Datasets:
Upload README.md with huggingface_hub
Browse files
README.md
ADDED
|
@@ -0,0 +1,223 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: cc-by-nc-4.0
|
| 3 |
+
language:
|
| 4 |
+
- fr
|
| 5 |
+
tags:
|
| 6 |
+
- inception
|
| 7 |
+
- uima
|
| 8 |
+
- annotation
|
| 9 |
+
task_categories:
|
| 10 |
+
- token-classification
|
| 11 |
+
configs:
|
| 12 |
+
- config_name: document_metadata
|
| 13 |
+
data_files:
|
| 14 |
+
- split: train
|
| 15 |
+
path: data/document_metadata/document_metadata_train.parquet
|
| 16 |
+
- split: dev
|
| 17 |
+
path: data/document_metadata/document_metadata_dev.parquet
|
| 18 |
+
- config_name: spans
|
| 19 |
+
data_files:
|
| 20 |
+
- split: train
|
| 21 |
+
path: data/spans/spans_train.parquet
|
| 22 |
+
- split: dev
|
| 23 |
+
path: data/spans/spans_dev.parquet
|
| 24 |
+
---
|
| 25 |
+
# Dataset Card for PARHAF-pseudo-annotated
|
| 26 |
+
|
| 27 |
+
<div align="center">
|
| 28 |
+
<p>
|
| 29 |
+
<a href="https://huggingface.co/spaces/HealthDataHub/PARTAGES" style="display:inline;"><img src="img/PARTAGES BASELINE_RVB.png" alt="Logo" style="height:20px;vertical-align:middle; margin-right:8px; display:inline; margin-bottom:0.2em; margin-top:0.2em;" title="PARTAGES project"; /></a>
|
| 30 |
+
<a href="https://www.etalab.gouv.fr/wp-content/uploads/2018/11/open-licence.pdf" style="display:inline;" title="Etalab 2.0 license"><img src="img/Logo-licence-ouverte2.svg" style="height:20px; display:inline; margin-bottom:0.2em; margin-top:0.2em;"/></a>
|
| 31 |
+
<a href="https://creativecommons.org/licenses/by/4.0/deed.en" style="display:inline;"><img src="https://mirrors.creativecommons.org/presskit/buttons/88x31/png/by.png" style="height:20px; display:inline; margin-bottom:0.2em; margin-top:0.2em;" title="CC BY 4.0 license"/></a>
|
| 32 |
+
</p>
|
| 33 |
+
</div>
|
| 34 |
+
|
| 35 |
+
|
| 36 |
+
## Dataset Description
|
| 37 |
+
|
| 38 |
+
- **Points of Contact:** WAJSBURT Perceval, KHALIL Youness
|
| 39 |
+
|
| 40 |
+
### Dataset Summary
|
| 41 |
+
|
| 42 |
+
PARHAF-pseudo-annotated is a subpart of the PARHAF corpus, an open French corpus of human-authored clinical reports of fictional patients.
|
| 43 |
+
|
| 44 |
+
It was created to support the development and evaluation of clinical NLP systems for the pseudonymization of identifying entities in clinical reports..
|
| 45 |
+
|
| 46 |
+
This dataset contains **training data only**. The test set will remain under embargo to enable future evaluations under controlled conditions, limiting the risk of LLM contamination through prior data exposure. **Please contact us for access to the test data.**
|
| 47 |
+
|
| 48 |
+
This training dataset is divided into a train split (80%) and a dev split (20%) to facilitate experimental design and reproducibility across teams. Teams are free to use the full training set or define a different split configuration.
