paperless-ngx/ansible/tasks/main.yml

---
- name: verify operating system
  fail:
    msg: Sorry, only Debian and Ubuntu supported at the moment.
  when: not(ansible_distribution == 'Debian' or ansible_distribution == 'Ubuntu')

- name: install base dependencies
  apt:
    update_cache: yes
    pkg:
      # paperless-ng
      - python3-pip
      - python3-dev
      - fonts-liberation
      - imagemagick
      - optipng
      - gnupg
      - libpq-dev
      - libmagic-dev
      - mime-support
      # OCRmyPDF
      - unpaper
      - ghostscript
      - icc-profiles-free
      - qpdf
      - liblept5
      - libxml2
      - pngquant
      - zlib1g
      - tesseract-ocr
      # dev
      - sudo
      - build-essential
      - python3-setuptools
      - python3-wheel

# upstream virtualenv in Ubuntu 20.04 is broken
# https://github.com/pypa/virtualenv/issues/1873
- name: install python virtualenv
  pip:
    name: virtualenv
    extra_args: --upgrade

- name: install ocr languages
  apt:
    pkg: "{{ paperlessng_ocr_languages | map('regex_replace', '^(.*)$', 'tesseract-ocr-\\1') | map('replace', '_', '-') | list }}"

- name: set up notesalexp repository key (for jbig2enc)
  apt_key:
    url: https://notesalexp.org/debian/alexp_key.asc
    state: present
  when: paperlessng_use_jbig2enc

- name: set up notesalexp repository (for jbig2enc)
  apt_repository:
    repo: "deb https://notesalexp.org/debian/{{ ansible_distribution_release }}/ {{ ansible_distribution_release }} main"
    state: present
  when: paperlessng_use_jbig2enc

- name: set up notesalexp repository pinning
  copy:
    content: |
      Package: *
      Pin: release o=notesalexp.org
      Pin-Priority: 1

      Package: jbig2enc
      Pin: release o=notesalexp.org
      Pin-Priority: 500
    dest: /etc/apt/preferences.d/notesalexp
  when: paperlessng_use_jbig2enc

- name: install jbig2enc
  apt:
    pkg: jbig2enc
    update_cache: yes
  when: paperlessng_use_jbig2enc

- name: install redis
  apt:
    pkg: redis-server
  when: paperlessng_redis_host == 'localhost' or paperlessng_redis_host == '127.0.0.1'

- name: enable redis
  systemd:
    name: redis-server
    enabled: yes
    masked: no
    state: started
  when: paperlessng_redis_host == 'localhost' or paperlessng_redis_host == '127.0.0.1'

- name: create paperless system group
  group:
    name: "{{ paperlessng_system_group }}"

- name: create paperless system user
  user:
    name: "{{ paperlessng_system_user }}"
    groups:
      - "{{ paperlessng_system_group }}"
    shell: /usr/sbin/nologin
    # GNUPG_HOME required due to paperless db.py
    create_home: yes

- block:
    - name: get latest release version
      uri:
        url: https://api.github.com/repos/jonaswinkler/paperless-ng/releases/latest
        method: GET
      register: latest_release
    - name: parse latest release version
      set_fact:
        paperlessng_version: "{{ latest_release.json['tag_name'] }}"
  when: paperlessng_version == "latest"

- name: check if version is ref
  fail:
    msg: "Specifying `paperlessng_version` as git ref may not work as expected!"
  ignore_errors: True  # Output failure (as warning), but don't consider play failed
  when: paperlessng_version.startswith('refs/')

- block:
    - name: sanitize version string
      set_fact:
        paperlessng_version: "{{ paperlessng_version | regex_replace('^ng-(\\d+\\.\\d+\\.\\d+)$', '\\1') }}"
    - name: get tag data
      uri:
        url: https://api.github.com/repos/jonaswinkler/paperless-ng/tags
        method: GET
      register: tags
    - name: get commit for target tag
      set_fact:
        paperlessng_commit: "{{ tags.json | json_query('[?name==`ng-' + paperlessng_version +'`] | [0].commit.sha') }}"
  when: paperlessng_version | regex_search("^(ng-)?(\d+\.\d+\.\d+)$")

