Accepting request 1160467 from home:bnavigator:branches:devel:languages:python

- Update to 3.8.1
  * Resolve RCE & XSS vulnerabilities in localhost WordNet Browser
  * Add Python 3.11 support
- Update nltk_data archive
- Drop port-2to3.patch
- Add nltk-pr3207-py312.patch for Python 3.12 support
  * gh#nltk/nltk#3207

OBS-URL: https://build.opensuse.org/request/show/1160467
OBS-URL: https://build.opensuse.org/package/show/devel:languages:python/python-nltk?expand=0&rev=45

nltk-3.8.1.zip

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1834da3d0682cba4f2cede2f9aad6b0fafb6461ba451db0efb6f9c39798d64d3
+size 4620388

nltk-3.8.zip

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:74b30826a37d78d53427105bbd037dd880251be269fca64ee530838a46ed55fc
-size 4619825

nltk_data.tar.xz

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:f41383a4774bf7227f4563f46543460ba07a6921f7bcc6185519e87ea9e4323f
+oid sha256:f79462ac99f414b4850943720bed4a59c1bb15bfc8f1ce16b26165da6db07680
-size 453871052
+size 393271816

port-2to3.patch

@@ -1,48 +0,0 @@
----
- nltk_data/corpora/pl196x/splitter.py                 |    4 ++--
- nltk_data/taggers/universal_tagset/universal_tags.py |    5 -----
- tools/find_deprecated.py                             |    2 +-
- 3 files changed, 3 insertions(+), 8 deletions(-)
-
---- a/nltk_data/corpora/pl196x/splitter.py
-+++ b/nltk_data/corpora/pl196x/splitter.py
-@@ -1,4 +1,4 @@
--#!/usr/bin/python
-+#!/usr/bin/python3
- 
- import sys, re
- 
-@@ -7,7 +7,7 @@ TEXTID = re.compile(r'<text id="(.*)">')
- 
- if __name__ == '__main__':
-     if len(sys.argv) != 2:
--        print 'One argument required: a pl196x corpus to split.'
-+        print('One argument required: a pl196x corpus to split.')
-         sys.exit()
- 
-     inputFileName = sys.argv[1]
---- a/nltk_data/taggers/universal_tagset/universal_tags.py
-+++ b/nltk_data/taggers/universal_tagset/universal_tags.py
-@@ -22,11 +22,6 @@ X - other: foreign words, typos, abbrevi
- @author: Nathan Schneider (nschneid)
- @since: 2011-05-06
- '''
--
--# Strive towards Python 3 compatibility
--from __future__ import print_function, unicode_literals, division
--from future_builtins import map, filter
--
- import re, glob
- from collections import defaultdict
- 
---- a/tools/find_deprecated.py
-+++ b/tools/find_deprecated.py
-@@ -29,7 +29,7 @@ import textwrap
- import tokenize
- from doctest import DocTestParser, register_optionflag
- 
--from cStringIO import StringIO
-+from io import StringIO
- 
- import nltk.corpus
- from nltk import defaultdict

python-nltk.changes

@@ -1,3 +1,14 @@
+-------------------------------------------------------------------
+Thu Mar 21 17:41:52 UTC 2024 - Ben Greiner <code@bnavigator.de>
+
+- Update to 3.8.1
+  * Resolve RCE & XSS vulnerabilities in localhost WordNet Browser
+  * Add Python 3.11 support
+- Update nltk_data archive
+- Drop port-2to3.patch
+- Add nltk-pr3207-py312.patch for Python 3.12 support
+  * gh#nltk/nltk#3207
+
 -------------------------------------------------------------------
 Tue Mar 28 08:36:04 UTC 2023 - pgajdos@suse.com
 

python-nltk.rpmlintrc

@@ -1 +1,2 @@
-addFilter("E: zero-length /usr/lib/python3\.\d/site-packages/nltk/tbl/api\.py")
+addFilter("E: zero-length /usr/lib/python3\.\d+/site-packages/nltk/tbl/api\.py")
+addFilter("explicit-lib-dependency python3\d*-joblib")

python-nltk.spec

@@ -1,7 +1,7 @@
 #
 # spec file for package python-nltk
 #
-# Copyright (c) 2023 SUSE LLC
+# Copyright (c) 2024 SUSE LLC
 #
 # All modifications and additions to the file contributed by third parties
 # remain the property of their copyright owners, unless otherwise agreed
@@ -16,55 +16,86 @@
 #
 