|
| 49 |
+
|
| 50 |
+
|
| 51 |
+
Each patient record was:
|
| 52 |
+
- written by a senior medical resident
|
| 53 |
+
- reviewed by another senior medical resident, from the same specialty
|
| 54 |
+
- annotated by a specialist of the use case
|
| 55 |
+
- curated by another specialist of the use case
|
| 56 |
+
|
| 57 |
+
|
| 58 |
+
#### Data statistics
|
| 59 |
+
|
| 60 |
+
### DATASET SUMMARY
|
| 61 |
+
|
| 62 |
+
| Indicator | Value |
|
| 63 |
+
| --- | --- |
|
| 64 |
+
| Complete Dataset | |
|
| 65 |
+
| Number of JSON files | 509 |
|
| 66 |
+
| Total annotations | 6976 |
|
| 67 |
+
| Average document length | 4021 characters |
|
| 68 |
+
| 80% Threshold (annotations) | 5580 |
|
| 69 |
+
|
| 70 |
+
### Complete dataset distribution by type
|
| 71 |
+
|
| 72 |
+
| Type | Annotations | % of total |
|
| 73 |
+
| --- | --- | --- |
|
| 74 |
+
| EntiteAnonymisation | 6976 | 100.00% |
|
| 75 |
+
|
| 76 |
+
### Complete dataset distribution by populated field
|
| 77 |
+
|
| 78 |
+
| Type.Field | Occurrences | % of total annot. |
|
| 79 |
+
| --- | --- | --- |
|
| 80 |
+
| EntiteAnonymisation.Category | 6976 | 100.00% |
|
| 81 |
+
| EntiteAnonymisation.RoleLOC | 87 | 1.25% |
|
| 82 |
+
| EntiteAnonymisation.RoleNUM | 8 | 0.11% |
|
| 83 |
+
| EntiteAnonymisation.RolePER | 2466 | 35.35% |
|
| 84 |
+
|
| 85 |
+
### Train / Dev Split Results
|
| 86 |
+
|
| 87 |
+
| Indicator | TRAIN | DEV |
|
| 88 |
+
| --- | --- | --- |
|
| 89 |
+
| Number of files | 408 | 101 |
|
| 90 |
+
| Number of annotations | 5580 | 1396 |
|
| 91 |
+
| Percentage of dataset | 79.99% | 20.01% |
|
| 92 |
+
| Avg. document length (chars) | 3957 | 4280 |
|
| 93 |
+
|
| 94 |
+
### Train / Dev distribution by type
|
| 95 |
+
|
| 96 |
+
| Type | TRAIN | DEV | % train of type |
|
| 97 |
+
| --- | --- | --- | --- |
|
| 98 |
+
| EntiteAnonymisation | 5580 | 1396 | 79.99% |
|
| 99 |
+
|
| 100 |
+
### Train / Dev distribution by populated field
|
| 101 |
+
|
| 102 |
+
| Type.Field | TRAIN | DEV | % train of field |
|
| 103 |
+
| --- | --- | --- | --- |
|
| 104 |
+
| EntiteAnonymisation.Category | 5580 | 1396 | 79.99% |
|
| 105 |
+
| EntiteAnonymisation.RoleLOC | 66 | 21 | 75.86% |
|
| 106 |
+
| EntiteAnonymisation.RoleNUM | 5 | 3 | 62.50% |
|
| 107 |
+
| EntiteAnonymisation.RolePER | 1971 | 495 | 79.93% |
|
| 108 |
+
|
| 109 |
+
### Train / Dev distribution by value
|
| 110 |
+
|
| 111 |
+
| Type.Field.Value | TRAIN | DEV | % train |
|
| 112 |
+
| --- | --- | --- | --- |
|
| 113 |
+
| EntiteAnonymisation.Category.IDENTIFYING_DATE | 2312 | 568 | 80.28% |
|
| 114 |
