This project aims to test the integration between Rasa and Unity, contributing to the Rasa repository itself, with the intention of helping the community.
-
Clone the repository.
-
Install the dependencies of Rasa, following the steps here.
-
Run the Rasa creation command, selecting the creation path of your preference:
rasa initImportant: After creating the Rasa project, access the Rasa\ folder in this repository, and copy all the files to the folder created with the command rasa init. This step is necessary as it modifies the training data, which will be used in the Unity project.
- Run the Rasa training command:
rasa train- Run the Rasa initialization command:
rasa run- With Rasa running, open the Unity project in the directory
RasaUnitythat is in this repository, and run the project. Another alternative is to extract theBuild.zipfile and run the project with the executableRasaUnity.exe.
The project is a Unity application that uses Rasa to carry out a conversation between the Player and an NPC, with the conversation focused on programming. In this sense, Rasa is used to identify the intention of the message sent by the user, and respond according to the identified intention. After this occurs, Rasa returns the response to Unity, which in turn displays the response to the user.
- Windows 10
- Python 3.9.13
- rasa 3.4.0
- rasa-sdk 3.4.0
- Unity 2021.3.16f1
OBS: Project developed as a way to contribute to the Rasa repository, showing a way to connect Rasa with Unity. Issue #14 (RasaHQ/contributors)
OBS 2: Another approach to using Rasa with Unity is in the branch ImageRasaChat