fix(basqueglue): load the f1 metric with evaluate (#4076)

datasets.load_metric was removed in datasets 3.0 and the unit-test job
resolves datasets==5.0.1, so the module-level

    from datasets import load_metric

makes lm_eval/tasks/basqueglue/utils.py unimportable. Five of the six
BasqueGLUE tasks resolve a !function from that module, so they now fail to
load at all, not just at scoring time:

    bec.yaml    bec2016eu        utils.micro_f1_score
    bhtc.yaml   bhtc_v2          utils.micro_f1_score
    coref.yaml  epec_koref_bin   utils.coref_doc_to_text
    vaxx.yaml   vaxx_stance      utils.vaxx_f1_score
    wic.yaml    wiceu            utils.process_wic_docs

Load the metric with evaluate instead, the same move #2351 made for
squadv2, with the import inside the function as that fix and
kobest/utils.py do. The metric and the compute() arguments are unchanged.

lm_eval/tasks/basqueglue/utils.py

@@ -1,8 +1,6 @@
 import html
 import re
 
-from datasets import load_metric
-
 
 def general_detokenize(string):
     string = re.sub(r"\s+([.,;:!?)])", r"\1", string)
@@ -60,7 +58,9 @@ def coref_doc_to_text(x):
 
 
 def micro_f1_score(items):
-    f1_metric = load_metric("f1")
+    import evaluate
+
+    f1_metric = evaluate.load("f1")
     golds, preds = list(zip(*items))
     f1_score = f1_metric.compute(references=golds, predictions=preds, average="micro")[
         "f1"
@@ -69,7 +69,9 @@ def micro_f1_score(items):
 
 
 def vaxx_f1_score(items):
-    f1_metric = load_metric("f1")
+    import evaluate
+
+    f1_metric = evaluate.load("f1")
     golds, preds = list(zip(*items))
     f1_class = f1_metric.compute(
         references=golds, predictions=preds, labels=[0, 2], average=None