+
| EntiteAnonymisation.Category.LAST_NAME | 1111 | 287 | 79.47% |
|
| 115 |
+
| EntiteAnonymisation.RolePER.Patient | 978 | 257 | 79.19% |
|
| 116 |
+
| EntiteAnonymisation.RolePER.Carer | 973 | 232 | 80.75% |
|
| 117 |
+
| EntiteAnonymisation.Category.FIRST_NAME | 860 | 208 | 80.52% |
|
| 118 |
+
| EntiteAnonymisation.Category.FAMILY_STATUS | 767 | 203 | 79.07% |
|
| 119 |
+
| EntiteAnonymisation.Category.PATIENT_SOCIAL_IDENTITY | 424 | 99 | 81.07% |
|
| 120 |
+
| EntiteAnonymisation.Category.CITY | 43 | 11 | 79.63% |
|
| 121 |
+
| EntiteAnonymisation.RoleLOC.Patient | 33 | 13 | 71.74% |
|
| 122 |
+
| EntiteAnonymisation.Category.COUNTRY | 19 | 9 | 67.86% |
|
| 123 |
+
| EntiteAnonymisation.RolePER.Other | 20 | 6 | 76.92% |
|
| 124 |
+
| EntiteAnonymisation.RoleLOC.Hospital | 18 | 4 | 81.82% |
|
| 125 |
+
| EntiteAnonymisation.Category.UNIDENTIFYING_DATE | 19 | 0 | 100.00% |
|
| 126 |
+
| EntiteAnonymisation.RoleLOC.Other | 15 | 4 | 78.95% |
|
| 127 |
+
| EntiteAnonymisation.Category.PATIENT_BIRTHDATE | 12 | 5 | 70.59% |
|
| 128 |
+
| EntiteAnonymisation.Category.PHONE_NUMBER | 5 | 3 | 62.50% |
|
| 129 |
+
| EntiteAnonymisation.Category.PATIENT_NATIONALITY | 4 | 2 | 66.67% |
|
| 130 |
+
| EntiteAnonymisation.RoleNUM.Carer | 4 | 1 | 80.00% |
|
| 131 |
+
| EntiteAnonymisation.Category.ADDRESS | 3 | 0 | 100.00% |
|
| 132 |
+
| EntiteAnonymisation.RoleNUM.Hospital | 1 | 1 | 50.00% |
|
| 133 |
+
| EntiteAnonymisation.Category.URL | 1 | 1 | 50.00% |
|
| 134 |
+
| EntiteAnonymisation.RoleNUM.Other | 0 | 1 | 0.00% |
|
| 135 |
+
|
| 136 |
+
|
| 137 |
+
#### Data Origin
|
| 138 |
+
|
| 139 |
+
The clinical reports are extracted from the PARHAF corpus. Please refer to PARHAF documentation for more information about this corpus.
|
| 140 |
+
|
| 141 |
+
|
| 142 |
+
|
| 143 |
+
### Languages
|
| 144 |
+
|
| 145 |
+
- fr_FR
|
| 146 |
+
|
| 147 |
+
|
| 148 |
+
## Dataset Structure
|
| 149 |
+
|
| 150 |
+
We distribute both a Hugging Face dataset and a standalone version of the corpus. The standalone dataset consists of a JSON file per patient report, in [UIMA CAS JSON format](https://inception-project.github.io/releases/39.7/docs/user-guide.html#sect_formats_uimajson). This format constitutes the canonical version of the corpus. The Hugging Face dataset (Parquet/Arrow) is a derived representation generated automatically from the JSON files.
|
| 151 |
+
|
| 152 |
+
Both formats therefore contain identical information and differ only in storage layout.
|
| 153 |
+
|
| 154 |
+
One dataset instance corresponds to **one report**.
