rigid 0.2.0
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- checksums.yaml +7 -0
- data/bin/rigid +4 -0
- data/rigid +16 -0
- data/vendor/PyYAML-3.12.dist-info/DESCRIPTION.rst +12 -0
- data/vendor/PyYAML-3.12.dist-info/INSTALLER +1 -0
- data/vendor/PyYAML-3.12.dist-info/METADATA +35 -0
- data/vendor/PyYAML-3.12.dist-info/RECORD +42 -0
- data/vendor/PyYAML-3.12.dist-info/WHEEL +5 -0
- data/vendor/PyYAML-3.12.dist-info/metadata.json +1 -0
- data/vendor/PyYAML-3.12.dist-info/top_level.txt +2 -0
- data/vendor/_yaml.so +0 -0
- data/vendor/click/__init__.py +98 -0
- data/vendor/click/__init__.pyc +0 -0
- data/vendor/click/_bashcomplete.py +83 -0
- data/vendor/click/_bashcomplete.pyc +0 -0
- data/vendor/click/_compat.py +642 -0
- data/vendor/click/_compat.pyc +0 -0
- data/vendor/click/_termui_impl.py +547 -0
- data/vendor/click/_termui_impl.pyc +0 -0
- data/vendor/click/_textwrap.py +38 -0
- data/vendor/click/_textwrap.pyc +0 -0
- data/vendor/click/_unicodefun.py +119 -0
- data/vendor/click/_unicodefun.pyc +0 -0
- data/vendor/click/_winconsole.py +273 -0
- data/vendor/click/_winconsole.pyc +0 -0
- data/vendor/click/core.py +1738 -0
- data/vendor/click/core.pyc +0 -0
- data/vendor/click/decorators.py +304 -0
- data/vendor/click/decorators.pyc +0 -0
- data/vendor/click/exceptions.py +201 -0
- data/vendor/click/exceptions.pyc +0 -0
- data/vendor/click/formatting.py +256 -0
- data/vendor/click/formatting.pyc +0 -0
- data/vendor/click/globals.py +48 -0
- data/vendor/click/globals.pyc +0 -0
- data/vendor/click/parser.py +426 -0
- data/vendor/click/parser.pyc +0 -0
- data/vendor/click/termui.py +539 -0
- data/vendor/click/termui.pyc +0 -0
- data/vendor/click/testing.py +322 -0
- data/vendor/click/testing.pyc +0 -0
- data/vendor/click/types.py +550 -0
- data/vendor/click/types.pyc +0 -0
- data/vendor/click/utils.py +415 -0
- data/vendor/click/utils.pyc +0 -0
- data/vendor/click-6.6.dist-info/DESCRIPTION.rst +3 -0
- data/vendor/click-6.6.dist-info/INSTALLER +1 -0
- data/vendor/click-6.6.dist-info/METADATA +16 -0
- data/vendor/click-6.6.dist-info/RECORD +41 -0
- data/vendor/click-6.6.dist-info/WHEEL +6 -0
- data/vendor/click-6.6.dist-info/metadata.json +1 -0
- data/vendor/click-6.6.dist-info/top_level.txt +1 -0
- data/vendor/easy_install.py +5 -0
- data/vendor/easy_install.pyc +0 -0
- data/vendor/pip-9.0.1.dist-info/DESCRIPTION.rst +39 -0
- data/vendor/pip-9.0.1.dist-info/INSTALLER +1 -0
- data/vendor/pip-9.0.1.dist-info/METADATA +69 -0
- data/vendor/pip-9.0.1.dist-info/RECORD +501 -0
- data/vendor/pip-9.0.1.dist-info/WHEEL +6 -0
- data/vendor/pip-9.0.1.dist-info/entry_points.txt +5 -0
- data/vendor/pip-9.0.1.dist-info/metadata.json +1 -0
- data/vendor/pip-9.0.1.dist-info/top_level.txt +1 -0
- data/vendor/pkg_resources/__init__.py +3051 -0
- data/vendor/pkg_resources/__init__.pyc +0 -0
- data/vendor/pkg_resources/_vendor/__init__.py +0 -0
- data/vendor/pkg_resources/_vendor/__init__.pyc +0 -0
- data/vendor/pkg_resources/_vendor/appdirs.py +552 -0
- data/vendor/pkg_resources/_vendor/appdirs.pyc +0 -0
- data/vendor/pkg_resources/_vendor/packaging/__about__.py +21 -0
- data/vendor/pkg_resources/_vendor/packaging/__about__.pyc +0 -0
