Skip to content
This repository has been archived by the owner on Jan 10, 2025. It is now read-only.

Latest commit

 

History

History
122 lines (104 loc) · 5.36 KB

Q&A.md

File metadata and controls

122 lines (104 loc) · 5.36 KB

About

Here are listed issues we have witnesed during installation and how to solve them and also notices to remember

Notice -

While editing the config.txt file or changing the values manually, do not end the URL's with / as this will cause errors within the system.

Issue - After logging you will be directed back to login page

Make sure TARA has the correct IP (IP of your buerokratt project)

Issue - chat authentication error
GlobalExceptionHandler: Writing error: ErrorResponseBody[error=InvalidDataAccessApiUsageException, message=No value supplied for the SQL parameter 'receivedFromName': No value registered for key 'receivedFromName']
Issue - Error during script deployment
Creating network "bykstack" with driver "bridge"
ERROR: failed to update bridge store for object type *bridge.networkConfiguration: open /var/lib/docker/network/files/local-kv.db: no such file or directory

Solution - sudo apt install --reinstall docker-ce

Issue
Library "TEEMAD" does not load.

  • Description - After logging in as admin into backoffice GUI, the library "TEEMAD" will not load and you can see the loading circle non-stop.

  • Cause - It is caused by non-functioning ssh private key. When using keygen command you create a key in OPENSSH format, that java does not undersand.
    Looking at private-ruuter logs, you should see following example lines:

Caused by: com.jcraft.jsch.JSchException: invalid privatekey: [B@67ca1612

and

Unable to call function getBlacklistedIntents  java.lang.reflect.InvocationTargetException: null

This is caused by your RSA SSH key wrong format. When using keygen the defalt format is OPENSSH but it should be RSA

  • Solution - After you have created your SSH private key, use the command
ssh-keygen -p -f /home/ubuntu/.ssh/id_rsa -m pem

This will change your key into RSA format.
In the private.urls.docker.json and public.urls.docker.json files, make sure the info is correct, most important, the training_bot_directory_name

For example
dmapper_url": "https://byk-dmapper:8443",
   "ruuter_url": "https://localhost:8443",
   "tim_url": "https://byk-tim:8443",
   "resql_url": "https://byk-resql:8443",
   "bot_url": "http://BOT_ADDRESS:5005",
   "training_url": "TRAINING_BOT_ADDRESS",
   "training_user": "ubuntu",
   "training_prv_key":"/home/ubuntu/.ssh/id_rsa",
   "training_bot_directory_name": "/srv/Buerokratt-onboarding/chatbot-installation/Installation-Guides/default-setup/chatbot-and-training/bot_training/chatbot",
Issue
Training the bot results in a "Trained with errors" issue

  • Description - In backoffice GUI, when you want to train a new bot model, it will result a issue, where it has trained, but with errors

  • Cause - Looking in the private-ruuter` logs, you should see the following error/errors

2022-11-09T08:49:07.236Z ERROR []  : Unable to call function areLatestTestResultsPositive

More info to come

  • Solution - Solution is to use /home/username/ folder as the homefolder for trainbot files.

  • Follow-up solution - To use other folders, rather then /home/, then make sure, that train.sh file has correct paths added and that private.docker.urls.json has correct info. Example below train.sh

#!/bin/bash
echo 'Started training' >> train.log
rm -rf -v /opt/bot_training/chatbot-train/data >> train.log
rm -rf -v /opt/bot_training/chatbot-train/models >> train.log
rm -f -v /opt/bot_training/chatbot-train/domain.yml >> train.log
cp -r -v  /opt/bot_training/chatbot/data chatbot-train >> train.log
cp -v /opt/bot_training/chatbot/domain.yml chatbot-train >> train.log
mkdir -v /opt/bot_training/chatbot-train/models >> train.log
docker compose up train-bot
docker compose up test-bot
docker compose down

deploy.sh

#!/bin/bash
# Replace $DIR with user folder, where training files are located in
# Replace $SSH_KEY with ssh keyname for bot system
# Replace $BOT_USER with username of user on bot system
# Replace $BOT_HOST with hostname of bot system
# Replace $BOT_CONTAINER_NAME with name of the bots container on bot system
DIR=/home/ubuntu/lasttest/Installation-Guides/default-setup/chatbot-and-training/bot/loba
BOT_HOST='buerokratt'
BOT_USER='ubuntu'
SSH_KEY='/home/ubuntu/.ssh/authorized_keys'
BOT_CONTAINER_NAME='byk-bot'
BOT_DATA_DIR='/opt/bot/loba'
cd $DIR || exit 1
mkdir -p models/latest-model
cp /opt/bot_training/chatbot-train/models/* models/latest-model/
cp -r /opt/bot_training/chatbot-train/results models/latest-model/
mv /opt/bot_training/models/latest-model models/model-$(date +%Y%m%d%H%M%S)
scp -i $SSH_KEY chatbot-train/models/* $BOT_USER@$BOT_HOST:$BOT_DATA_DIR/models/
ssh $BOT_HOST -i $SSH_KEY -l $BOT_USER "docker restart $BOT_CONTAINER_NAME"

private-ruuter.docker.urls.json

"training_bot_directory_name": "chatbot",
Issue
AUTH FAIL

When your library "Teemad" is not loading and training the bot fails, take a look at your private ruuter logs. If you spot an error auth failed then check if the key type ssh-rsa is in PubkeyAcceptedAlgorithms

Issue

When training module in GUI tells, that "training module is flawed" and cannot be published and private-ruuter logs shw error cause "Cannot find the file" Make sure that your bot is running