|
| 155 |
+
|
| 156 |
+
|
| 157 |
+
### Data Fields
|
| 158 |
+
|
| 159 |
+
|
| 160 |
+
| Path | Type | Description | Possible values |
|
| 161 |
+
| ---- | ---- | ----------- | --------------- |
|
| 162 |
+
| **document_metadata** | | | |
|
| 163 |
+
| `report` | string | Identifiant unique du rapport | |
|
| 164 |
+
| `full_text` | string | Texte intégral du rapport | |
|
| 165 |
+
| **spans** | | | |
|
| 166 |
+
| `report` | string | Identifiant du rapport | |
|
| 167 |
+
| `span_id` | integer | Identifiant de l'annotation | |
|
| 168 |
+
| `span_type` | string | Type de l'entité annotée | `EntiteAnonymisation` |
|
| 169 |
+
| `begin` | integer | Offset de début | |
|
| 170 |
+
| `end` | integer | Offset de fin | |
|
| 171 |
+
| `span_text` | string | Texte de l'entité | |
|
| 172 |
+
| `attribute_Categorie` | string | Categorie | `ADDRESS`, `CITY`, `COUNTRY`, `FAMILY_STATUS`, `FIRST_NAME`, `IDENTIFYING_DATE`, `LAST_NAME`, `PATIENT_BIRTHDATE`, `PATIENT_NATIONALITY`, `PATIENT_SOCIAL_IDENTITY` |
|
| 173 |
+
| `attribute_RolePER` | string | RolePER | `Carer`, `Other`, `Patient` |
|
| 174 |
+
| `attribute_RoleLOC` | string | RoleLOC | `Hospital`, `Other`, `Patient` |
|
| 175 |
+
| `attribute_RoleNUM` | string | RoleNUM | `Carer`, `Hospital` |
|
| 176 |
+
|
| 177 |
+
### Data Splits
|
| 178 |
+
|
| 179 |
+
Only the training set is released here. The remaining portion of the corpus will be temporarily embargoed to enable future evaluations under controlled conditions, thereby limiting the risk of large language model contamination through prior exposure to the data. You can evaluate your system on the test set through the CodaBench platform.
|
| 180 |
+
|
| 181 |
+
### Annotation Guidelines
|
| 182 |
+
You can find the detailed annotation protocol here: [annotation_guidelines.pdf](guidelines/annotation_guidelines.pdf)
|
| 183 |
+
|
| 184 |
+
### Hugging Face dataset
|
| 185 |
+
|
| 186 |
+
The Hugging Face dataset and the standalone JSON files provide the same information but use different data representations. On one hand, the Hugging Face dataset is stored in an optimized columnar Parquet format where data is structured into three distinct configurations: document_metadata for full texts, spans for annotations, and relations for links between spans. This flat table structure is specifically designed for efficient machine learning workflows and easy filtering. On the other hand, the standalone JSON files follow the original Inception/UIMA CAS nested structure where each document contains all its associated layers in a single file. While the organization shifts from a nested document-based view to an optimized relational view, the underlying data content remains strictly identical.
|
| 187 |
+
You can easily load and explore the different parts of this dataset using the Hugging Face datasets library as shown in the following example:
|
| 188 |
+
|
| 189 |
+
```python
|
| 190 |
+
from datasets import load_dataset
|
| 191 |
+
|
| 192 |
+
# Load document metadata and access the first report
|
| 193 |
+
ds_meta = load_dataset("HealthDataHub/PARHAF-pseudo-annotated", name="document_metadata")
|
| 194 |
+
first_doc = ds_meta["train"][0]
|
| 195 |
+
print(f"Report ID: {first_doc['report']}")
|
| 196 |
+
|
| 197 |
+
# Load annotations and iterate through the first samples
|
| 198 |
+
ds_spans = load_dataset("HealthDataHub/PARHAF-pseudo-annotated", name="spans")
|
| 199 |
+
for span in ds_spans["train"].select(range(5)):
|
| 200 |
+
print(f"Type: {span['span_type']} | Text: {span['span_text']}")
|
| 201 |
+
|
| 202 |
+
# Load and display a relation between entities (if they exist)
|
| 203 |
+
ds_rel = load_dataset("HealthDataHub/PARHAF-pseudo-annotated", name="relations")
|
| 204 |
+
if len(ds_rel["train"]) > 0:
|
| 205 |
+
rel = ds_rel["train"][0]
|
| 206 |
+
print(f"Relation: {rel['source_text']} -> {rel['target_text']}")
|
| 207 |
+
```
|
| 208 |
+
|
| 209 |
+
|
| 210 |
+
### Licensing Information
|
| 211 |
+
|
| 212 |
+
This dataset is released under licenses:
|
| 213 |
+
|
| 214 |
+
- CC BY 4.0
|
| 215 |
+
- Etalab 2.0
|
| 216 |
+
|
| 217 |
+
### Citation Information
|
| 218 |
+
|
| 219 |
+
[More Information Needed]
|
| 220 |
+
|
| 221 |
+
|
| 222 |
+
|
| 223 |
+
|