- data/vendor/pkg_resources/_vendor/packaging/__init__.py +14 -0
- data/vendor/pkg_resources/_vendor/packaging/__init__.pyc +0 -0
- data/vendor/pkg_resources/_vendor/packaging/_compat.py +30 -0
- data/vendor/pkg_resources/_vendor/packaging/_compat.pyc +0 -0
- data/vendor/pkg_resources/_vendor/packaging/_structures.py +68 -0
- data/vendor/pkg_resources/_vendor/packaging/_structures.pyc +0 -0
- data/vendor/pkg_resources/_vendor/packaging/markers.py +287 -0
- data/vendor/pkg_resources/_vendor/packaging/markers.pyc +0 -0
- data/vendor/pkg_resources/_vendor/packaging/requirements.py +127 -0
- data/vendor/pkg_resources/_vendor/packaging/requirements.pyc +0 -0
- data/vendor/pkg_resources/_vendor/packaging/specifiers.py +774 -0
- data/vendor/pkg_resources/_vendor/packaging/specifiers.pyc +0 -0
- data/vendor/pkg_resources/_vendor/packaging/utils.py +14 -0
- data/vendor/pkg_resources/_vendor/packaging/utils.pyc +0 -0
- data/vendor/pkg_resources/_vendor/packaging/version.py +393 -0
- data/vendor/pkg_resources/_vendor/packaging/version.pyc +0 -0
- data/vendor/pkg_resources/_vendor/pyparsing.py +5696 -0
- data/vendor/pkg_resources/_vendor/pyparsing.pyc +0 -0
- data/vendor/pkg_resources/_vendor/six.py +868 -0
- data/vendor/pkg_resources/_vendor/six.pyc +0 -0
- data/vendor/pkg_resources/extern/__init__.py +73 -0
- data/vendor/pkg_resources/extern/__init__.pyc +0 -0
- data/vendor/requests/__init__.py +86 -0
- data/vendor/requests/__init__.pyc +0 -0
- data/vendor/requests/adapters.py +503 -0
- data/vendor/requests/adapters.pyc +0 -0
- data/vendor/requests/api.py +148 -0
- data/vendor/requests/api.pyc +0 -0
- data/vendor/requests/auth.py +252 -0
- data/vendor/requests/auth.pyc +0 -0
- data/vendor/requests/cacert.pem +5616 -0
- data/vendor/requests/certs.py +25 -0
- data/vendor/requests/certs.pyc +0 -0
- data/vendor/requests/compat.py +66 -0
- data/vendor/requests/compat.pyc +0 -0
- data/vendor/requests/cookies.py +540 -0
- data/vendor/requests/cookies.pyc +0 -0
- data/vendor/requests/exceptions.py +114 -0
- data/vendor/requests/exceptions.pyc +0 -0
- data/vendor/requests/hooks.py +34 -0
- data/vendor/requests/hooks.pyc +0 -0
- data/vendor/requests/models.py +873 -0
- data/vendor/requests/models.pyc +0 -0
- data/vendor/requests/packages/__init__.py +36 -0
- data/vendor/requests/packages/__init__.pyc +0 -0
- data/vendor/requests/packages/chardet/__init__.py +32 -0
- data/vendor/requests/packages/chardet/__init__.pyc +0 -0
- data/vendor/requests/packages/chardet/big5freq.py +925 -0
- data/vendor/requests/packages/chardet/big5freq.pyc +0 -0
- data/vendor/requests/packages/chardet/big5prober.py +42 -0
- data/vendor/requests/packages/chardet/big5prober.pyc +0 -0
- data/vendor/requests/packages/chardet/chardetect.py +80 -0
- data/vendor/requests/packages/chardet/chardetect.pyc +0 -0
- data/vendor/requests/packages/chardet/chardistribution.py +231 -0
- data/vendor/requests/packages/chardet/chardistribution.pyc +0 -0
- data/vendor/requests/packages/chardet/charsetgroupprober.py +106 -0
- data/vendor/requests/packages/chardet/charsetgroupprober.pyc +0 -0
- data/vendor/requests/packages/chardet/charsetprober.py +62 -0
