How to setup the server

Fill the needed data inside the instance/config.toml file.
Then, you can use docker compose to run the development server

docker compose up -d --build

How to use the transcribe api

You can use the following curl request to transcribe a video

curl -X POST "http://localhost:5000/api/transcribe" \
  -H "X-API-PASSCODE: YOUR_PASSCODE" \
  -F "video=@/absolute/path/to/video.mp4" \
  -F "language=it"

with YOUR_PASSCODE being the secret_passcode set in the API section of instance/config.toml. For the video, you can use one of the videos provided in the sample_videos/.
In this case the video is loaded from the filesystem. However, it can also be sent as a part of the HTTPS request.

How to use the generate_qa api

For example, you can use the following curl request

curl -X POST "http://localhost:5000/api/generate_qa" \  
  -H "X-API-PASSCODE: YOUR_PASSCODE" \  
  -F "language=it" \  
  -F "competency_level=medium" \  
  -F "num_questions=10" \  
  -F "num_multiple_answers=4" \  
  -F "topic=Alessandro Magno" \  
  -F "transcription=<transcription_example.json"  

with YOUR_PASSCODE being the secret_passcode set in the API section of instance/config.toml.
The request will return a list of dictionaries, each containing the question text, the possible answers, and the position (in python indexing) of the correct answer.

How to use the evaluate api

curl -X POST "http://localhost:5000/api/evaluate" \  
  -H "X-API-PASSCODE: YOUR_PASSCODE" \  
  -F "language=it" \  
  -F "competency_level=medium" \  
  -F "topic=Alessandro Magno" \  
  -F "transcription=<transcription_example.json" \  
  -F "questions=<questions.json" \  
  -F "user_answers=<user_answers.json" \  
  -F "correct_answers=<correct_answers.json"  

with YOUR_PASSCODE being the secret_passcode set in the API section of instance/config.toml.
The request will return the list of correct answers, the weak topics identified and a feedback message indicating which part needs to be revised.