- block:
    - name: check if version is branch
      uri:
        url: "https://api.github.com/repos/jonaswinkler/paperless-ng/branches/{{ paperlessng_version }}"
        method: GET
        status_code: [200, 404]
      register: branch
    - name: get commit for target branch
      set_fact:
        paperlessng_commit: "{{ branch.json | json_query('commit.sha') }}"
      when: branch.status == 200
    - block:
        - name: check if version is commit-or-ref
          uri:
            url: "https://api.github.com/repos/jonaswinkler/paperless-ng/commits/{{ paperlessng_version }}"
            method: GET
            status_code: [200, 404, 422]
          register: commit
        - name: get commit for target commit-or-ref
          set_fact:
            paperlessng_commit: "{{ commit.json | json_query('sha') }}"
          when: commit.status == 200
        - name: fail
          fail:
            msg: "Can not determine commit from `paperlessng_version=={{ paperlessng_version }}`!"
          when: commit.status != 200
      when: branch.status == 404
  when: not(paperlessng_version | regex_search("^(ng-)?(\d+\.\d+\.\d+)$"))

- name: check for paperless-ng installation
  command:
    cmd: "cat {{ paperlessng_directory }}/.installed_version"
  changed_when: '"No such file or directory" in paperlessng_current_commit.stderr or paperlessng_current_commit.stdout != paperlessng_commit | string'
  failed_when: false
  ignore_errors: yes
  register: paperlessng_current_commit

- name: register current state
  set_fact:
    fresh_installation: '{{ "No such file or directory" in paperlessng_current_commit.stderr }}'
    update_installation: '{{ "No such file or directory" not in paperlessng_current_commit.stderr and paperlessng_current_commit.stdout != paperlessng_commit | string }}'
    reconfigure_only: "{{ paperlessng_current_commit.stdout == paperlessng_commit | string }}"

- block:
    - name: backup current paperless-ng installation
      copy:
        src: "{{ paperlessng_directory }}"
        remote_src: yes
        dest: "{{ paperlessng_directory }}-{{ ansible_date_time.iso8601 }}/"
    - name: remove current paperless sources
      file:
        path: "{{ paperlessng_directory }}/{{ item }}"
        state: absent
      with_items:
        - docker
        - docs
        - scripts
        - src
        - static
  when: update_installation

- block:
    - name: create paperless-ng directory and set permissions
      file:
        path: "{{ paperlessng_directory }}"
        state: directory
        owner: "{{ paperlessng_system_user }}"
        group: "{{ paperlessng_system_group }}"
        mode: "750"
    - name: create temporary directory
      become: yes
      become_user: "{{ paperlessng_system_user }}"
      tempfile:
        state: directory
        path: "{{ paperlessng_directory }}"
      register: tempdir
    - name: check if version is available as release archive
      uri:
        url: "https://github.com/jonaswinkler/paperless-ng/releases/download/ng-{{ paperlessng_version }}/paperless-ng-{{ paperlessng_version }}.tar.xz"
        method: GET
        status_code: [200, 404]
      register: release_archive
    - name: install paperless-ng from source
      include_tasks: install-source.yml
      when: release_archive.status == 404
    - name: install paperless-ng from release archive
      include_tasks: install-release.yml
      when: release_archive.status == 200
    - name: change owner and permissions of paperless-ng
      command:
        cmd: "{{ item }}"
        warn: false
      with_items:
        - "chown -R {{ paperlessng_system_user }}:{{ paperlessng_system_group }} {{ tempdir.path }}"
        - "find {{ tempdir.path }} -type d -exec chmod 0750 {} ;"
        - "find {{ tempdir.path }} -type f -exec chmod 0640 {} ;"
    - name: move paperless-ng
      command:
        cmd: "cp -a {{ tempdir.path }}/paperless-ng/. {{ paperlessng_directory }}"
    - name: store commit hash of installed version
      copy:
        content: "{{ paperlessng_commit }}"
        dest: "{{ paperlessng_directory }}/.installed_version"
        owner: "{{ paperlessng_system_user }}"
        group: "{{ paperlessng_system_group }}"
        mode: "0440"
    - name: remove temporary directory
      file:
        path: "{{ tempdir.path }}"
        state: absent
  when: not reconfigure_only