- data/vendor/requests/packages/chardet/charsetprober.pyc +0 -0
- data/vendor/requests/packages/chardet/codingstatemachine.py +61 -0
- data/vendor/requests/packages/chardet/codingstatemachine.pyc +0 -0
- data/vendor/requests/packages/chardet/compat.py +34 -0
- data/vendor/requests/packages/chardet/compat.pyc +0 -0
- data/vendor/requests/packages/chardet/constants.py +39 -0
- data/vendor/requests/packages/chardet/constants.pyc +0 -0
- data/vendor/requests/packages/chardet/cp949prober.py +44 -0
- data/vendor/requests/packages/chardet/cp949prober.pyc +0 -0
- data/vendor/requests/packages/chardet/escprober.py +86 -0
- data/vendor/requests/packages/chardet/escprober.pyc +0 -0
- data/vendor/requests/packages/chardet/escsm.py +242 -0
- data/vendor/requests/packages/chardet/escsm.pyc +0 -0
- data/vendor/requests/packages/chardet/eucjpprober.py +90 -0
- data/vendor/requests/packages/chardet/eucjpprober.pyc +0 -0
- data/vendor/requests/packages/chardet/euckrfreq.py +596 -0
- data/vendor/requests/packages/chardet/euckrfreq.pyc +0 -0
- data/vendor/requests/packages/chardet/euckrprober.py +42 -0
- data/vendor/requests/packages/chardet/euckrprober.pyc +0 -0
- data/vendor/requests/packages/chardet/euctwfreq.py +428 -0
- data/vendor/requests/packages/chardet/euctwfreq.pyc +0 -0
- data/vendor/requests/packages/chardet/euctwprober.py +41 -0
- data/vendor/requests/packages/chardet/euctwprober.pyc +0 -0
- data/vendor/requests/packages/chardet/gb2312freq.py +472 -0
- data/vendor/requests/packages/chardet/gb2312freq.pyc +0 -0
- data/vendor/requests/packages/chardet/gb2312prober.py +41 -0
- data/vendor/requests/packages/chardet/gb2312prober.pyc +0 -0
- data/vendor/requests/packages/chardet/hebrewprober.py +283 -0
- data/vendor/requests/packages/chardet/hebrewprober.pyc +0 -0
- data/vendor/requests/packages/chardet/jisfreq.py +569 -0
- data/vendor/requests/packages/chardet/jisfreq.pyc +0 -0
- data/vendor/requests/packages/chardet/jpcntx.py +227 -0
- data/vendor/requests/packages/chardet/jpcntx.pyc +0 -0
- data/vendor/requests/packages/chardet/langbulgarianmodel.py +229 -0
- data/vendor/requests/packages/chardet/langbulgarianmodel.pyc +0 -0
- data/vendor/requests/packages/chardet/langcyrillicmodel.py +329 -0
- data/vendor/requests/packages/chardet/langcyrillicmodel.pyc +0 -0
- data/vendor/requests/packages/chardet/langgreekmodel.py +225 -0
- data/vendor/requests/packages/chardet/langgreekmodel.pyc +0 -0
- data/vendor/requests/packages/chardet/langhebrewmodel.py +201 -0
- data/vendor/requests/packages/chardet/langhebrewmodel.pyc +0 -0
- data/vendor/requests/packages/chardet/langhungarianmodel.py +225 -0
- data/vendor/requests/packages/chardet/langhungarianmodel.pyc +0 -0
- data/vendor/requests/packages/chardet/langthaimodel.py +200 -0
- data/vendor/requests/packages/chardet/langthaimodel.pyc +0 -0
- data/vendor/requests/packages/chardet/latin1prober.py +139 -0
- data/vendor/requests/packages/chardet/latin1prober.pyc +0 -0
- data/vendor/requests/packages/chardet/mbcharsetprober.py +86 -0
- data/vendor/requests/packages/chardet/mbcharsetprober.pyc +0 -0
- data/vendor/requests/packages/chardet/mbcsgroupprober.py +54 -0
