = [
-        unit2parent_relation[unit].label if unit in unit2parent_relation else "none"
-        for unit in document.units
-    ]
-    parent_offsets = [
-        unit2idx[unit2parent_relation[unit].tail] - idx if unit in unit2parent_relation else 0
-        for idx, unit in enumerate(document.units)
-    ]
-    labels = [
-        f"{unit_label}-{role}-{parent_offset}"
-        for unit_label, role, parent_offset in zip(unit_labels, roles, parent_offsets)
-    ]
-
-    tag_sequence = labels_and_spans_to_bio_tags(
-        labels=labels,
-        spans=[(unit.start, unit.end) for unit in document.units],
-        sequence_length=len(document.tokens),
-    )
-    bio_tags, unit_labels, roles, parent_offsets = zip(
-        *[tag.split("-", maxsplit=3) for tag in tag_sequence]
-    )
-
-    data = {
-        "token": list(document.tokens),
-        "unit-bio": unit_bio.str2int(bio_tags),
-        "unit-label": unit_label.str2int(unit_labels),
-        "role": relation.str2int(roles),
-        "parent-offset": [int(idx_str) for idx_str in parent_offsets],
-    }
-    result = {"id": document.id, "data": data}
-    return result
-
-
-def convert_to_text_document_with_labeled_spans_and_binary_relations(
-    document: SciDTBArgminDocument,
-) -> TextDocumentWithLabeledSpansAndBinaryRelations:
-    doc_simplified = document.as_type(
-        SimplifiedSciDTBArgminDocument,
-        field_mapping={"units": "labeled_spans", "relations": "binary_relations"},
-    )
-    result = token_based_document_to_text_based(
-        doc_simplified,
-        result_document_type=TextDocumentWithLabeledSpansAndBinaryRelations,
-        join_tokens_with=" ",
-    )
-    return result
-
-
-class SciDTBArgmin(GeneratorBasedBuilder):
-    DOCUMENT_TYPE = SciDTBArgminDocument
-
-    DOCUMENT_CONVERTERS = {
-        TextDocumentWithLabeledSpansAndBinaryRelations: convert_to_text_document_with_labeled_spans_and_binary_relations
-    }
-
-    BASE_DATASET_PATH = "DFKI-SLT/scidtb_argmin"
-    BASE_DATASET_REVISION = "8c02587edcb47ab5b102692bd10bfffd1844a09b"
-
-    BUILDER_CONFIGS = [datasets.BuilderConfig(name="default")]
-
-    DEFAULT_CONFIG_NAME = "default"
-
-    def _generate_document_kwargs(self, dataset):
-        return {
-            "unit_bio": dataset.features["data"].feature["unit-bio"],
-            "unit_label": dataset.features["data"].feature["unit-label"],
-            "relation": dataset.features["data"].feature["role"],
-        }
-
-    def _generate_document(self, example, unit_bio, unit_label, relation):
-        return example_to_document(
-            example,
-            unit_bio=unit_bio,
-            unit_label=unit_label,
-            relation=relation,
-        )
+import dataclasses
+import logging
+from typing import Any, Dict, List, Tuple
+
+import datasets
+from pie_modules.document.processing import token_based_document_to_text_based
+from pytorch_ie.annotations import BinaryRelation, LabeledSpan
+from pytorch_ie.core import AnnotationList, annotation_field
+from pytorch_ie.documents import (
+    TextDocumentWithLabeledSpansAndBinaryRelations,
+    TokenBasedDocument,
+)
+from pytorch_ie.utils.span import bio_tags_to_spans
+
+from pie_datasets import GeneratorBasedBuilder
+
+log = logging.getLogger(__name__)
+
+
+def labels_and_spans_to_bio_tags(
+    labels: List[str], spans: List[Tuple[int, int]], sequence_length: int
+) -> List[str]:
+    bio_tags = ["O"] * sequence_length
+    for label, (start, end) in zip(labels, spans):
+        bio_tags[start] = f"B-{label}"
+        for i in range(start + 1, end):
+            bio_tags[i] = f"I-{label}"
+    return bio_tags
+
+
+@dataclasses.dataclass
+class SciDTBArgminDocument(TokenBasedDocument):
+    units: AnnotationList[LabeledSpan] = annotation_field(target="tokens")
+    relations: AnnotationList[BinaryRelation] = annotation_field(target="units")
+
+
+@dataclasses.dataclass
+class SimplifiedSciDTBArgminDocument(TokenBasedDocument):
+    labeled_spans: AnnotationList[LabeledSpan] = annotation_field(target="tokens")
+    binary_relations: AnnotationList[BinaryRelation] = annotation_field(target="labeled_spans")
+
+
+def example_to_document(
+    example: Dict[str, Any],
+    unit_bio: datasets.ClassLabel,
+    unit_label: datasets.ClassLabel,
+    relation: datasets.ClassLabel,
+):
+    document = SciDTBArgminDocument(id=example["id"], tokens=tuple(example["data"]["token"]))
+    bio_tags = unit_bio.int2str(example["data"]["unit-bio"])