- name: create paperless-ng directories and set permissions
  file:
    path: "{{ item }}"
    state: directory
    owner: "{{ paperlessng_system_user }}"
    group: "{{ paperlessng_system_group }}"
    mode: "750"
  with_items:
    - "{{ paperlessng_consumption_dir }}"
    - "{{ paperlessng_data_dir }}"
    - "{{ paperlessng_media_root }}"
    - "{{ paperlessng_staticdir }}"

- name: rename initial config
  command:
    cmd: "mv -f {{ paperlessng_directory }}/paperless.conf {{ paperlessng_directory }}/paperless.conf.template"
    removes: "{{ paperlessng_directory }}/paperless.conf"

- name: configure paperless-ng
  lineinfile:
    path: "{{ paperlessng_directory }}/paperless.conf.template"
    regexp: "^#?{{ item.regexp }}="
    line: "{{ item.line }}"
  with_items:
    # Required services
    - regexp: PAPERLESS_REDIS
      line: "PAPERLESS_REDIS=redis://{{ paperlessng_redis_host }}:{{ paperlessng_redis_port }}"
    # Paths and folders
    - regexp: PAPERLESS_CONSUMPTION_DIR
      line: "PAPERLESS_CONSUMPTION_DIR={{ paperlessng_consumption_dir }}"
    - regexp: PAPERLESS_DATA_DIR
      line: "PAPERLESS_DATA_DIR={{ paperlessng_data_dir }}"
    - regexp: PAPERLESS_MEDIA_ROOT
      line: "PAPERLESS_MEDIA_ROOT={{ paperlessng_media_root }}"
    - regexp: PAPERLESS_STATICDIR
      line: "PAPERLESS_STATICDIR={{ paperlessng_staticdir }}"
    - regexp: PAPERLESS_FILENAME_FORMAT
      line: "PAPERLESS_FILENAME_FORMAT={{ paperlessng_filename_format }}"
    - regexp: PAPERLESS_LOGGING_DIR
      line: "PAPERLESS_LOGGING_DIR={{ paperlessng_logging_dir }}"
    # Hosting & Security
    - regexp: PAPERLESS_SECRET_KEY
      line: "PAPERLESS_SECRET_KEY={{ paperlessng_secret_key }}"
    - regexp: PAPERLESS_ALLOWED_HOSTS
      line: "PAPERLESS_ALLOWED_HOSTS={{ paperlessng_allowed_hosts }}"
    - regexp: PAPERLESS_CORS_ALLOWED_HOSTS
      line: "PAPERLESS_CORS_ALLOWED_HOSTS={{ paperlessng_cors_allowed_hosts }}"
    - regexp: PAPERLESS_FORCE_SCRIPT_NAME
      line: "PAPERLESS_FORCE_SCRIPT_NAME={{ paperlessng_force_script_name }}"
    - regexp: PAPERLESS_STATIC_URL
      line: "PAPERLESS_STATIC_URL={{ paperlessng_static_url }}"
    - regexp: PAPERLESS_AUTO_LOGIN_USERNAME
      line: "PAPERLESS_AUTO_LOGIN_USERNAME={{ paperlessng_auto_login_username }}"
    - regexp: PAPERLESS_COOKIE_PREFIX
      line: "PAPERLESS_COOKIE_PREFIX={{ paperlessng_cookie_prefix }}"
    - regexp: PAPERLESS_ENABLE_HTTP_REMOTE_USER
      line: "PAPERLESS_ENABLE_HTTP_REMOTE_USER={{ paperlessng_enable_http_remote_user }}"
    # OCR settings
    - regexp: PAPERLESS_OCR_LANGUAGE
      line: "PAPERLESS_OCR_LANGUAGE={{ paperlessng_ocr_languages | join('+') | replace('-','_') }}"
    - regexp: PAPERLESS_OCR_MODE
      line: "PAPERLESS_OCR_MODE={{ paperlessng_ocr_mode }}"
    - regexp: PAPERLESS_OCR_CLEAN
      line: "PAPERLESS_OCR_CLEAN={{ paperlessng_ocr_clean }}"
