Merge pull request #244656 from jokatzke/trafilatura (bc48aabf) · Commits · nix / nixpkgs

maintainers/maintainer-list.nix

+6 −0

Original line number	Diff line number	Diff line
		@@ -9162,6 +9162,12 @@
		fingerprint = "7249 70E6 A661 D84E 8B47 678A 0590 93B1 A278 BCD0";
		}];
		};
		jokatzke = {
		email = "jokatzke@fastmail.com";
		github = "jokatzke";
		githubId = 46931073;
		name = "Jonas Katzke";
		};
		joko = {
		email = "ioannis.koutras@gmail.com";
		github = "jokogr";

0 → 100644

+54 −0

Original line number	Diff line number	Diff line
		{ lib
		, buildPythonPackage
		, fetchPypi
		, langcodes
		, pytestCheckHook
		, tld
		, urllib3
		, pythonOlder
		}:

		buildPythonPackage rec {
		pname = "courlan";
		version = "0.9.5";
		format = "setuptools";

		disabled = pythonOlder "3.6";

		src = fetchPypi {
		inherit pname version;
		hash = "sha256-ONw1suO/H11RbQDVGsEuveVD40F8a+b2oic8D8W1s1M=";
		};

		propagatedBuildInputs = [
		langcodes
		tld
		urllib3
		];

		nativeCheckInputs = [
		pytestCheckHook
		];

		# disable tests that require an internet connection
		disabledTests = [
		"test_urlcheck"
		];

		# nixify path to the courlan binary in the test suite
		postPatch = ''
		substituteInPlace tests/unit_tests.py \
		--replace "\"courlan --help\"" "\"$out/bin/courlan --help\"" \
		--replace "courlan_bin = \"courlan\"" "courlan_bin = \"$out/bin/courlan\""
		'';

		pythonImportsCheck = [ "courlan" ];

		meta = with lib; {
		description = "Clean, filter and sample URLs to optimize data collection";
		homepage = "https://github.com/adbar/courlan";
		changelog = "https://github.com/adbar/courlan/blob/v${version}/HISTORY.md";
		license = licenses.gpl3Plus;
		maintainers = with maintainers; [ jokatzke ];
		};
		}

0 → 100644

+56 −0

Original line number	Diff line number	Diff line
		{ lib
		, buildPythonPackage
		, fetchPypi
		, pythonOlder
		, charset-normalizer
		, dateparser
		, lxml
		, pytestCheckHook
		, python-dateutil
		, urllib3
		, backports-datetime-fromisoformat
		}:

		buildPythonPackage rec {
		pname = "htmldate";
		version = "1.6.0";
		format = "setuptools";

		disabled = pythonOlder "3.6";

		src = fetchPypi {
		inherit pname version;
		hash = "sha256-WCfI9iahaACinlfoGIo9MtCwjKTHvWYlN7c7u/IsRaY=";
		};

		propagatedBuildInputs = [
		charset-normalizer
		dateparser
		lxml
		python-dateutil
		urllib3
		] ++ lib.optionals (pythonOlder "3.7") [
		backports-datetime-fromisoformat
		];

		nativeCheckInputs = [
		pytestCheckHook
		];

		# disable tests that require an internet connection
		disabledTests = [
		"test_input"
		"test_cli"
		"test_download"
		];

		pythonImportsCheck = [ "htmldate" ];

		meta = with lib; {
		description = "Fast and robust extraction of original and updated publication dates from URLs and web pages";
		homepage = "https://htmldate.readthedocs.io";
		changelog = "https://github.com/adbar/htmldate/blob/v${version}/CHANGELOG.md";
		license = licenses.gpl3Plus;
		maintainers = with maintainers; [ jokatzke ];
		};
		}

0 → 100644

+43 −0

Original line number	Diff line number	Diff line
		{ lib
		, buildPythonPackage
		, fetchFromGitHub
		, pytestCheckHook
		, lxml
		}:

		buildPythonPackage rec {
		pname = "justext";
		version = "3.0.0";
		format = "setuptools";

		src = fetchFromGitHub {
		owner = "miso-belica";
		repo = "jusText";
		rev = "refs/tags/v${version}";
		hash = "sha256-WNxDoM5666tEHS9pMl5dOoig4S7dSYaCLZq71tehWqw=";
		};

		propagatedBuildInputs = [
		lxml
		];

		nativeCheckInputs = [
		pytestCheckHook
		];

		# patch out coverage report
		postPatch = ''
		substituteInPlace setup.cfg \
		--replace " --cov=justext --cov-report=term-missing --no-cov-on-fail" ""
		'';

		pythonImportsCheck = [ "justext" ];

		meta = with lib; {
		description = "Heuristic based boilerplate removal tool";
		homepage = "https://github.com/miso-belica/jusText";
		changelog = "https://github.com/miso-belica/jusText/blob/v${version}/CHANGELOG.rst";
		license = licenses.bsd2;
		maintainers = with maintainers; [ jokatzke ];
		};
		}

0 → 100644

+43 −0

Original line number	Diff line number	Diff line
		{ lib
		, buildPythonPackage
		, fetchPypi
		, pythonOlder
		, numpy
		, pytestCheckHook
		}:

		buildPythonPackage rec {
		pname = "py3langid";
		version = "0.2.2";
		format = "setuptools";

		disabled = pythonOlder "3.6";

		src = fetchPypi {
		inherit pname version;
		hash = "sha256-tN4B2tfnAfKdIWoJNeheCWzIZ1kD0j6oRFsrtfCQuW8=";
		};

		propagatedBuildInputs = [
		numpy
		];

		nativeCheckInputs = [
		pytestCheckHook
		];

		# nixify path to the courlan binary in the test suite
		postPatch = ''
		substituteInPlace tests/test_langid.py --replace "'langid'" "'$out/bin/langid'"
		'';

		pythonImportsCheck = [ "py3langid" ];

		meta = with lib; {
		description = "Fork of the language identification tool langid.py, featuring a modernized codebase and faster execution times";
		homepage = "https://github.com/adbar/py3langid";
		changelog = "https://github.com/adbar/py3langid/blob/v${version}/HISTORY.rst";
		license = licenses.bsd3;
		maintainers = with maintainers; [ jokatzke ];
		};
		}