+    unit_labels = unit_label.int2str(example["data"]["unit-label"])
+    roles = relation.int2str(example["data"]["role"])
+    tag_sequence = [
+        f"{bio}-{label}|{role}|{parent_offset}"
+        for bio, label, role, parent_offset in zip(
+            bio_tags, unit_labels, roles, example["data"]["parent-offset"]
+        )
+    ]
+    spans_with_label = sorted(
+        bio_tags_to_spans(tag_sequence), key=lambda label_and_span: label_and_span[1][0]
+    )
+    labels, spans = zip(*spans_with_label)
+    span_unit_labels, span_roles, span_parent_offsets = zip(
+        *[label.split("|") for label in labels]
+    )
+
+    units = [
+        LabeledSpan(start=start, end=end + 1, label=label)
+        for (start, end), label in zip(spans, span_unit_labels)
+    ]
+    document.units.extend(units)
+
+    # The relation direction is as in "f{head} {relation_label} {tail}"
+    relations = []
+    for idx, parent_offset in enumerate(span_parent_offsets):
+        if span_roles[idx] != "none":
+            relations.append(
+                BinaryRelation(
+                    head=units[idx], tail=units[idx + int(parent_offset)], label=span_roles[idx]
+                )
+            )
+
+    document.relations.extend(relations)
+
+    return document
+
+
+def document_to_example(
+    document: SciDTBArgminDocument,
+    unit_bio: datasets.ClassLabel,
+    unit_label: datasets.ClassLabel,
+    relation: datasets.ClassLabel,
+) -> Dict[str, Any]:
+    unit2idx = {unit: idx for idx, unit in enumerate(document.units)}
+    unit2parent_relation = {relation.head: relation for relation in document.relations}
+
+    unit_labels = [unit.label for unit in document.units]
+    roles = [
+        unit2parent_relation[unit].label if unit in unit2parent_relation else "none"
+        for unit in document.units
+    ]
+    parent_offsets = [
+        unit2idx[unit2parent_relation[unit].tail] - idx if unit in unit2parent_relation else 0
+        for idx, unit in enumerate(document.units)
+    ]
+    labels = [
+        f"{unit_label}-{role}-{parent_offset}"
+        for unit_label, role, parent_offset in zip(unit_labels, roles, parent_offsets)
+    ]
+
+    tag_sequence = labels_and_spans_to_bio_tags(
+        labels=labels,
+        spans=[(unit.start, unit.end) for unit in document.units],
+        sequence_length=len(document.tokens),
+    )
+    bio_tags, unit_labels, roles, parent_offsets = zip(
+        *[tag.split("-", maxsplit=3) for tag in tag_sequence]
+    )
+
+    data = {
+        "token": list(document.tokens),
+        "unit-bio": unit_bio.str2int(bio_tags),
+        "unit-label": unit_label.str2int(unit_labels),
+        "role": relation.str2int(roles),
+        "parent-offset": [int(idx_str) for idx_str in parent_offsets],
+    }
+    result = {"id": document.id, "data": data}
+    return result
+
+
+def convert_to_text_document_with_labeled_spans_and_binary_relations(
+    document: SciDTBArgminDocument,
+) -> TextDocumentWithLabeledSpansAndBinaryRelations:
+    doc_simplified = document.as_type(
+        SimplifiedSciDTBArgminDocument,
+        field_mapping={"units": "labeled_spans", "relations": "binary_relations"},
+    )
+    result = token_based_document_to_text_based(
+        doc_simplified,
+        result_document_type=TextDocumentWithLabeledSpansAndBinaryRelations,
+        join_tokens_with=" ",
+    )
+    return result
+
+
+class SciDTBArgmin(GeneratorBasedBuilder):
+    DOCUMENT_TYPE = SciDTBArgminDocument
+
+    DOCUMENT_CONVERTERS = {
+        TextDocumentWithLabeledSpansAndBinaryRelations: convert_to_text_document_with_labeled_spans_and_binary_relations
+    }
+
+    BASE_DATASET_PATH = "DFKI-SLT/scidtb_argmin"
+    BASE_DATASET_REVISION = "8c02587edcb47ab5b102692bd10bfffd1844a09b"
+
+    BUILDER_CONFIGS = [datasets.BuilderConfig(name="default")]
+
+    DEFAULT_CONFIG_NAME = "default"
+
+    def _generate_document_kwargs(self, dataset):
+        return {
+            "unit_bio": dataset.features["data"].feature["unit-bio"],
+            "unit_label": dataset.features["data"].feature["unit-label"],
+            "relation": dataset.features["data"].feature["role"],
+        }
+
+    def _generate_document(self, example, unit_bio, unit_label, relation):
+        return example_to_document(
+            example,
+            unit_bio=unit_bio,
+            unit_label=unit_label,
+            relation=relation,
+        )