    - regexp: PAPERLESS_OCR_DESKEW
      line: "PAPERLESS_OCR_DESKEW={{ paperlessng_ocr_deskew }}"
    - regexp: PAPERLESS_OCR_ROTATE_PAGES
      line: "PAPERLESS_OCR_ROTATE_PAGES={{ paperlessng_ocr_rotate_pages }}"
    - regexp: PAPERLESS_OCR_ROTATE_PAGES_THRESHOLD
      line: "PAPERLESS_OCR_ROTATE_PAGES_THRESHOLD={{ paperlessng_ocr_rotate_pages_threshold }}"
    - regexp: PAPERLESS_OCR_OUTPUT_TYPE
      line: "PAPERLESS_OCR_OUTPUT_TYPE={{ paperlessng_ocr_output_type }}"
    - regexp: PAPERLESS_OCR_PAGES
      line: "PAPERLESS_OCR_PAGES={{ paperlessng_ocr_pages }}"
    - regexp: PAPERLESS_OCR_IMAGE_DPI
      line: "PAPERLESS_OCR_IMAGE_DPI={{ paperlessng_ocr_image_dpi }}"
    - regexp: PAPERLESS_OCR_USER_ARGS
      line: "PAPERLESS_OCR_USER_ARGS={{ paperlessng_ocr_user_args | combine({'jbig2_lossy': true} if paperlessng_big2enc_lossy else {}) | to_json }}"
    # Tika settings
    - regexp: PAPERLESS_TIKA_ENABLED
      line: "PAPERLESS_TIKA_ENABLED={{ paperlessng_tika_enabled }}"
    - regexp: PAPERLESS_TIKA_ENDPOINT
      line: "PAPERLESS_TIKA_ENDPOINT={{ paperlessng_tika_endpoint }}"
    - regexp: PAPERLESS_TIKA_GOTENBERG_ENDPOINT
      line: "PAPERLESS_TIKA_GOTENBERG_ENDPOINT={{ paperlessng_tika_gotenberg_endpoint }}"
    # Software tweaks
    - regexp: PAPERLESS_TIME_ZONE
      line: "PAPERLESS_TIME_ZONE={{ paperlessng_time_zone }}"
    - regexp: PAPERLESS_CONSUMER_POLLING
      line: "PAPERLESS_CONSUMER_POLLING={{ paperlessng_consumer_polling }}"
    - regexp: PAPERLESS_CONSUMER_DELETE_DUPLICATES
      line: "PAPERLESS_CONSUMER_DELETE_DUPLICATES={{ paperlessng_consumer_delete_duplicates }}"
    - regexp: PAPERLESS_CONSUMER_RECURSIVE
      line: "PAPERLESS_CONSUMER_RECURSIVE={{ paperlessng_consumer_recursive }}"
    - regexp: PAPERLESS_CONSUMER_SUBDIRS_AS_TAGS
      line: "PAPERLESS_CONSUMER_SUBDIRS_AS_TAGS={{ paperlessng_consumer_subdirs_as_tags }}"
    - regexp: PAPERLESS_CONVERT_MEMORY_LIMIT
      line: "PAPERLESS_CONVERT_MEMORY_LIMIT={{ paperlessng_convert_memory_limit }}"
    - regexp: PAPERLESS_CONVERT_TMPDIR
      line: "PAPERLESS_CONVERT_TMPDIR={{ paperlessng_convert_tmpdir }}"
    - regexp: PAPERLESS_OPTIMIZE_THUMBNAILS
      line: "PAPERLESS_OPTIMIZE_THUMBNAILS={{ paperlessng_optimize_thumbnails }}"
    - regexp: PAPERLESS_PRE_CONSUME_SCRIPT
      line: "PAPERLESS_PRE_CONSUME_SCRIPT={{ paperlessng_pre_consume_script }}"
    - regexp: PAPERLESS_POST_CONSUME_SCRIPT
      line: "PAPERLESS_POST_CONSUME_SCRIPT={{ paperlessng_post_consume_script }}"
    - regexp: PAPERLESS_FILENAME_DATE_ORDER
      line: "PAPERLESS_FILENAME_DATE_ORDER={{ paperlessng_filename_date_order }}"
    - regexp: PAPERLESS_THUMBNAIL_FONT_NAME
      line: "PAPERLESS_THUMBNAIL_FONT_NAME={{ paperlessng_thumbnail_font_name }}"
    - regexp: PAPERLESS_IGNORE_DATES
      line: "PAPERLESS_IGNORE_DATES={{ paperlessng_ignore_dates }}"
  no_log: yes