- data/vendor/requests/packages/chardet/mbcsgroupprober.pyc +0 -0
- data/vendor/requests/packages/chardet/mbcssm.py +572 -0
- data/vendor/requests/packages/chardet/mbcssm.pyc +0 -0
- data/vendor/requests/packages/chardet/sbcharsetprober.py +120 -0
- data/vendor/requests/packages/chardet/sbcharsetprober.pyc +0 -0
- data/vendor/requests/packages/chardet/sbcsgroupprober.py +69 -0
- data/vendor/requests/packages/chardet/sbcsgroupprober.pyc +0 -0
- data/vendor/requests/packages/chardet/sjisprober.py +91 -0
- data/vendor/requests/packages/chardet/sjisprober.pyc +0 -0
- data/vendor/requests/packages/chardet/universaldetector.py +170 -0
- data/vendor/requests/packages/chardet/universaldetector.pyc +0 -0
- data/vendor/requests/packages/chardet/utf8prober.py +76 -0
- data/vendor/requests/packages/chardet/utf8prober.pyc +0 -0
- data/vendor/requests/packages/urllib3/__init__.py +96 -0
- data/vendor/requests/packages/urllib3/__init__.pyc +0 -0
- data/vendor/requests/packages/urllib3/_collections.py +324 -0
- data/vendor/requests/packages/urllib3/_collections.pyc +0 -0
- data/vendor/requests/packages/urllib3/connection.py +330 -0
- data/vendor/requests/packages/urllib3/connection.pyc +0 -0
- data/vendor/requests/packages/urllib3/connectionpool.py +866 -0
- data/vendor/requests/packages/urllib3/connectionpool.pyc +0 -0
- data/vendor/requests/packages/urllib3/contrib/__init__.py +0 -0
- data/vendor/requests/packages/urllib3/contrib/__init__.pyc +0 -0
- data/vendor/requests/packages/urllib3/contrib/appengine.py +231 -0
- data/vendor/requests/packages/urllib3/contrib/appengine.pyc +0 -0
- data/vendor/requests/packages/urllib3/contrib/ntlmpool.py +115 -0
- data/vendor/requests/packages/urllib3/contrib/ntlmpool.pyc +0 -0
- data/vendor/requests/packages/urllib3/contrib/pyopenssl.py +358 -0
- data/vendor/requests/packages/urllib3/contrib/pyopenssl.pyc +0 -0
- data/vendor/requests/packages/urllib3/contrib/socks.py +172 -0
- data/vendor/requests/packages/urllib3/contrib/socks.pyc +0 -0
- data/vendor/requests/packages/urllib3/exceptions.py +209 -0
- data/vendor/requests/packages/urllib3/exceptions.pyc +0 -0
- data/vendor/requests/packages/urllib3/fields.py +178 -0
- data/vendor/requests/packages/urllib3/fields.pyc +0 -0
- data/vendor/requests/packages/urllib3/filepost.py +94 -0
- data/vendor/requests/packages/urllib3/filepost.pyc +0 -0
- data/vendor/requests/packages/urllib3/packages/__init__.py +5 -0
- data/vendor/requests/packages/urllib3/packages/__init__.pyc +0 -0
- data/vendor/requests/packages/urllib3/packages/ordered_dict.py +259 -0
- data/vendor/requests/packages/urllib3/packages/ordered_dict.pyc +0 -0
- data/vendor/requests/packages/urllib3/packages/six.py +868 -0
- data/vendor/requests/packages/urllib3/packages/six.pyc +0 -0
- data/vendor/requests/packages/urllib3/packages/ssl_match_hostname/__init__.py +13 -0
- data/vendor/requests/packages/urllib3/packages/ssl_match_hostname/__init__.pyc +0 -0
- data/vendor/requests/packages/urllib3/packages/ssl_match_hostname/_implementation.py +105 -0
- data/vendor/requests/packages/urllib3/packages/ssl_match_hostname/_implementation.pyc +0 -0
