Remove config names as yaml keys (#4367)

* validate YAML part

* fix bad YAML keys

* don't have config names as yaml keys

* style

* Remove hacking license tags when mirroring datasets on the Hub

* fix test

* fix style

* more fixes

* remove dots in QA column mapping keys

Commit from https://github.com/huggingface/datasets/commit/3f3013441e4af9bf4efa752b5e6e184a588b7a80

README.md

@@ -1,67 +1,38 @@
 ---
 annotations_creators:
-  monolingual:
-  - no-annotation
-  monolingual_raw:
-  - found
-  parallel:
-  - expert-generated
-  parallel_raw:
-  - expert-generated
+- expert-generated
+- found
+- no-annotation
 language_creators:
 - found
 languages:
-  monolingual:
-  - chr
-  - en
-  monolingual_raw:
-  - chr
-  parallel:
-  - chr
-  - en
-  parallel_raw:
-  - chr
-  - en
+- chr
+- en
 licenses:
 - other-different-license-per-source
 multilinguality:
-  monolingual:
-  - multilingual
-  monolingual_raw:
-  - monolingual
-  parallel:
-  - translation
-  parallel_raw:
-  - translation
+- monolingual
+- multilingual
+- translation
 size_categories:
-  monolingual:
-  - 100K<n<1M
-  monolingual_raw:
-  - 1K<n<10K
-  parallel:
-  - 10K<n<100K
-  parallel_raw:
-  - 10K<n<100K
+- 100K<n<1M
+- 10K<n<100K
+- 1K<n<10K
 source_datasets:
 - original
 task_categories:
-  monolingual:
-  - translation
-  monolingual_raw:
-  - text-generation
-  - fill-mask
-  parallel:
-  - translation
-  parallel_raw:
-  - translation
+- fill-mask
+- text-generation
+- translation
 task_ids:
-  monolingual: []
-  monolingual_raw:
-  - language-modeling
-  - masked-language-modeling
-  parallel: []
-  parallel_raw: []
+- language-modeling
+- masked-language-modeling
 paperswithcode_id: chren
+configs:
+- monolingual
+- monolingual_raw
+- parallel
+- parallel_raw
 ---
 
 # Dataset Card for ChrEn