Update app.py

app.py

@@ -3,22 +3,19 @@ import numpy as np
 import os
 from huggingface_hub import hf_hub_download
 from camel_tools.data import CATALOGUE
+from camel_tools.tagger.default import DefaultTagger
+from camel_tools.disambig.bert import BERTUnfactoredDisambiguator
 
-
-
-CATALOGUE.download_package("light",
-                           recursive=True,
-                           force=True,
-                           print_status=True)
 def predict_label(text):
     
     ip = text.split()
     ip_len = [len(ip)]
     
-    span_scores = extract_spannet_scores(span_model,ip,ip_len, pos_col=1, task_col=2)
+    span_scores = extract_spannet_scores(span_model,ip,ip_len)
     span_pooled_scores = pool_span_scores(span_scores, ip_len)
 
-    msa_span_scores = extract_spannet_scores(msa_span_model,ip,ip_len, pos_col=1, task_col=2, pos='not none')
+    pos_tags = tagger.tag(ip)
+    msa_span_scores = extract_spannet_scores(msa_span_model,ip,ip_len,pos=pos_tags)
     msa_pooled_scores = pool_span_scores(msa_span_scores, ip_len)
 
     ensemble_span_scores = [score for scores in [span_scores, msa_span_scores] for score in scores]
@@ -41,21 +38,22 @@ if __name__ == '__main__':
                     filename=file,
                     local_dir='src',
                     token=space_key)
+        
+    CATALOGUE.download_package("light",
+                           recursive=True,
+                           force=True,
+                           print_status=True)
     
     from src.predict import extract_spannet_scores,extract_ent_scores,pool_span_scores,pool_ent_scores
     from src.network import SpanNet, EntNet
     from src.validate import entities_from_token_classes
 
-    from camel_tools.disambig.mle import MLEDisambiguator
-    from camel_tools.tagger.default import DefaultTagger
-
-    mled = MLEDisambiguator.pretrained()
-    tagger = DefaultTagger(mled, 'pos')
 
-    print(tagger.tag('ذهبت الى المدرسة'.split()))
+    diasmbig = BERTUnfactoredDisambiguator.pretrained('msa')
+    tagger = DefaultTagger(diasmbig, 'pos')
     
     span_path = 'models/span.model'
-    msa_span_path = 'models/msa.best.model'
+    msa_span_path = 'new_models/msa.best.model'
     entity_path= 'models/entity.msa.model'
     
     span_model = SpanNet.load_model(span_path)