- data/vendor/requests/packages/urllib3/poolmanager.py +367 -0
- data/vendor/requests/packages/urllib3/poolmanager.pyc +0 -0
- data/vendor/requests/packages/urllib3/request.py +151 -0
- data/vendor/requests/packages/urllib3/request.pyc +0 -0
- data/vendor/requests/packages/urllib3/response.py +530 -0
- data/vendor/requests/packages/urllib3/response.pyc +0 -0
- data/vendor/requests/packages/urllib3/util/__init__.py +46 -0
- data/vendor/requests/packages/urllib3/util/__init__.pyc +0 -0
- data/vendor/requests/packages/urllib3/util/connection.py +144 -0
- data/vendor/requests/packages/urllib3/util/connection.pyc +0 -0
- data/vendor/requests/packages/urllib3/util/request.py +72 -0
- data/vendor/requests/packages/urllib3/util/request.pyc +0 -0
- data/vendor/requests/packages/urllib3/util/response.py +74 -0
- data/vendor/requests/packages/urllib3/util/response.pyc +0 -0
- data/vendor/requests/packages/urllib3/util/retry.py +300 -0
- data/vendor/requests/packages/urllib3/util/retry.pyc +0 -0
- data/vendor/requests/packages/urllib3/util/ssl_.py +320 -0
- data/vendor/requests/packages/urllib3/util/ssl_.pyc +0 -0
- data/vendor/requests/packages/urllib3/util/timeout.py +242 -0
- data/vendor/requests/packages/urllib3/util/timeout.pyc +0 -0
- data/vendor/requests/packages/urllib3/util/url.py +217 -0
- data/vendor/requests/packages/urllib3/util/url.pyc +0 -0
- data/vendor/requests/sessions.py +712 -0
- data/vendor/requests/sessions.pyc +0 -0
- data/vendor/requests/status_codes.py +91 -0
- data/vendor/requests/status_codes.pyc +0 -0
- data/vendor/requests/structures.py +105 -0
- data/vendor/requests/structures.pyc +0 -0
- data/vendor/requests/utils.py +817 -0
- data/vendor/requests/utils.pyc +0 -0
- data/vendor/requests-2.11.1.dist-info/DESCRIPTION.rst +1294 -0
- data/vendor/requests-2.11.1.dist-info/INSTALLER +1 -0
- data/vendor/requests-2.11.1.dist-info/METADATA +1323 -0
- data/vendor/requests-2.11.1.dist-info/RECORD +170 -0
- data/vendor/requests-2.11.1.dist-info/WHEEL +6 -0
- data/vendor/requests-2.11.1.dist-info/metadata.json +1 -0
- data/vendor/requests-2.11.1.dist-info/top_level.txt +1 -0
- data/vendor/rigid/__init__.py +1 -0
- data/vendor/rigid/__init__.pyc +0 -0
- data/vendor/rigid/api.py +129 -0
- data/vendor/rigid/api.pyc +0 -0
- data/vendor/rigid/commands/__init__.py +260 -0
- data/vendor/rigid/commands/__init__.pyc +0 -0
- data/vendor/rigid/commands/deploy.py +0 -0
- data/vendor/rigid/commands/deploy.pyc +0 -0
- data/vendor/rigid/deploy.py +70 -0
- data/vendor/rigid/deploy.pyc +0 -0
- data/vendor/rigid/file_scanner.py +63 -0
- data/vendor/rigid/file_scanner.pyc +0 -0
- data/vendor/rigid/utils.py +17 -0
- data/vendor/rigid/utils.pyc +0 -0
- data/vendor/rigid-0.2.0.dist-info/DESCRIPTION.rst +3 -0
- data/vendor/rigid-0.2.0.dist-info/INSTALLER +1 -0
- data/vendor/rigid-0.2.0.dist-info/METADATA +23 -0
- data/vendor/rigid-0.2.0.dist-info/RECORD +49 -0
- data/vendor/rigid-0.2.0.dist-info/WHEEL +5 -0
- data/vendor/rigid-0.2.0.dist-info/entry_points.txt +3 -0
- data/vendor/rigid-0.2.0.dist-info/metadata.json +1 -0
- data/vendor/rigid-0.2.0.dist-info/top_level.txt +2 -0
- data/vendor/setuptools-28.8.0.dist-info/DESCRIPTION.rst +243 -0
- data/vendor/setuptools-28.8.0.dist-info/INSTALLER +1 -0
