Real-time suggestions for internal and external product documentation.
(Not strictly required but good practice. Could use a docker container as well)
conda create --name docusync python=3.8.10
conda activate docusync
python populate_documentation.py
npm install
cd ..
cd server
pip install -r requirements.txt
cd ..
USERNAME = "<username>"
PASSWORD = "<password>"
OPENAI_API_KEY = "<api_key>"
Register and create a database on MonogoDB Atlas. For this project, the name of the database is 'docusync' and the names of the collections are 'documentation' & 'suggestions'.
Replace the URI in server/config.py
with your own database URI.
cd server/direct_db_edits
python populate_documentation.py
python populate_suggestions.py
cd ..
cd client
npm start
cd ..
cd server
python app.py
cd ..
Go to localhost:3000
Devpost Entry for the HackerDojo Generative AI hackathon
- You can use the
server/direct_db_edits/reset_pending.py
script to get back all accepted suggestions (setspending
to True so that it is displayed in the diff screen) - You can use the
server/direct_db_edits/delete_answers.py
script to delete 'answer', 'old_doc' and 'new_doc' fields from all suggestions.