Point to raw pre-merge export (provenance)
Browse files
README.md
CHANGED
|
@@ -132,3 +132,12 @@ from datasets import load_dataset
|
|
| 132 |
ds = load_dataset("felix453/interpretive-canons-instances", "wortlaut")
|
| 133 |
print(ds["test"][0])
|
| 134 |
```
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 132 |
ds = load_dataset("felix453/interpretive-canons-instances", "wortlaut")
|
| 133 |
print(ds["test"][0])
|
| 134 |
```
|
| 135 |
+
|
| 136 |
+
## Raw annotations (provenance)
|
| 137 |
+
|
| 138 |
+
The raw, pre-merge Label Studio export these data are built from is available at
|
| 139 |
+
[`felix453/interpretive-canons-raw-export`](https://huggingface.co/datasets/felix453/interpretive-canons-raw-export).
|
| 140 |
+
It retains two fields the postprocessing merge does not preserve: the per-reading
|
| 141 |
+
determinate-content judgment `bestimmt_moeglicheDeutung` (dropped by the merge)
|
| 142 |
+
and the un-flattened list of concrete provisions (the merge joins the entries
|
| 143 |
+
into one unsplittable string). Join by Label Studio region `id`.
|