- data/vendor/setuptools-28.8.0.dist-info/METADATA +272 -0
- data/vendor/setuptools-28.8.0.dist-info/RECORD +143 -0
- data/vendor/setuptools-28.8.0.dist-info/WHEEL +6 -0
- data/vendor/setuptools-28.8.0.dist-info/dependency_links.txt +2 -0
- data/vendor/setuptools-28.8.0.dist-info/entry_points.txt +63 -0
- data/vendor/setuptools-28.8.0.dist-info/metadata.json +1 -0
- data/vendor/setuptools-28.8.0.dist-info/top_level.txt +3 -0
- data/vendor/setuptools-28.8.0.dist-info/zip-safe +1 -0
- data/vendor/tests/__init__.py +0 -0
- data/vendor/tests/__init__.pyc +0 -0
- data/vendor/tests/integration/__init__.py +0 -0
- data/vendor/tests/integration/__init__.pyc +0 -0
- data/vendor/tests/integration/test_app.py +63 -0
- data/vendor/tests/integration/test_app.pyc +0 -0
- data/vendor/tests/integration/test_apps.py +27 -0
- data/vendor/tests/integration/test_apps.pyc +0 -0
- data/vendor/tests/integration/test_deploy.py +128 -0
- data/vendor/tests/integration/test_deploy.pyc +0 -0
- data/vendor/tests/integration/test_domains.py +35 -0
- data/vendor/tests/integration/test_domains.pyc +0 -0
- data/vendor/tests/integration/test_login.py +37 -0
- data/vendor/tests/integration/test_login.pyc +0 -0
- data/vendor/tests/integration/test_promote.py +24 -0
- data/vendor/tests/integration/test_promote.pyc +0 -0
- data/vendor/tests/integration/test_token.py +33 -0
- data/vendor/tests/integration/test_token.pyc +0 -0
- data/vendor/tests/integration/test_whoami.py +24 -0
- data/vendor/tests/integration/test_whoami.pyc +0 -0
- data/vendor/tests/test_deploy.py +33 -0
- data/vendor/tests/test_deploy.pyc +0 -0
- data/vendor/tests/test_file_scanner.py +89 -0
- data/vendor/tests/test_file_scanner.pyc +0 -0
- data/vendor/tests/utils.py +78 -0
- data/vendor/tests/utils.pyc +0 -0
- data/vendor/wheel-0.30.0a0.dist-info/DESCRIPTION.rst +325 -0
- data/vendor/wheel-0.30.0a0.dist-info/INSTALLER +1 -0
- data/vendor/wheel-0.30.0a0.dist-info/LICENSE.txt +22 -0
- data/vendor/wheel-0.30.0a0.dist-info/METADATA +357 -0
- data/vendor/wheel-0.30.0a0.dist-info/RECORD +86 -0
- data/vendor/wheel-0.30.0a0.dist-info/WHEEL +6 -0
- data/vendor/wheel-0.30.0a0.dist-info/entry_points.txt +6 -0
- data/vendor/wheel-0.30.0a0.dist-info/metadata.json +1 -0
- data/vendor/wheel-0.30.0a0.dist-info/top_level.txt +1 -0
- data/vendor/yaml/__init__.py +315 -0
- data/vendor/yaml/__init__.pyc +0 -0
- data/vendor/yaml/composer.py +139 -0
- data/vendor/yaml/composer.pyc +0 -0
- data/vendor/yaml/constructor.py +675 -0
- data/vendor/yaml/constructor.pyc +0 -0
- data/vendor/yaml/cyaml.py +85 -0
- data/vendor/yaml/cyaml.pyc +0 -0
- data/vendor/yaml/dumper.py +62 -0
- data/vendor/yaml/dumper.pyc +0 -0
- data/vendor/yaml/emitter.py +1140 -0
- data/vendor/yaml/emitter.pyc +0 -0
- data/vendor/yaml/error.py +75 -0
- data/vendor/yaml/error.pyc +0 -0
- data/vendor/yaml/events.py +86 -0
- data/vendor/yaml/events.pyc +0 -0
- data/vendor/yaml/loader.py +40 -0
- data/vendor/yaml/loader.pyc +0 -0
- data/vendor/yaml/nodes.py +49 -0
- data/vendor/yaml/nodes.pyc +0 -0
- data/vendor/yaml/parser.py +589 -0
- data/vendor/yaml/parser.pyc +0 -0
- data/vendor/yaml/reader.py +190 -0
- data/vendor/yaml/reader.pyc +0 -0
- data/vendor/yaml/representer.py +486 -0
- data/vendor/yaml/representer.pyc +0 -0
