This is a backend implementation for a chatbot using the open-source machine learning framework Rasa.
-
Ask for current or next talk in a specific room
- Example: "Inform me about the next lecture in room A2"
- Response : "The next talk in room A2 is 'Addressing students’ eco-anxiety when teaching sustainability in computing education' by John Doe, Joseph Doe on Monday, 01 January 2024 at 00:00:00."
-
Ask for the time and date a specific article is presented
- Example: "When is the article Addressing students’ eco-anxiety when teaching sustainability in computing education presented?"
- Response : "John Doe, Joseph Doe will present the article 'Addressing students’ eco-anxiety when teaching sustainability in computing education' on Monday, 01 January 2024 at 00:00:00 at location A2. You can find more information about the talk at https://example.com/talk.html."
-
Ask for the time and date a specific speaker is presenting
- Example: "What time is the talk by John Doe?"
- Response : "John Doe will be presenting the following talks: Ethical implications of artificial on Monday, 01 January 2018 at 00:00:00 in room A2 Addressing students’ eco-anxiety when teaching sustainability in computing education on Monday, 01 January 2024 at 00:00:00 in room A2"
-
Ask for the conference schedule
- Example: "When is the conference?"
- Response : "The event has ended. It was held from Tuesday 01 January 2019 00:00:00 to Tuesday 01 January 2019 00:00:00"
-
Ask for the venue address
- Example: "What is the address of the venue?"
- Response : "The venue is located at Université de Rennes - Campus Beaulieu, 263 Av. Général Leclerc, 35000 Rennes. You can find it on Google Maps here: https://www.google.com/maps?q=Universit%C3%A9+de+Rennes+-+Campus+Beaulieu%2C+263+Av.+G%C3%A9n%C3%A9ral+Leclerc%2C+35000+Rennes"
-
Ask for restaurant recommendations
- Example: "Where can I eat?"
- Response : "Here is a list of restaurants near the venue (Université de Rennes - Campus Beaulieu, 263 Av. Général Leclerc, 35000 Rennes): https://www.google.com/maps/search/restaurants+near+Universit%C3%A9+de+Rennes+-+Campus+Beaulieu%2C+263+Av.+G%C3%A9n%C3%A9ral+Leclerc%2C+35000+Rennes"
-
Ask for things to do in the city
- Example: "What are the things to see in this area?"
- Response : "Here is a list of things to see near the venue (Université de Rennes - Campus Beaulieu, 263 Av. Général Leclerc, 35000 Rennes): https://www.google.com/maps/search/things+to+see+near+Universit%C3%A9+de+Rennes+-+Campus+Beaulieu%2C+263+Av.+G%C3%A9n%C3%A9ral+Leclerc%2C+35000+Rennes"
-
Ask when and where the next meal is
- Example: "When is the next meal?"
- Response : "The next meal is Lunch on Tuesday, 06 June 2023 at 12:00:00 in the Cafeteria. Enjoy your meal!"
-
Ask when is the next talk
- Example: "When is the next talk?"
- Response : "Please provide either a room location or the name of the speaker for the next talk."
- Example: "Lukas Koedijk"
- Response : "Lukas Koedijk will be presenting the following talks: Finding Significant Differences in the Energy Consumption when Comparing Programming Languages and Programs. on Tuesday, 06 June 2023 at 09:30:00 in room B1"
-
Ask when and where the conference reception is taking place
-
Retrieve the link to an event
-
Ask for the registration desk's opening hours
-
Ask for the link to the conference proceedings
-
Ask for (or the link to) a campus map
-
Ask how to get to the reception or gala
-
Ask for directions to a specific location
The backend consists of two parts: the Rasa NLU model and the Rasa actions server.
graph LR
B(Chatbot)
B --> C[Rasa NLU model and HTTP endpoint]
C --> E[Actions Server]
E --> C
Download a pretrained model from the GitHub artifacts and place it in the models folder.
Or
To train the model, run the following command:
rasa trainTo start the actions server, run the following command in the root directory of the project:
rasa run actionsTo start the backend server, run the following command in the root directory of the project:
export ACTION_SERVER=localhost
export ACTION_PORT=5055
rasa run --enable-api --cors "*"This will start the server at http://localhost:5005.
To send a message to the chatbot, send a POST request to the /webhooks/rest/webhook endpoint with the following body:
{
"sender": "test",
"message": "Hello"
}The sender field is used to identify the user. The message field is the message that the user sent.
The response will be a JSON array with the following structure:
[
{
"recipient_id": "test",
"text": "Hello, how can I help you?"
}
]The recipient_id field is the same as the sender field in the request. The text field is the response from the chatbot.
This project is licensed under the Apache License 2.0 - see the LICENSE file for details.