AMI Corpus JSON Data Manupulator

Getting Started

1. Install Python 3:

   Go to install Python 3

2. Install Required Python Packages:

   $ pip install -r requirements.txt
   

3. Download Dataset

   Estimated Size: 1.94 GiB.

   $ python download_dataset.py
   

4. Update Dataset

   $ python download_dataset.py
   

---

Usage

python main.py [-h] [--meetings MEETINGS] [--transcript] [--dialogacts] [--extractivesummary] [--abstractivesummary]

-h: Help

--meetings: Meeting ids as comma-separated values

--transcript: Play Audio with Transcript

--dialogacts: Play Audio with Dialog Acts

--extsummary: Play Audio with Extractive Summary

--abssummary: Play Audio with Abstractive Summary

---

Examples

1. Play All Meeting Audios with Transcript:

$ python main.py --transcript

2. Play Specific Meeting Audio with Transcript:

$ python main.py --meetings ES2002a --transcript

3. Play Specific Meetings Audio with Transcript:

$ python main.py --meetings ES2002a,ES2002b,ES2003a --transcript

4. Play Specific Meeting Audio with Dialog Acts:

$ python main.py --meetings ES2002a --dialogacts

5. Get Extractive Summary in Specific Meeting:

$ python main.py --meetings ES2002a --extsummary

6. Get Abstractive Summary in Specific Meeting:

$ python main.py --meetings ES2002a --abssummary

Release Notes

v1.3 - 2021-06-06

• Added Confidence Level Data
• Updated Abstractive Summary Data

v1.2 - 2020-10-29

• Added Summary Links
• Added Arguments
• Added Discussions

v1.1 - 2020-10-12

• Added Transcripts
• Added Dialog Acts
• Added Adjacency Pairs
• Added Word Segmentations
• Added Topic Segmentations
• Added Decision Points
• Added Extractive Summaries
• Added Abstractive Summaries

v1.0 - 2020-05-20

• Initial Release with Meeting Audios only

License

Licensed under The MIT License (MIT)