From b96dd66941fa4c4f995eb0e57ac9e2771d513b88 Mon Sep 17 00:00:00 2001
From: Zhaofeng Li <hello@zhaofeng.li>
Date: Sun, 6 Jul 2025 11:16:32 -0600
Subject: [PATCH] cdxj-indexer: init at 1.4.6

---
 .../python-modules/cdxj-indexer/default.nix   | 57 +++++++++++++++++++
 pkgs/top-level/all-packages.nix               |  2 +
 pkgs/top-level/python-packages.nix            |  2 +
 3 files changed, 61 insertions(+)
 create mode 100644 pkgs/development/python-modules/cdxj-indexer/default.nix

diff --git a/pkgs/development/python-modules/cdxj-indexer/default.nix b/pkgs/development/python-modules/cdxj-indexer/default.nix
new file mode 100644
index 000000000000..f0a28354411d
--- /dev/null
+++ b/pkgs/development/python-modules/cdxj-indexer/default.nix
@@ -0,0 +1,57 @@
+{
+  lib,
+  buildPythonPackage,
+  fetchFromGitHub,
+  setuptools,
+  warcio,
+  surt,
+  py3amf,
+  multipart,
+  pytestCheckHook,
+}:
+
+buildPythonPackage rec {
+  pname = "cdxj-indexer";
+  version = "1.4.6";
+  pyproject = true;
+
+  src = fetchFromGitHub {
+    owner = "webrecorder";
+    repo = "cdxj-indexer";
+    tag = "v${version}";
+    hash = "sha256-E3b/IfjngyXhWvRYP9CkQGvBFeC8pAm4KxZA9MwOo4s=";
+  };
+
+  build-system = [
+    setuptools
+  ];
+
+  dependencies = [
+    warcio
+    surt
+    py3amf
+    multipart
+  ];
+
+  pythonRemoveDeps = [
+    # Transitive dependency that does not need to be pinned
+    # Proposed fix in <https://github.com/webrecorder/cdxj-indexer/pull/25>
+    "idna"
+  ];
+
+  nativeCheckInputs = [
+    pytestCheckHook
+  ];
+
+  pythonImportsCheck = [
+    "cdxj_indexer"
+  ];
+
+  meta = {
+    description = "CDXJ Indexing of WARC/ARCs";
+    homepage = "https://github.com/webrecorder/cdxj-indexer";
+    license = lib.licenses.asl20;
+    maintainers = with lib.maintainers; [ zhaofengli ];
+    mainProgram = "cdxj-indexer";
+  };
+}
diff --git a/pkgs/top-level/all-packages.nix b/pkgs/top-level/all-packages.nix
index 03fb17ebc456..1a886e62864f 100644
--- a/pkgs/top-level/all-packages.nix
+++ b/pkgs/top-level/all-packages.nix
@@ -11949,6 +11949,8 @@ with pkgs;
 
   cdparanoia = cdparanoiaIII;
 
+  cdxj-indexer = with python3Packages; toPythonApplication cdxj-indexer;
+
   chromium = callPackage ../applications/networking/browsers/chromium (config.chromium or { });
 
   chuck = callPackage ../applications/audio/chuck {
diff --git a/pkgs/top-level/python-packages.nix b/pkgs/top-level/python-packages.nix
index aa01cb5e833d..27348052e871 100644
--- a/pkgs/top-level/python-packages.nix
+++ b/pkgs/top-level/python-packages.nix
@@ -2332,6 +2332,8 @@ self: super: with self; {
 
   cddlparser = callPackage ../development/python-modules/cddlparser { };
 
+  cdxj-indexer = callPackage ../development/python-modules/cdxj-indexer { };
+
   celery = callPackage ../development/python-modules/celery { };
 
   celery-batches = callPackage ../development/python-modules/celery-batches { };