Merge pull request #1130 from edx/will/ora2-ai-grading

Install NLTK for edxapp workers

playbooks/roles/edxapp/meta/main.yml

@@ -6,6 +6,7 @@ dependencies:
     rbenv_dir: "{{ edxapp_app_dir }}"
     rbenv_ruby_version: "{{ edxapp_ruby_version }}"
   - devpi
+  - nltk
   - role: user
     user_info:
       - name: "{{ EDXAPP_AUTOMATOR_NAME }}"

playbooks/roles/nltk/defaults/main.yml

+---
+NLTK_DATA_DIR: "/usr/local/share/nltk_data"
+
+# Once the file is downloaded, it won't be downloaded again,
+# so if you need to version the data files, you should upload
+# your own version of the files with the version appended to the filename.
+NLTK_DATA:
+  - { path: "taggers/maxent_treebank_pos_tagger",
+      url: "http://nltk.github.com/nltk_data/packages/taggers/maxent_treebank_pos_tagger.zip" }
+  - { path: "corpora/stopwords",
+      url: "http://nltk.github.com/nltk_data/packages/corpora/stopwords.zip" }
+  - { path: "corpora/wordnet",
+      url: "http://nltk.github.com/nltk_data/packages/corpora/wordnet.zip" }

playbooks/roles/nltk/tasks/main.yml

+---
+
+- name: create the nltk data directory and subdirectories
+  file: path={{ NLTK_DATA_DIR }}/{{ item.path|dirname }} state=directory
+  with_items: NLTK_DATA
+  tags:
+    - deploy
+
+- name: download nltk data
+  get_url: >
+    dest={{ NLTK_DATA_DIR }}/{{ item.url|basename }}
+    url={{ item.url }}
+  with_items: NLTK_DATA
+  register: nltk_download
+  tags:
+    - deploy
+
+- name: unarchive nltk data
+  unarchive: >
+    src={{ NLTK_DATA_DIR }}/{{ item.url|basename }}
+    dest={{ NLTK_DATA_DIR }}/{{ item.path|dirname }}
+  with_items: NLTK_DATA
+  when: nltk_download|changed
+  tags:
+    - deploy

playbooks/roles/ora/defaults/main.yml

@@ -11,10 +11,6 @@ ora_venv_dir: "{{ ora_venvs_dir }}/ora"
 ora_venv_bin: "{{ ora_venv_dir }}/bin"
 ora_user: "ora"
 ora_deploy_path: "{{ ora_venv_bin }}:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin"
-ora_nltk_data_dir: "{{ ora_data_dir}}/nltk_data"
-
-ora_nltk_download_url: http://edx-static.s3.amazonaws.com/nltk/nltk-data-20131113.tar.gz
-ora_nltk_tmp_file: "{{ ora_data_dir }}/nltk.tmp.tar.tz"
 
 ora_source_repo: https://github.com/edx/edx-ora.git
 ora_version: 'master'
@@ -143,7 +139,6 @@ ora_auth_config:
 
 ora_environment:
   SERVICE_VARIANT: ora
-  NLTK_DATA: $ora_nltk_data_dir
   LANG: $ORA_LANG
   PATH: $ora_deploy_path
 

playbooks/roles/ora/meta/main.yml

 ---
 dependencies:
   - supervisor
+  - nltk

playbooks/roles/ora/tasks/ease.yml

@@ -43,17 +43,3 @@
   notify:
     - restart ora
     - restart ora_celery
-
-- name: download and install nltk
-  shell: |
-      set -e
-      curl -o {{ ora_nltk_tmp_file }} {{ ora_nltk_download_url }}
-      tar zxf {{ ora_nltk_tmp_file }}
-      rm -f {{ ora_nltk_tmp_file }}
-      touch {{ ora_nltk_download_url|basename }}-installed
-        creates={{ ora_data_dir }}/{{ ora_nltk_download_url|basename }}-installed
-        chdir={{ ora_data_dir }}
-  sudo_user: "{{ common_web_user }}"
-  notify:
-    - restart ora
-    - restart ora_celery

playbooks/roles/ora/templates/ora.conf.j2

@@ -5,7 +5,7 @@ command={{ ora_venv_bin }}/gunicorn --preload -b {{ ora_gunicorn_host }}:{{ ora_
 user={{ common_web_user }}
 directory={{ ora_code_dir }}
 
-environment=PID=/var/run/gunicorn/edx-ora.pid,WORKERS={{ ora_gunicorn_workers }},PORT={{ ora_gunicorn_port }},ADDRESS={{ ora_gunicorn_host }},LANG={{ ORA_LANG }},DJANGO_SETTINGS_MODULE=edx_ora.aws,SERVICE_VARIANT=ora,NLTK_DATA={{ ora_nltk_data_dir }}
+environment=PID=/var/run/gunicorn/edx-ora.pid,WORKERS={{ ora_gunicorn_workers }},PORT={{ ora_gunicorn_port }},ADDRESS={{ ora_gunicorn_host }},LANG={{ ORA_LANG }},DJANGO_SETTINGS_MODULE=edx_ora.aws,SERVICE_VARIANT=ora
 
 stdout_logfile={{ supervisor_log_dir }}/%(program_name)-stdout.log
 stderr_logfile={{ supervisor_log_dir }}/%(program_name)-stderr.log