updated legal mc4 dataset

Multi_Legal_Pile.py

@@ -29,7 +29,7 @@ CASELAW = "caselaw"
 LEGISLATION = "legislation"
 CONTRACTS = "contracts"
 OTHER = "other"
-MC4_LEGAL = "mc4-legal"
+LEGAL_MC4 = "legal-mc4"
 
 # Native MultiLegalPile
 _DATA_NAME = {
@@ -319,13 +319,40 @@ _PILE_OF_LAW_NAMES = {
             "validation": ["validation.doj_guidance"]
         }
     },
-    MC4_LEGAL: {}
+    LEGAL_MC4: {}
+}
+
+_LEGAL_MC4_SHARDS = {
+    "bg": 0,
+    "cs": 2,
+    "da": 0,
+    "de": 8,
+    "el": 0,
+    "en": 1,
+    "es": 9,
+    "et": 1,
+    "fi": 0,
+    "fr": 2,
+    "ga": 0,
+    # "hr", # hr is not present in mc4
+    "hu": 0,
+    "it": 3,
+    "lt": 0,
+    "lv": 0,
+    "mt": 0,
+    "nl": 0,
+    "pl": 2,
+    "pt": 1,
+    "ro": 0,
+    "sk": 0,
+    "sl": 0,
+    "sv": 0,
 }
 
 _LANGUAGES = ["bg", "cs", "da", "de", "el", "en", "es", "et", "fi", "fr", "ga", "hr",
               "hu", "it", "lt", "lv", "mt", "nl", "pl", "pt", "ro", "sk", "sl", "sv"]
 
-_TYPES = [CASELAW, CONTRACTS, LEGISLATION, OTHER, MC4_LEGAL]
+_TYPES = [CASELAW, CONTRACTS, LEGISLATION, OTHER, LEGAL_MC4]
 
 _JURISDICTIONS = ["Austria", "Belgium", "Brazil", "Bulgaria", "Croatia", "Czechia", "Denmark", "Estonia", "Finland",
                   "France", "Germany", "Greece", "Hungary", "Ireland", "Italy", "Latvia", "Lithuania", "Luxembourg",
@@ -341,7 +368,7 @@ class MultiLegalPileConfig(datasets.BuilderConfig):
         Args:
             name: combination of language and type with _
             language:       One of bg,cs,da,de,el,en,es,et,fi,fr,ga,hr,hu,it,lt,lv,mt,nl,pl,pt,ro,sk,sl,sv or all
-            type:           One of caselaw,contracts,legislation,other,mc4_legal or all
+            type:           One of caselaw,contracts,legislation,other,legal-mc4 or all
           **kwargs: keyword arguments forwarded to super.
         """
         super(MultiLegalPileConfig, self).__init__(**kwargs)
@@ -411,18 +438,17 @@ class MultiLegalPile(datasets.GeneratorBasedBuilder):
                         info["filepath"] = download_url("joelito/eurlex_resources", f"{language}/intagr")
                         data_infos.append(info.copy())
 
-                    # add mc4_legal data
-                    if type == 'mc4_legal':
+                    # add legal-mc4 data
+                    if type == LEGAL_MC4:
                         info["jurisdiction"] = "N/A"
-                        if language in ["de", "en", "es"]:  # here we need to chunk because the files are too large
-                            for idx in [0, 1]:
-                                info["filepath"] = download_url("joelito/mc4_legal", f"{language}_{idx}")
-                                data_infos.append(info.copy())
-                        elif language in ["hr"]:
-                            pass  # nothing there for Croatian
+                        if language == "hr":
+                            pass
                         else:
-                            info["filepath"] = download_url("joelito/mc4_legal", f"{language}")
-                            data_infos.append(info.copy())
+                            for split in ["train", "validation"]:
+                                shards = range(_LEGAL_MC4_SHARDS[language] + 1) if split == "train" else [0]
+                                for shard in shards:
+                                    info["filepath"] = download_url("joelito/legal-mc4", f"{language}.{split}.{shard}")
+                                    data_infos.append(info.copy())
 
                     # add pile-of-law data
                     if language == 'en':

README.md

@@ -126,11 +126,11 @@ all_legislation').
 
 The file format is jsonl.xz and there is one split available ("train").
 
-The complete dataset (716GB) consists of four large subsets:
+The complete dataset (689GB) consists of four large subsets:
 
 - Native Multi Legal Pile (112GB)
 - Eurlex Resources (179GB)
-- MC4 Legal (133GB)
+- Legal MC4 (106GB)
 - Pile of Law (292GB)
 
 #### Native Multilingual Legal Pile data
@@ -172,9 +172,9 @@ The complete dataset (716GB) consists of four large subsets:
 
 See [Eurlex Resources](https://huggingface.co/datasets/joelito/eurlex_resources#data-instances) for more information.
 
-#### MC4 Legal
+#### Legal-MC4
 
-See [MC4 Legal](https://huggingface.co/datasets/joelito/mc4_legal#data-instances) for more information.
+See [Legal-MC4](https://huggingface.co/datasets/joelito/legal-mc4#data-instances) for more information.
 
 #### Pile-of-Law