Commit ebdfdae1 authored by Martin Puppe's avatar Martin Puppe
Browse files

nixos/paperless: download NLTK data 

Since version 1.10.0 paperless-ngx depends on the NLTK library which is
used to pre-process data for machine learning. NLTK needs certain
data for stemming, stopword removal etc. This data has to be downloaded
first. This commit introduces a new systemd service that does the
downloading.
parent fa7fbe56

nixos/modules/services/misc/paperless.nix

+30 −0
Original line number Diff line number Diff line
@@ -6,6 +6,7 @@ let 
  pkg = cfg.package;



  defaultUser = "paperless";

  nltkDir = "/var/cache/paperless/nltk";



  # Don't start a redis instance if the user sets a custom redis connection

  enableRedis = !hasAttr "PAPERLESS_REDIS" cfg.extraConfig;
@@ -15,6 +16,7 @@ let 
    PAPERLESS_DATA_DIR = cfg.dataDir;

    PAPERLESS_MEDIA_ROOT = cfg.mediaDir;

    PAPERLESS_CONSUMPTION_DIR = cfg.consumptionDir;

    PAPERLESS_NLTK_DIR = nltkDir;

    GUNICORN_CMD_ARGS = "--bind=${cfg.address}:${toString cfg.port}";

  } // optionalAttrs (config.time.timeZone != null) {

    PAPERLESS_TIME_ZONE = config.time.timeZone;
@@ -49,6 +51,7 @@ let 
      cfg.dataDir

      cfg.mediaDir

    ];

    CacheDirectory = "paperless";

    CapabilityBoundingSet = "";

    # ProtectClock adds DeviceAllow=char-rtc r

    DeviceAllow = "";
@@ -293,6 +296,33 @@ in 
      };

    };



    # Download NLTK corpus data

    systemd.services.paperless-download-nltk-data = {

      wantedBy = [ "paperless-scheduler.service" ];

      before = [ "paperless-scheduler.service" ];

      after = [ "network-online.target" ];

      serviceConfig = defaultServiceConfig // {

        User = cfg.user;

        Type = "oneshot";

        # Enable internet access

        PrivateNetwork = false;

        # Restrict write access

        BindPaths = [];

        BindReadOnlyPaths = [

          "/nix/store"

          "-/etc/resolv.conf"

          "-/etc/nsswitch.conf"

          "-/etc/ssl/certs"

          "-/etc/static/ssl/certs"

          "-/etc/hosts"

          "-/etc/localtime"

        ];

        ExecStart = let pythonWithNltk = pkg.python.withPackages (ps: [ ps.nltk ]); in ''

          ${pythonWithNltk}/bin/python -m nltk.downloader -d '${nltkDir}' punkt snowball_data stopwords

        '';

      };

    };



    systemd.services.paperless-consumer = {

      description = "Paperless document consumer";

      # Bind to `paperless-scheduler` so that the consumer never runs