Update fincat_utils.py

fincat_utils.py

@@ -29,26 +29,7 @@ def extract_context_words(x, window = 6):
 """The following functions have been created with inspiration from https://github.com/arushiprakash/MachineLearning/blob/main/BERT%20Word%20Embeddings.ipynb"""
 
 def bert_text_preparation(text, tokenizer):
- """Preparing the input for BERT
- 
- Takes a string argument and performs
- pre-processing like adding special tokens,
- tokenization, tokens to ids, and tokens to
- segment ids. All tokens are mapped to seg-
- ment id = 1.
- 
- Args:
- text (str): Text to be converted
- tokenizer (obj): Tokenizer object
- to convert text into BERT-re-
- adable tokens and ids
- 
- Returns:
- list: List of BERT-readable tokens
- obj: Torch tensor with token ids
- obj: Torch tensor segment ids
- 
- """
+
  marked_text = "[CLS] " + text + " [SEP]"
  tokenized_text = tokenizer.tokenize(marked_text)
  indexed_tokens = tokenizer.convert_tokens_to_ids(tokenized_text)
@@ -61,22 +42,7 @@ def bert_text_preparation(text, tokenizer):
  return tokenized_text, tokens_tensor, segments_tensors
  
 def get_bert_embeddings(tokens_tensor, segments_tensors, model):
- """Get embeddings from an embedding model
- 
- Args:
- tokens_tensor (obj): Torch tensor size [n_tokens]
- with token ids for each token in text
- segments_tensors (obj): Torch tensor size [n_tokens]
- with segment ids for each token in text
- model (obj): Embedding model to generate embeddings
- from token and segment ids
- 
- Returns:
- list: List of list of floats of size
- [n_tokens, n_embedding_dimensions]
- containing embeddings for each token
- """
- 
+
  # Gradient calculation id disabled
  # Model is in inference mode
  with torch.no_grad():
@@ -106,5 +72,5 @@ def bert_embedding_extract(context_text, word):
  word_embedding_all.append(word_embedding)
  word_embedding_mean = np.array(word_embedding_all).mean(axis=0)
  return word_embedding_mean
- except:
- return ['None']
+ except:
+ return ['None']