- errbot or install it
- python > 3
- virtualenv
You will have to install virtualenv on to have a nice python environment to work in. Having that you can clone the repository where ever you whant to:
git clone git@github.com:mayflower/err-rasa.git plugins/rasa-plugin
cd plugins/rasa-pluginI would create a python environment there:
virtualenv --python `which python3` .errbot-ve
.errbot-ve/bin/pip install errbot
# install required python packages
.errbot-ve/bin/pip install -r requirements.txtThe folder lives inside the .gitignore and doing so we have a nice errbot installed inside our environment.
Now you have to prepare your working folder:
# create an empty folder for later data
mkdir data/
# create a demo config files with most needed parameters
cp config_slack.py config.py
# set `BOT_EXTRA_PLUGIN_DIR` value to `./plugin` - inside your working errbot instance you have to adopt this oneTo have the german language available through spacy we have to download it:
# install and link german language pack for spacy (the analyzer)
python -m spacy download deWe prepared a tool with name chatitot to create training data from a markdown file. So to use those data you have to run
a npm install and then run
# creates config/training_data.json from config/chatito
npm run chatito
# builds nlu model from config/training_data.json
npm run build:nlu-model
# builds dialoque model form config/stories.md and config/training_data.json
npm run build:dialog-model
# run trainer to get a better conversation
npm run train:dialogNow grab your Slack-API-Key from Slack apps overview (There you have to create an app for you for Errbot if not done yet). If you click then into one app, the api token key is listed between the app properties.
Then pass the key to the configuration as the last task and you can start your bot (BOT_IDENTITY in config.py).
# run errbot
.errbot-ve/bin/errbotUsing the -T or --text command will open a text based version inside your console. This command will open a connection to your slack.