score-using-embedding fails a lot

[saggu]

Attaching 5 files, and their auxiliary files, the command to be used is

tl align-page-rank $file \
            / string-similarity -i --method symmetric_monge_elkan:tokenizer=word -o monge_elkan \
            / string-similarity -i --method symmetric_monge_elkan:tokenizer=word -c label_clean kg_aliases -o monge_elkan_aliases \
            / string-similarity -i --method jaro_winkler -o jaro_winkler \
            / string-similarity -i --method levenshtein -o levenshtein \
            / string-similarity -i --method jaccard:tokenizer=word -c kg_descriptions context -o des_cont_jaccard \
            / normalize-scores -c des_cont_jaccard / smallest-qnode-number \
            / mosaic-features -c kg_labels --num-char --num-tokens \
            / create-singleton-feature -o singleton \
            / vote-by-classifier  \
            --prob-threshold 0.995 \
            --model $classifier_model_path \
            / score-using-embedding \
            --column-vector-strategy centroid-of-lof \
            --lof-strategy ems-mv \
            -o lof-graph-embedding-score \
            --embedding-file $embedding_file \
            / generate-reciprocal-rank  \
            -c lof-graph-embedding-score \
            -o lof-reciprocal-rank \
            / compute-tf-idf  \
            --feature-file $class_count_file \
            --feature-name class_count \
            --singleton-column singleton \
            -o lof_class_count_tf_idf_score \
            / compute-tf-idf \
            --feature-file $property_count_file \
            --feature-name property_count \
            --singleton-column singleton \
            -o lof_property_count_tf_idf_score \
            / context-match --context-file $context_file \
            -o context_score \
            > $output_file

Error:

Error Message:  Traceback (most recent call last):
  File "/Users/amandeep/Github/table-linker/tl/cli/score-using-embedding.py", line 74, in run
    vector_transformer.process_vectors()
  File "/Users/amandeep/Github/table-linker/tl/features/external_embedding.py", line 158, in process_vectors
    if not self._centroid_of_lof():
  File "/Users/amandeep/Github/table-linker/tl/features/external_embedding.py", line 328, in _centroid_of_lof
    lof_pred = clf.fit_predict(vectors)
  File "/Users/amandeep/Github/table-linker/tl_env/lib/python3.9/site-packages/sklearn/neighbors/_lof.py", line 246, in _fit_predict
    return self.fit(X)._predict()
  File "/Users/amandeep/Github/table-linker/tl_env/lib/python3.9/site-packages/sklearn/neighbors/_lof.py", line 265, in fit
    self._fit(X)
  File "/Users/amandeep/Github/table-linker/tl_env/lib/python3.9/site-packages/sklearn/neighbors/_base.py", line 514, in _fit
    raise ValueError(
ValueError: Expected n_neighbors > 0. Got 0

data.zip

[saggu]

file for classifier_model_path can be downloaded from https://github.com/usc-isi-i2/table-linker-pipelines/blob/main/table-linker-full-pipeline/models/weighted_lr.pkl