To get started with Ganimedes
, clone this repository locally and install the following prerequisites:
-
Create conda environment
conda create --name ganimedes --clone base
conda activate ganimedes
-
Install dependencies
cd functions/src
pip install -r requirements.txt
-
Compile OpenKE
cd model/kb/OpenKE
mkdir release
bash make.sh
-
Train Knowledge Embedding
mkdir -p ../artifacts/kge ../artifacts/dicts
python example_train_transe.py
python ../../../../../scripts/create_dicts.py
-
Firebase Authentication Set-up
- Follow the procedure available at (https://firebase.google.com/docs/admin/setup)
- Make a new directory
auth
and paste the json file - In the file
main.py
, change the name of the json authentication file to reflect this change
-
Start Local Server (DEBUG mode)
- Launch
Python: Flask
onmain_emulate.py
file . It should start the server on port 5000
- Launch
-
VSCode and install python extension
Currently, this project uses Jest to run its Unit Tests for each cloud functions. The tests consists basically on mocked calls (mock folder).
To run the test, execute the previous section steps from 1 to 4. After that, execute the following at functions folder:
npm run test
curl -X POST -H "Content-Type: application/json" -d @./functions/test/mock/getConceptDescription.json http://localhost:5000/dialogflow-fulfillment
-
Create make.sh
-
Fix get explanation and context
-
Move tests to pytest (deco supertest .ts)
-
Rethink deploy