install

- https://github.com/seatgeek/fuzzywuzzy
- pip install fuzzywuzzy

1. smi_cleaning.py

   • input : .smi file
   • output : smi_cleaning.txt
   • remove unnecessary lines and
     tag from smi file.
   • seperate two conversation in one capture

2. json_cleaning.py

   • read ['transcript'] element from json file.
   • remove unnecessary line. ex) ( ) [ ] pattern

3. smi_superset.py

   • input : smi_cleaning.txt
   • output : smi_superset1.txt (Data structure: [("caption", start_t,end_t),("caption", start_t, end_t)...] )

4.trans_smi_matching.py - input : json_cleaning.txt, smi_cleaning.txt - fuzzy string matching: https://github.com/seatgeek/fuzzywuzzy - output: sample_result_ver_1.txt

5.trans_smi_matching2.py - output : sample_result_ver_2.txt