Create Sudo user

STEP 1. ssh into your server.

ssh root@<server ip>

STEP 2. Use the adduser command to add a new user to your system and set the password for the user

adduser texttrack

STEP 3. Use the usermod command to add the user to the sudo group.

usermod -aG sudo texttrack

STEP 4. Test sudo access on new user account. Use the su command to switch to the new user account.

su texttrack or su - texttrack

Then execute any command with sudo to test the user.

sudo mkdir abc

Installing Ruby

Make sure you are logged in with new sudo user (not as root user)

Adding Node.js 10 repository

curl -sL https://deb.nodesource.com/setup_10.x | sudo -E bash -

Adding Yarn repository

curl -sS https://dl.yarnpkg.com/debian/pubkey.gpg | sudo apt-key add -
echo "deb https://dl.yarnpkg.com/debian/ stable main" | sudo tee /etc/apt/sources.list.d/yarn.list
sudo add-apt-repository ppa:chris-lea/redis-server

Refresh packege

sudo apt-get update

Install our dependencies for compiiling Ruby along with Node.js and Yarn

sudo apt-get install git-core curl zlib1g-dev build-essential libssl-dev libreadline-dev libyaml-dev libsqlite3-dev sqlite3 libxml2-dev libxslt1-dev libcurl4-openssl-dev software-properties-common libffi-dev dirmngr gnupg apt-transport-https ca-certificates redis-server redis-tools nodejs yarn

Install Rbenv

git clone https://github.com/rbenv/rbenv.git ~/.rbenv
echo 'export PATH="$HOME/.rbenv/bin:$PATH"' >> ~/.bashrc
echo 'eval "$(rbenv init -)"' >> ~/.bashrc
git clone https://github.com/rbenv/ruby-build.git ~/.rbenv/plugins/ruby-build
echo 'export PATH="$HOME/.rbenv/plugins/ruby-build/bin:$PATH"' >> ~/.bashrc
git clone https://github.com/rbenv/rbenv-vars.git ~/.rbenv/plugins/rbenv-vars
exec $SHELL
rbenv install 2.6.3
sudo chown -R texttrack.texttrack ~/.rbenv
rbenv global 2.6.3
ruby -v
#ruby 2.6.3

This installs the latest Bundler, currently 2.x.

gem install bundler

For older apps that require Bundler 1.x, you can install it as well.

gem install bundler -v 1.17.3

Test and make sure bundler is installed correctly, you should see a version number.

bundle -v
#Bundler version 2.0

Install rails

gem install rails

Some errors and solutions.

for example, Rails application : deepspeech-web (rails root dir) sudo user : texttrack

Error: if you get error related .ruby-version file
solution: delete .ruby-version file from your rails application root directory

Error: Your Ruby version is 2.6.3, but your Gemfile specified 2.6.1
solution: open your Gemfile and change the Gemfile version to 2.6.3

Error: There was an error while trying to write to `<path-to-rails-app>/deepspeech-web/Gemfile.lock`. It is likely that you need to grant write permissions for that path.

solution
sudo chown -R <user>:<user> path-to-rails_root

Example:
sudo chown -R texttrack:texttrack /usr/local/deepspeech-web

Running Rails Application

Go to rails root directory and execute following commands

Install dependencies

bundle install

Execute migration to create database

rake db:migrate

Security

** Create credentials and add apikey

touch credentials.yaml

* you can generate your own apikey. You don't need to go anywhere to get apikey.
sudo nano credentials.yaml (refer example-credentails.yaml in the working dir)

You can use port_number = 3000 or 4000

Install Mozilla deepspeech

1. Install dependencies

sudo apt-get install build-essential
sudo apt-get install aptitude
sudo apt-get install libstdc++6
sudo apt-get install libsox-dev
cd /usr/lib/x86_64-linux-gnu
sudo ln -s libsox.so.3 libsox.so.2
export PATH="$PATH:/usr/lib/x86_64-linux-gnu"

2. Install Python

sudo apt-get install python3-dev

#other dependencies
sudo apt-get install pkg-config zip g++ zlib1g-dev unzip wget
export PYTHON_INCLUDE_PATH="/usr/include/python3.6m"
export PYTHON_BIN_PATH="/usr/bin/python3.6"

3. Install Bazel and update PATH variable 3.1 Download Bazel 0.15 compatible to tensorflow 1.12.0

Create installation dir called deepspeech

sudo mkdir deepspeech
cd deepspeech
sudo mkdir temp
sudo apt-get install wget
sudo wget https://github.com/bazelbuild/bazel/releases/download/0.15.0/bazel-0.15.0-installer-linux-x86_64.sh

3.2 Run the Bazel installer

sudo chmod +x bazel-0.15.0-installer-linux-x86_64.sh
sudo ./bazel-0.15.0-installer-linux-x86_64.sh –-user --bin=$HOME/bin

#if you get an error in above command then remove user flag
$user/deepspeech: sudo ./bazel-0.15.0-installer-linux-x86_64.sh
$user/deepspeech: export PATH="$PATH:/home/<user>/bin"
#you will get path after successfully executing command
#for example: make sure you have /home/user/bin dir in your system
#The --user flag installs Bazel to the $HOME/bin directory on your system and sets the .bazelrc path to $HOME/.bazelrc.

3.3 Set up your environment

export PATH="$PATH:$HOME/bin"

4. Build deepspeech native 4.1 DeepSpeech clone and checkout

sudo apt-get install git
sudo git clone https://github.com/dabinat/DeepSpeech.git
cd DeepSpeech/
sudo git checkout timing-info
cd ..

