From addbb0edd29c948ffebfc7f1636b975c0e39b6e7 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Sandro=20J=C3=A4ckel?= <sandro.jaeckel@gmail.com>
Date: Mon, 26 Aug 2024 00:09:02 +0200
Subject: [PATCH] python312Packages.llama-index-core: use nltk-data from nix
 package

This also doesn't fetch the tip of branch which is not reproducible
---
 .../llama-index-core/default.nix              | 20 ++++---------------
 1 file changed, 4 insertions(+), 16 deletions(-)

diff --git a/pkgs/development/python-modules/llama-index-core/default.nix b/pkgs/development/python-modules/llama-index-core/default.nix
index f6dab9b1f0fb..85ad5d57d8f6 100644
--- a/pkgs/development/python-modules/llama-index-core/default.nix
+++ b/pkgs/development/python-modules/llama-index-core/default.nix
@@ -6,13 +6,13 @@
   deprecated,
   dirtyjson,
   fetchFromGitHub,
-  fetchzip,
   fsspec,
   jsonpath-ng,
   llamaindex-py-client,
   nest-asyncio,
   networkx,
   nltk,
+  nltk-data,
   numpy,
   openai,
   pandas,
@@ -33,18 +33,6 @@
   typing-inspect,
 }:
 
-let
-  stopwords = fetchzip {
-    url = "https://raw.githubusercontent.com/nltk/nltk_data/gh-pages/packages/corpora/stopwords.zip";
-    hash = "sha256-tX1CMxSvFjr0nnLxbbycaX/IBnzHFxljMZceX5zElPY=";
-  };
-
-  punkt = fetchzip {
-    url = "https://raw.githubusercontent.com/nltk/nltk_data/gh-pages/packages/tokenizers/punkt.zip";
-    hash = "sha256-SKZu26K17qMUg7iCFZey0GTECUZ+sTTrF/pqeEgJCos=";
-  };
-in
-
 buildPythonPackage rec {
   pname = "llama-index-core";
   version = "0.10.60";
@@ -67,12 +55,12 @@ buildPythonPackage rec {
   # Setting `NLTK_DATA` to a writable path can also solve this problem, but it needs to be done in
   # every package that depends on `llama-index-core` for `pythonImportsCheck` not to fail, so this
   # solution seems more elegant.
-  patchPhase = ''
+  postPatch = ''
     mkdir -p llama_index/core/_static/nltk_cache/corpora/stopwords/
-    cp -r ${stopwords}/* llama_index/core/_static/nltk_cache/corpora/stopwords/
+    cp -r ${nltk-data.stopwords}/corpora/stopwords/* llama_index/core/_static/nltk_cache/corpora/stopwords/
 
     mkdir -p llama_index/core/_static/nltk_cache/tokenizers/punkt/
-    cp -r ${punkt}/* llama_index/core/_static/nltk_cache/tokenizers/punkt/
+    cp -r ${nltk-data.punkt}/tokenizers/punkt/* llama_index/core/_static/nltk_cache/tokenizers/punkt/
   '';
 
   build-system = [ poetry-core ];