about summary refs log tree commit diff
path: root/corp/russian/data-import/default.nix
diff options
context:
space:
mode:
authorVincent Ambo <mail@tazj.in>2023-01-20T08·54+0300
committertazjin <tazjin@tvl.su>2023-01-21T17·49+0000
commitee0c0ee95103fa10e227a1976149d20e6944001c (patch)
tree9348ea54fc9ccedb01355f80c9d7c4bf55aa4e9c /corp/russian/data-import/default.nix
parent1a84e3e6d800fed1709364174d4a615b1d146ad0 (diff)
chore(corp/data-import): make OR data archive available in env r/5728
Change-Id: Idacf42743051eae0cf7010f952a4f91af17ad708
Reviewed-on: https://cl.tvl.fyi/c/depot/+/7892
Reviewed-by: tazjin <tazjin@tvl.su>
Tested-by: BuildkiteCI
Diffstat (limited to 'corp/russian/data-import/default.nix')
-rw-r--r--corp/russian/data-import/default.nix23
1 files changed, 15 insertions, 8 deletions
diff --git a/corp/russian/data-import/default.nix b/corp/russian/data-import/default.nix
index c2fc1bf1cb5e..cf358874dce6 100644
--- a/corp/russian/data-import/default.nix
+++ b/corp/russian/data-import/default.nix
@@ -9,22 +9,29 @@ let
   # mirrored input data from OpenCorpora, as of 2023-01-17.
   #
   # This data is licensed under CC-BY-SA.
-  inputDataArchive = pkgs.fetchurl {
+  openCorporaArchive = pkgs.fetchurl {
     name = "dict.opcorpora.xml.bz";
-    url = "https://tazj.in/blobs/dict.opcorpora.xml.bz2";
+    url = "https://tazj.in/blobs/opencorpora-20230117.xml.bz2";
     sha256 = "04n5g43fkfc93z6xlwf2qfdrfdfl562pc2ivdb3cbbbsy56gkqg6";
   };
 
-  inputData = pkgs.runCommand "dict.opcorpora.xml" { } ''
-    ${pkgs.bzip2}/bin/bunzip2 -k -c ${inputDataArchive} > $out
+  openCorpora = pkgs.runCommand "dict.opcorpora.xml" { } ''
+    ${pkgs.bzip2}/bin/bunzip2 -k -c ${openCorporaArchive} > $out
   '';
 
+  openRussianArchive = pkgs.fetchzip {
+    name = "openrussian-20230117";
+    url = "https://tazj.in/blobs/openrussian-20230117.tar.xz";
+    sha256 = "06jl7i23cx58a0n2626hb82xlzimixvnxp7lxdw0g664kv9bmw25";
+  };
+
   # development shell with native deps
   shell = pkgs.mkShell {
     inherit buildInputs;
 
-    # make OPENCORPORA_DATA available in the environment
-    OPENCORPORA_DATA = inputData;
+    # make datasets available in the environment
+    OPENCORPORA_DATA = openCorpora;
+    OPENRUSSIAN_DATA = openRussianArchive;
   };
 
 in
@@ -33,11 +40,11 @@ lib.fix (self: depot.third_party.naersk.buildPackage {
   inherit buildInputs;
 
   passthru = depot.nix.readTree.drvTargets {
-    inherit shell inputData;
+    inherit shell openCorpora;
 
     # target that actually builds an entire database
     database = pkgs.runCommand "tvl-russian-db.sqlite" { } ''
-      ${self}/bin/data-import ${inputData} $out
+      ${self}/bin/data-import ${openCorpora} $out
     '';
   };
 })