- name: configure paperless-ng database [sqlite]
  lineinfile:
    path: "{{ paperlessng_directory }}/paperless.conf.template"
    regexp: "^#?PAPERLESS_DBHOST=(.*)$"
    line: '#PAPERLESS_DBHOST=\1'
    backrefs: yes
  when: paperlessng_db_type == 'sqlite'

- name: configure paperless-ng database [postgresql]
  lineinfile:
    path: "{{ paperlessng_directory }}/paperless.conf.template"
    regexp: "^#?{{ item.regexp }}="
    line: "{{ item.line }}"
  with_items:
    - regexp: PAPERLESS_DBHOST
      line: "PAPERLESS_DBHOST={{ paperlessng_db_host }}"
    - regexp: PAPERLESS_DBPORT
      line: "PAPERLESS_DBPORT={{ paperlessng_db_port }}"
    - regexp: PAPERLESS_DBNAME
      line: "PAPERLESS_DBNAME={{ paperlessng_db_name }}"
    - regexp: PAPERLESS_DBUSER
      line: "PAPERLESS_DBUSER={{ paperlessng_db_user }}"
    - regexp: PAPERLESS_DBPASS
      line: "PAPERLESS_DBPASS={{ paperlessng_db_pass }}"
    - regexp: PAPERLESS_DBSSLMODE
      line: "PAPERLESS_DBSSLMODE={{ paperlessng_db_sslmode }}"
  when: paperlessng_db_type == 'postgresql'
  no_log: yes

- name: deploy paperless-ng configuration
  copy:
    src: "{{ paperlessng_directory }}/paperless.conf.template"
    remote_src: yes
    dest: /etc/paperless.conf
    owner: root
    group: root
    mode: "0644"
  register: configuration

- name: create paperlessng venv
  become: yes
  become_user: "{{ paperlessng_system_user }}"
  command:
    cmd: "python3 -m virtualenv {{ paperlessng_virtualenv }} -p /usr/bin/python3"
    creates: "{{ paperlessng_virtualenv }}"
  register: venv

- block:
    - name: install paperlessng requirements
      become: yes
      become_user: "{{ paperlessng_system_user }}"
      pip:
        requirements: "{{ paperlessng_directory }}/requirements.txt"
        executable: "{{ paperlessng_virtualenv }}/bin/pip3"
        extra_args: --upgrade
    - name: migrate database schema
      become: yes
      become_user: "{{ paperlessng_system_user }}"
      command: "{{ paperlessng_virtualenv }}/bin/python3 {{ paperlessng_directory }}/src/manage.py migrate"
      register: database_schema
      changed_when: '"No migrations to apply." not in database_schema.stdout'
  when: not reconfigure_only

- name: configure paperless superuser
  become: yes
  become_user: "{{ paperlessng_system_user }}"
  # "manage.py createsuperuser" only works on interactive TTYs
  vars:
    creation_script: |
      from django.contrib.auth.models import User
      from django.contrib.auth.hashers import get_hasher

      if User.objects.filter(username='{{ paperlessng_superuser_name }}').exists():
          user = User.objects.get(username='{{ paperlessng_superuser_name }}')
          old = user.__dict__.copy()

          user.is_superuser = True
          user.email = '{{ paperlessng_superuser_email }}'
          user.set_password('{{ paperlessng_superuser_password }}')
          user.save()
          new = user.__dict__