 
-%{!?python_module:%define python_module() python-%{**} python3-%{**}}
-%define pyname nltk
-%define skip_python2 1
 Name:           python-nltk
-Version:        3.8
+Version:        3.8.1
 Release:        0
 Summary:        Natural Language Toolkit
 License:        Apache-2.0
 URL:            http://nltk.org/
-Source0:        https://files.pythonhosted.org/packages/source/n/nltk/%{pyname}-%{version}.zip
+# SourceRepository: https://github.com/nltk/nltk
-# Downloaded NLTK data via python3 -m nltk.downloader,
+Source0:        https://files.pythonhosted.org/packages/source/n/nltk/nltk-%{version}.zip
-# then unzip downloaded zip archive.
+# Download/Update NLTK data:
+#     quilt setup python-nltk.spec
+#     pushd nltk-?.?.?
+#     python3 -m nltk.downloader -d nltk_data tests \
+#          averaged_perceptron_tagger_ru \
+#          brown \
+#          cess_cat \
+#          cess_esp \
+#          conll2007 \
+#          floresta \
+#          gutenberg \
+#          inaugural \
+#          indian \
+#          large_grammars \
+#          nombank.1.0 \
+#          omw-1.4 \
+#          pl196x \
+#          ptb \
+#          punkt \
+#          rte \
+#          sinica_treebank \
+#          stopwords \
+#          treebank \
+#          udhr \
+#          universal_tagset \
+#          wordnet \
+#          wordnet_ic \
+#          words
+#     tar -cJf ../nltk_data.tar.xz nltk_data
+#     popd
 # see https://www.nltk.org/data.html for more details
 Source1:        nltk_data.tar.xz
 Source99:       python-nltk.rpmlintrc
 # PATCH-FIX-UPSTREAM skip-networked-test.patch gh#nltk/nltk#2969 mcepl@suse.com
 # skip tests requiring network connection
 Patch0:         skip-networked-test.patch
-# PATCH-FIX-UPSTREAM port-2to3.patch bsc#[0-9]+ mcepl@suse.com
+# PATCH-FIX-UPSTREAM nltk-pr3207-py312.patch gh#nltk/nltk#3207
-# port scripts in nltk_data to Python 3
+Patch1:         nltk-pr3207-py312.patch
-Patch1:         port-2to3.patch
+BuildRequires:  %{python_module base >= 3.7}
-BuildRequires:  %{python_module regex}
+BuildRequires:  %{python_module pip}
 BuildRequires:  %{python_module setuptools}
+BuildRequires:  %{python_module wheel}
 BuildRequires:  %{pythons}
 BuildRequires:  fdupes
 BuildRequires:  python-rpm-macros
 BuildRequires:  unzip
-# For testing
+# SECTION runtime
-BuildRequires:  %{python_module tk}
+BuildRequires:  %{python_module regex >= 2021.8.3}
 BuildRequires:  %{python_module click}
-BuildRequires:  %{python_module pytest}
-# BuildRequires:  %%{python_module gensim}
 BuildRequires:  %{python_module joblib}
+BuildRequires:  %{python_module tqdm}
+# /SECTION
+# SECTION test
+BuildRequires:  %{python_module tk}
 BuildRequires:  %{python_module Jinja2}
 BuildRequires:  %{python_module matplotlib}
 BuildRequires:  %{python_module numpy}
 BuildRequires:  %{python_module pyparsing}
 BuildRequires:  %{python_module pytest-cov}
 BuildRequires:  %{python_module pytest-mock}
+BuildRequires:  %{python_module pytest}
 BuildRequires:  %{python_module python-crfsuite}
-BuildRequires:  %{python_module regex}
 BuildRequires:  %{python_module requests}
 BuildRequires:  %{python_module scikit-learn}
 BuildRequires:  %{python_module scipy}
 BuildRequires:  %{python_module text-unidecode}
-BuildRequires:  %{python_module tqdm}
 BuildRequires:  %{python_module twython}
-#
+# /SECTION
-Requires:       python-regex
+Requires:       python-regex >= 2021.8.3
+Requires:       python-click
+Requires:       python-joblib
+Requires:       python-tqdm
 Recommends:     python-gensim
 Recommends:     python-matplotlib
 Recommends:     python-numpy
@@ -87,10 +118,7 @@ Python modules, data sets and tutorials supporting research and
 development in Natural Language Processing.
 
 %prep
-%autosetup -p1 -a1 -n %{pyname}-%{version}
+%autosetup -p1 -a1 -n nltk-%{version}
-
-# Remove obsolete scripts
-rm tools/nltk_term_index.py tools/run_doctests.py nltk_data/corpora/semcor/semcor.py
 
 # Fix EOL
 sed -i 's/\r/\n/g; s/\n$//' \
@@ -120,14 +148,13 @@ sed -E -i "s|#![[:space:]]*%{_bindir}/env python|#!%{_bindir}/python3|" \
     setup.py \
     tools/global_replace.py \
     nltk_data/corpora/pl196x/splitter.py \
-    tools/find_deprecated.py \
+    tools/find_deprecated.py
-    tools/svnmime.py
 
 %build
-%python_build
+%pyproject_wheel
 
 %install
-%python_install
+%pyproject_install
 %python_clone -a %{buildroot}%{_bindir}/nltk
 
 %{python_expand %fdupes %{buildroot}%{$python_sitelib}/
@@ -148,8 +175,8 @@ export NLTK_DATA=$(readlink -f ./nltk_data/)
 %files %{python_files}
 %doc README.md
 %license LICENSE.txt
-%{python_sitelib}/%{pyname}/
+%{python_sitelib}/nltk/
-%{python_sitelib}/%{pyname}-%{version}-py%{python_version}.egg-info/
+%{python_sitelib}/nltk-%{version}.dist-info/
 %python_alternative %{_bindir}/nltk
 
 %changelog