Accepting request 1121649 from devel:languages:python

OBS-URL: https://build.opensuse.org/request/show/1121649
OBS-URL: https://build.opensuse.org/package/show/openSUSE:Factory/python-identify?expand=0&rev=10

0001-use-editdistance-not-ukkonen.patch

@@ -0,0 +1,61 @@
+From: Matthias Fehring <buschmann23@opensuse.org>
+Date: 2023-11-01 09:44:00 +0100
+Subject: Use editdistance instead of ukkonen
+Upstream: never
+
+Upstream switched from editdistance_s to ukkonen that is currently
+not packaged for openSUSE.
+
+---
+ identify/identify.py |   10 ++++------
+ setup.cfg            |    2 +-
+ 2 files changed, 5 insertions(+), 7 deletions(-)
+
+--- a/identify/identify.py	2023-10-28 19:19:41.000000000 +0200
++++ b/identify/identify.py	2023-11-01 09:39:16.942021416 +0100
+@@ -243,7 +243,7 @@
+     3. check exact text match with existing licenses
+     4. failing that use edit distance
+     """
+-    import ukkonen  # `pip install identify[license]`
++    import editdistance  # `pip install identify[license]`
+ 
+     with open(filename, encoding='UTF-8') as f:
+         contents = f.read()
+@@ -253,8 +253,6 @@
+     min_edit_dist = sys.maxsize
+     min_edit_dist_spdx = ''
+ 
+-    cutoff = math.ceil(.05 * len(norm))
+-
+     # try exact matches
+     for spdx, text in licenses.LICENSES:
+         norm_license = _norm_license(text)
+@@ -265,13 +263,13 @@
+         if norm and abs(len(norm) - len(norm_license)) / len(norm) > .05:
+             continue
+ 
+-        edit_dist = ukkonen.distance(norm, norm_license, cutoff)
+-        if edit_dist < cutoff and edit_dist < min_edit_dist:
++        edit_dist = editdistance.distance(norm, norm_license)
++        if edit_dist < min_edit_dist:
+             min_edit_dist = edit_dist
+             min_edit_dist_spdx = spdx
+ 
+     # if there's less than 5% edited from the license, we found our match
+-    if norm and min_edit_dist < cutoff:
++    if norm and min_edit_dist / len(norm) < .05:
+         return min_edit_dist_spdx
+     else:
+         # no matches :'(
+--- a/setup.cfg	2023-10-28 19:19:41.000000000 +0200
++++ b/setup.cfg	2023-11-01 10:13:07.904157653 +0100
+@@ -31,7 +31,7 @@
+ 
+ [options.extras_require]
+ license =
+-    ukkonen
++    editdistance
+ 
+ [options.package_data]
+ identify =

identify-2.2.13.tar.gz

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:cd06ccd046da3ece6bbc0a581173c8b67322bbfe6a80a2f52c5915cac560333f
-size 101136

identify-2.5.31.tar.gz

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4d119d9e1cff305bbcb0743f106daaada7035d171c7c9fd21883b5f414d2f422
+size 101612

python-identify.changes

@@ -1,3 +1,20 @@
+-------------------------------------------------------------------
+Wed Nov  1 09:17:44 UTC 2023 - Matthias Fehring <buschmann23@opensuse.org>
+
+- update to version 2.5.31
+  * no changelog available, alternatively have a look at
+    https://github.com/pre-commit/identify/compare/v2.2.13...v2.5.31
+- use editdistance instead of ukkonen
+  * Upstream changed from editdistance_s to ukkonen for performance
+    reasons. Ukkonen is not now available for openSUSE. So switch
+    back to editdistance for now.
+  * Add 0001-use-editdistance-not-ukkonen.patch
+- spec file changes
+  * remove not required python-setuptools from Requires
+  * upstream URL has changed to https://github.com/pre-commit/identify
+  * change requirement for python-editdistance from Suggests to
+    Recommends
+
 -------------------------------------------------------------------
 Wed Aug 18 12:47:19 UTC 2021 - John Paul Adrian Glaubitz <adrian.glaubitz@suse.com>
 

python-identify.spec

@@ -1,7 +1,7 @@
 #
 # spec file for package python-identify
 #
-# Copyright (c) 2021 SUSE LLC
+# Copyright (c) 2023 SUSE LLC
 #
 # All modifications and additions to the file contributed by third parties
 # remain the property of their copyright owners, unless otherwise agreed
@@ -16,24 +16,24 @@
 #
 
 
-%{?!python_module:%define python_module() python-%{**} python3-%{**}}
 Name:           python-identify
-Version:        2.2.13
+Version:        2.5.31
 Release:        0
 Summary:        File identification library for Python
 License:        MIT
 Group:          Development/Languages/Python
-URL:            https://github.com/chriskuehl/identify
-Source:         https://github.com/chriskuehl/identify/archive/v%{version}.tar.gz#/identify-%{version}.tar.gz
+URL:            https://github.com/pre-commit/identify
+Source:         https://github.com/pre-commit/identify/archive/v%{version}.tar.gz#/identify-%{version}.tar.gz
+# PATCH-FIX-OPENSUSE 0001-use-editdistance-not-ukkonen.patch -- ukkonen not packaged for opensuse now
+Patch1:         0001-use-editdistance-not-ukkonen.patch
 BuildRequires:  %{python_module editdistance}
 BuildRequires:  %{python_module pytest}
 BuildRequires:  %{python_module setuptools}
 BuildRequires:  fdupes
 BuildRequires:  python-rpm-macros
-Requires:       python-setuptools
 Requires(post): update-alternatives
 Requires(postun):update-alternatives
-Suggests:       python-editdistance
+Recommends:     python-editdistance
 BuildArch:      noarch
 %python_subpackages
 
@@ -42,9 +42,7 @@ File identification library for Python, including license file SPDX identifier.
 
 %prep
 %setup -q -n identify-%{version}
-# stick with editdistance as it generally has more functionality
-sed -i 's/editdistance_s.distance/editdistance.eval/' identify/identify.py
-sed -i 's/editdistance_s/editdistance/' identify/identify.py
+%autopatch -p1
 
 %build
 %python_build
@@ -67,6 +65,7 @@ sed -i 's/editdistance_s/editdistance/' identify/identify.py
 %doc README.md
 %license LICENSE
 %python_alternative %{_bindir}/identify-cli
-%{python_sitelib}/*
+%{python_sitelib}/identify
+%{python_sitelib}/identify-%{version}-*-info
 
 %changelog