Convert to parquet (loading scripts are no longer supported)

`datasets` no longer supports dataset loading scripts, so `load_dataset('biglam/contentious_contexts')` fails with `RuntimeError: Dataset scripts are no longer supported`.

This PR converts the dataset to parquet and removes the loading script. The join and schema match the script exactly: one row per extract, with every annotator response carried in both Dutch and English translation alongside the free-text suggestions. All response strings map cleanly to the script's English lookup (asserted in the build).

The card's Loading section said loading was broken, which this PR fixes, so that paragraph is replaced with a working snippet. Nothing else in the card is touched.

Source data unchanged: https://github.com/cultural-ai/ConConCor

README.md

@@ -26,6 +26,43 @@ task_categories:
 task_ids:
 - sentiment-scoring
 - multi-label-classification
+dataset_info:
+  features:
+  - name: extract_id
+    dtype: string
+  - name: text
+    dtype: string
+  - name: target
+    dtype: string
+  - name: annotator_responses_english
+    list:
+    - name: id
+      dtype: string
+    - name: response
+      dtype: string
+  - name: annotator_responses_dutch
+    list:
+    - name: id
+      dtype: string
+    - name: response
+      dtype: string
+  - name: annotator_suggestions
+    list:
+    - name: id
+      dtype: string
+    - name: suggestion
+      dtype: string
+  splits:
+  - name: train
+    num_bytes: 2199414
+    num_examples: 2720
+  download_size: 1940545
+  dataset_size: 2199414
+configs:
+- config_name: default
+  data_files:
+  - split: train
+    path: data/train-*
 ---
 
 # Contentious Contexts Corpus (ConConCor)

data/train-00000-of-00001.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cbfc97851b492f569c346cbc16cd825c9eb4b7242ed1eb55d8eb10c4788e5bb6
+size 1940545