          algorithm, iterations, old_salt, old_hash = old['password'].split('$')
          new_password_old_salt = get_hasher(algorithm).encode(password='{{ paperlessng_superuser_password }}', salt=old_salt, iterations=int(iterations))
          _, _, _, new_hash = new_password_old_salt.split('$')
          if not (old_hash == new_hash and old['is_superuser'] == new['is_superuser'] and old['email'] == new['email']):
              print('changed')
      else:
          User.objects.create_superuser('{{ paperlessng_superuser_name }}', '{{ paperlessng_superuser_email }}', '{{ paperlessng_superuser_password }}')
          print('changed')
  command: '{{ paperlessng_virtualenv }}/bin/python3 {{ paperlessng_directory }}/src/manage.py shell -c "{{ creation_script }}"'
  register: superuser
  changed_when: superuser.stdout == 'changed'
  no_log: yes

- name: set ownership and permissions on paperlessng venv
  file:
    path: "{{ paperlessng_virtualenv }}"
    state: directory
    recurse: yes
    owner: "{{ paperlessng_system_user }}"
    group: "{{ paperlessng_system_group }}"
    mode: g-w,o-rwx
  when: venv.changed or not reconfigure_only

- name: configure ghostscript for PDF
  lineinfile:
    path: /etc/ImageMagick-6/policy.xml
    regexp: '(\s+)<policy domain="coder" rights=".*" pattern="PDF" />'
    line: '\1<policy domain="coder" rights="read|write" pattern="PDF" />'
    backrefs: yes

- name: configure gunicorn web server
  lineinfile:
    path: "{{ paperlessng_directory }}/gunicorn.conf.py"
    regexp: '^bind = '
    line: "bind = '{{ paperlessng_listen_address }}:{{ paperlessng_listen_port }}'"

- name: configure systemd services
  ini_file:
    path: "{{ paperlessng_directory }}/scripts/{{ item[0] }}"
    section: "Service"
    option: "{{ item[1].option  }}"
    value: "{{ item[1].value }}"
  with_nested:
    - [
        paperless-consumer.service,
        paperless-scheduler.service,
        paperless-webserver.service,
      ]
    - [
        # https://www.freedesktop.org/software/systemd/man/systemd.exec.html
        { option: "User", value: "{{ paperlessng_system_user }}" },
        { option: "Group", value: "{{ paperlessng_system_group }}" },
        { option: "WorkingDirectory", value: "{{ paperlessng_directory }}/src" },
        { option: "ProtectSystem", value: "full" },
        { option: "NoNewPrivileges", value: "true" },
        { option: "PrivateUsers", value: "true" },
        { option: "PrivateDevices", value: "true" },
      ]

- name: configure paperless-consumer service
  ini_file:
    path: "{{ paperlessng_directory }}/scripts/paperless-consumer.service"
    section: "Service"
    option: "ExecStart"
    value: "{{ paperlessng_virtualenv }}/bin/python3 manage.py document_consumer"

- name: configure paperless-scheduler service
  ini_file:
    path: "{{ paperlessng_directory }}/scripts/paperless-scheduler.service"
    section: "Service"
    option: "ExecStart"
    value: "{{ paperlessng_virtualenv }}/bin/python3 manage.py qcluster"

- name: configure paperless-webserver service
  ini_file:
    path: "{{ paperlessng_directory }}/scripts/paperless-webserver.service"
    section: "Service"
    option: "ExecStart"
    value: "{{ paperlessng_virtualenv }}/bin/gunicorn -c {{ paperlessng_directory }}/gunicorn.conf.py paperless.asgi:application"

- name: copy systemd services
  copy:
    src: "{{ paperlessng_directory }}/scripts/{{ item }}"
    remote_src: yes
    dest: "/etc/systemd/system/{{ item }}"
  with_items:
    - paperless-consumer.service
    - paperless-scheduler.service
    - paperless-webserver.service
  register: paperless_services

- name: reload systemd daemon
  systemd:
    name: "{{ item }}"
    state: restarted
    daemon_reload: yes
  with_items:
    - paperless-consumer
    - paperless-scheduler
    - paperless-webserver
  when: paperless_services.changed or configuration.changed

- name: enable paperlessng services
  systemd:
    name: "{{ item }}"
    enabled: yes
    masked: no
    state: started
  with_items:
    - paperless-consumer
    - paperless-scheduler
    - paperless-webserver