Retrain with parquet files

[brabbit61]

• Extracts the corrected entities from the parquet file and store as json
• Populates the data_path folder (passed as an arg) with the JSON files
• Can process a combination of txt and json files during training (simultaneously).

[codecov]

Codecov Report

Patch coverage: 44.00% and project coverage change: -0.69 ⚠️

Comparison is base (97e145e) 48.01% compared to head (c319474) 47.33%.

Additional details and impacted files

@@            Coverage Diff             @@
##              dev      #87      +/-   ##
==========================================
- Coverage   48.01%   47.33%   -0.69%     
==========================================
  Files          19       19              
  Lines        1918     1971      +53     
==========================================
+ Hits          921      933      +12     
- Misses        997     1038      +41     

Impacted Files	Coverage Δ
src/preprocessing/labelling_preprocessing.py	18.06% <ø> (-0.53%)	⬇️
src/preprocessing/labelling_data_split.py	65.66% <43.54%> (-15.70%)	⬇️
.../hf_token_classification/huggingface_preprocess.py	75.47% <46.15%> (-9.64%)	⬇️

☔ View full report in Codecov by Sentry.
📢 Do you have feedback about the report comment? Let us know in this issue.

[tieandrews]

See PR #86 for these changes cherry picked into the organized structured