Upload download_eurovoc_filter_by_language.py

download_eurovoc_filter_by_language.py

@@ -0,0 +1,32 @@
+from tqdm import tqdm
+import datasets
+from tqdm import tqdm 
+import pandas as pd
+
+configs = datasets.get_dataset_config_names("EuropeanParliament/Eurovoc")
+ 
+data = []
+for c in tqdm(configs):
+    dataset = datasets.load_dataset("EuropeanParliament/Eurovoc", c)
+    data.append(dataset)
+
+dataset = datasets.concatenate_datasets([d['train'] for d in data])
+
+# Downloads in ~1h, then 20 mins to process
+# 100%|██████████| 334/334 [19:36<00:00,  3.52s/it]
+
+def eng_only(dataset):
+    df = dataset['train'].to_pandas()#.lang.unique()
+    df = df[df["lang"] == "eng"]
+    return df
+
+eng_dfs = []
+
+for i in tqdm(data):
+    eng_dfs.append(eng_only(i))
+
+# 100%|██████████| 334/334 [04:25<00:00,  1.26it/s]
+
+df = pd.concat(eng_dfs)
+
+df.to_parquet("CELLAR_EN_16_05_2024.parquet")