Update playbook to download nltk data, remove debian package listing

playbooks/roles/mlapi/tasks/main.yml

@@ -12,27 +12,27 @@
   command: sysctl vm.overcommit_memory=1
   sudo: yes
 - name: set permissions on $app_base_dir sgid for edx
-  file: path=$app_base_dir owner=ubuntu group=edx mode=2775 state=directory
-  file: path={{venv_dir}} owner=ubuntu group=edx mode=2775 state=directory
+  file: path=$app_base_dir owner={{remote_user}} group=edx mode=2775 state=directory
+  file: path={{venv_dir}} owner={{remote_user}} group=edx mode=2775 state=directory
   sudo: True
 - name: Install git so that we can clone repos
   apt: pkg=git install_recommends=yes state=present
   sudo: True
 - name: create edx virtualenv directory
-  file: path={{ml_api_dir}} owner=ubuntu group=edx mode=2775 state=directory
+  file: path={{ml_api_dir}} owner={{remote_user}} group=edx mode=2775 state=directory
   sudo: True
 - name: create edx virtualenv directory
-  file: path={{ml_dir}} owner=ubuntu group=edx mode=2775 state=directory
+  file: path={{ml_dir}} owner={{remote_user}} group=edx mode=2775 state=directory
   sudo: True
 - name: git checkout ml-service-api repo into $app_base_dir
   git: dest={{ml_api_dir}} repo={{mlapi_source_repo}}
 - name: git checkout ml-service-api repo into $app_base_dir
   git: dest={{ml_dir}} repo={{ml_source_repo}}
 - name: install ml-service-api apt packages
-  command: xargs -a $app_base_dir/ml-service-api/apt-packages.txt apt-get install
+  command: xargs -a $app_base_dir/ml-service-api/apt-packages.txt apt-get install -y
   sudo: yes
 - name: install machine-learning apt packages
-  command: xargs -a $app_base_dir/machine-learning/apt-packages.txt apt-get install
+  command: xargs -a $app_base_dir/machine-learning/apt-packages.txt apt-get install -y
   sudo: yes
 - name : install python pre-requirements
   pip: requirements="{{ml_api_dir}}/pre-requirements.txt" virtualenv="{{venv_dir}}" state=present
@@ -42,6 +42,13 @@
   pip: requirements="{{ml_api_dir}}/requirements.txt" virtualenv="{{venv_dir}}" state=present
 - name : install python requirements
   pip: requirements="{{ml_dir}}/requirements.txt" virtualenv="{{venv_dir}}" state=present
+- name: render nltk template
+  template: src=install_nltk_data.sh.j2 dest=$app_base_dir/install_nltk_data.sh
+- name: install nltk data
+  action: script $app_base_sir/install_nltk_data.sh
+- name: set permissions on $app_base_dir sgid for edx
+  file: path={{nltk_data_dir}} owner={{remote_user}} group=edx mode=2775 state=directory
+  sudo: True
 - name: django syncdb
   django_manage: >
       command=syncdb

playbooks/roles/mlapi/templates/install_nltk_data.sh.j2

+if [ ! -d "{{nltk_data_dir}}"]; then
+    source {{venv_dir}}/bin/activate
+    python -m nltk.downloader -d {{nltk_data_dir}} all
+fi
\ No newline at end of file

playbooks/roles/mlapi/vars/main.yml

@@ -3,30 +3,5 @@ ml_source_repo: git@github.com:edx/machine-learning.git
 ml_dir: $app_base_dir/machine-learning
 ml_api_dir: $app_base_dir/ml-service-api
 ml_api_settings: ml_service_api.settings
-mlapi_debian_pkgs:
-  - python
-  - gcc
-  - g++
-  - aspell
-  - python-virtualenv
-  - build-essential
-  - python-dev
-  - gfortran
-  - libfreetype6-dev
-  - libpng12-dev
-  - libxml2-dev
-  - libxslt1-dev
-  - libreadline6
-  - libreadline6-dev
-  - redis-server
-  - python-pip
-  - ipython
-  - nginx
-  - libmysqlclient-dev
-  - libblas3gf
-  - libblas-dev
-  - liblapack3gf
-  - liblapack-dev
-  - libatlas-base-dev
-  - curl
-  - yui-compressor
\ No newline at end of file
+nltk_data_dir: /usr/share/nltk_data
+remote_user: ubuntu
\ No newline at end of file