Remove old train.parquet file and update new file with larger size; enhance convert.py to drop duplicates and reset index; add PTT article crawler script

Add token counting and text processing scripts with sample data files