4.2 Tensorflow clone and checkout

sudo git clone https://github.com/mozilla/tensorflow.git
cd tensorflow/
sudo git checkout origin/r1.12

4.3 create a symbolic link to the DeepSpeech native_client directory.

sudo ln -s ../DeepSpeech/native_client ./

5. Build DeepSpeech 5.1 make sure you have a python

sudo apt-get install python

5.2 execute build command and have a cup of tea because usually it takes 20-25 mins to build. However, time is depending on server configurations

sudo bazel build --config=monolithic -c opt --copt=-O3 --copt="-D_GLIBCXX_USE_CXX11_ABI=0" --copt=-fvisibility=hidden //native_client:libdeepspeech.so //native_client:generate_trie

5.3 set environment variable

export TFDIR=~/workspace/tensorflow

5.4 make sure you have deepspeech file in your DeepSpeech dir and tensaflow dir. If file exist in both dir then execute following command.

cd ../DeepSpeech/native_client
make deepspeech

6. Copy binaries into deepspeech/temp folder

cp deepspeech ~/deepspeech/temp/deepspeech
cp ~/deepspeech/tensorflow/bazel-bin/native_client/libdeepspeech.so ~/deepspeech/temp/libdeepspeech.so
cd ~/deepspeech/temp

7. Download model and audio files

sudo curl -LO https://github.com/mozilla/DeepSpeech/releases/download/v0.6.1/deepspeech-0.6.1-models.tar.gz
sudo tar xvf deepspeech-0.6.1-models.tar.gz
sudo curl -LO https://github.com/mozilla/DeepSpeech/releases/download/v0.6.1/audio-0.6.1.tar.gz
sudo tar xvf audio-0.6.1.tar.gz

8. If you want to update the deepspeech version in future (depending on if you want to use the gpu or not - only install 1 both pip & pip3)

pip install deepspeech==0.6.1
pip3 install deepspeech==0.6.1

pip install deepspeech-gpu==0.6.1
pip3 install deepspeech-gpu==0.6.1

Now we need to replace our old deepspeech with the new one in home/deepspeech/temp

cd /home/deepspeech/temp
sudo mv deepspeech deepspeech_old
which deepspeech

which deepspeech will give you the link to where the new one is located (now to copy that to /home/deepspeech/temp)

sudo cp /home/texttrack/.local/bin/deepspeech /home/deepspeech/temp

9. Check if deepspeech native is working

cd /home/deepspeech/temp
./deepspeech --model deepspeech-0.6.1-models/output_graph.pbmm --lm deepspeech-0.6.1-models/lm.binary --trie deepspeech-0.6.1-models/trie --audio audio/2830-3980-0043.wav

* if you have installed models on the different path then you need to update the setting.yaml file

At this point, you should get words for sample audios inside the temp/audio directory.

Redis server

sudo apt-get update
sudo apt-get upgrade

1. Install & enable Redis server

sudo apt-get install redis-server
sudo systemctl enable redis-server.service

2. Configure redis server

sudo vim /etc/redis/redis.conf
#you can use nano instead of vim
#copy following lines or find and remove comments for these lines
maxmemory 256mb
maxmemory-policy allkeys-lru
#save and exit

3. Restart service

sudo systemctl restart redis-server.service

4. Install redis php extension

sudo apt-get install php-redis

5. Test the connection

redis-cli ping
#you should get response "PONG"

6. Important commands

#restart redis server
sudo systemctl restart redis
#redis-cli commands
redis-cli info
redis-cli stats
redis-cli server

Install faktory worker

sudo wget https://github.com/contribsys/faktory/releases/download/v1.0.1-1/faktory_1.0.1-1_amd64.deb

sudo dpkg -i faktory_1.0.1-1_amd64.deb

sudo cat /etc/faktory/password (Manually find password if ever needed)

Set password for worker and Service

Find password

cat /etc/faktory/password

Set password

Open working_dir/service/deepspeech_worker.service and find this line given below. Environment=LANG=en_US.UTF-8 FAKTORY_PROVIDER=FAKTORY_URL FAKTORY_URL=tcp://:<password>@localhost:7419 Replace your pasword with . save & exit Do the same thing for working_dir/service/deepspeech_service.service

Copy /usr/local/deepspeech-web/service/.service to /etc/systemd/system

cd /usr/local/deepspeech-web/service/
sudo cp *.service /etc/systemd/system

Start all services

sudo systemctl enable deepspeech_rails
sudo systemctl start deepspeech_rails

sudo systemctl enable deepspeech_service
sudo systemctl start deepspeech_service

sudo systemctl enable deepspeech_worker
sudo systemctl start deepspeech_worker

Restart/stop services

sudo systemctl restart service-name
sudo systemctl stop service-name

Check status

sudo systemctl status service-name

Edit service

sudo systemctl edit service-name --full
sudo systemctl daemon-reload (After editing)
sudo journalctl -u service-name.service

Check logs

sudo journalctl -u service-name.service

Clear logs

sudo journalctl --rotate
sudo journalctl --vacuum-time=1s

Keep your code clean with Rubocop

1. Install rubocop

gem install rubocop

2. Run rubocop to find coding offences

cd <working_dir>
rubocop

3. Safe auto corrections

rubocop --safe-auto-correct --disable-uncorrectable