- data/vendor/yaml/resolver.py +227 -0
- data/vendor/yaml/resolver.pyc +0 -0
- data/vendor/yaml/scanner.py +1453 -0
- data/vendor/yaml/scanner.pyc +0 -0
- data/vendor/yaml/serializer.py +111 -0
- data/vendor/yaml/serializer.pyc +0 -0
- data/vendor/yaml/tokens.py +104 -0
- data/vendor/yaml/tokens.pyc +0 -0
- metadata +407 -0
Binary file
|
@@ -0,0 +1,42 @@
|
|
1
|
+
######################## BEGIN LICENSE BLOCK ########################
|
2
|
+
# The Original Code is Mozilla Communicator client code.
|
3
|
+
#
|
4
|
+
# The Initial Developer of the Original Code is
|
5
|
+
# Netscape Communications Corporation.
|
6
|
+
# Portions created by the Initial Developer are Copyright (C) 1998
|
7
|
+
# the Initial Developer. All Rights Reserved.
|
8
|
+
#
|
9
|
+
# Contributor(s):
|
10
|
+
# Mark Pilgrim - port to Python
|
11
|
+
#
|
12
|
+
# This library is free software; you can redistribute it and/or
|
13
|
+
# modify it under the terms of the GNU Lesser General Public
|
14
|
+
# License as published by the Free Software Foundation; either
|
15
|
+
# version 2.1 of the License, or (at your option) any later version.
|
16
|
+
#
|
17
|
+
# This library is distributed in the hope that it will be useful,
|
18
|
+
# but WITHOUT ANY WARRANTY; without even the implied warranty of
|
19
|
+
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
|
20
|
+
# Lesser General Public License for more details.
|
21
|
+
#
|
22
|
+
# You should have received a copy of the GNU Lesser General Public
|
23
|
+
# License along with this library; if not, write to the Free Software
|
24
|
+
# Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA
|
25
|
+
# 02110-1301 USA
|
26
|
+
######################### END LICENSE BLOCK #########################
|
27
|
+
|
28
|
+
from .mbcharsetprober import MultiByteCharSetProber
|
29
|
+
from .codingstatemachine import CodingStateMachine
|
30
|
+
from .chardistribution import Big5DistributionAnalysis
|
31
|
+
from .mbcssm import Big5SMModel
|
32
|
+
|
33
|
+
|
34
|
+
class Big5Prober(MultiByteCharSetProber):
|
35
|
+
def __init__(self):
|
36
|
+
MultiByteCharSetProber.__init__(self)
|
37
|
+
self._mCodingSM = CodingStateMachine(Big5SMModel)
|
38
|
+
self._mDistributionAnalyzer = Big5DistributionAnalysis()
|
39
|
+
self.reset()
|
40
|
+
|
41
|
+
def get_charset_name(self):
|
42
|
+
return "Big5"
|
Binary file
|
@@ -0,0 +1,80 @@
|
|
1
|
+
#!/usr/bin/env python
|
2
|
+
"""
|
3
|
+
Script which takes one or more file paths and reports on their detected
|
4
|
+
encodings
|
5
|
+
|
6
|
+
Example::
|
7
|
+
|
8
|
+
% chardetect somefile someotherfile
|
9
|
+
somefile: windows-1252 with confidence 0.5
|
10
|
+
someotherfile: ascii with confidence 1.0
|
11
|
+
|
12
|
+
If no paths are provided, it takes its input from stdin.
|
13
|
+
|
14
|
+
"""
|
15
|
+
|
16
|
+
from __future__ import absolute_import, print_function, unicode_literals
|
17
|
+
|
18
|
+
import argparse
|
19
|
+
import sys
|
20
|
+
from io import open
|
21
|
+
|
22
|
+
from chardet import __version__
|
23
|
+
from chardet.universaldetector import UniversalDetector
|
24
|
+
|
25
|
+
|
26
|
+
def description_of(lines, name='stdin'):
|
27
|
+
"""
|
28
|
+
Return a string describing the probable encoding of a file or
|
29
|
+
list of strings.
|
30
|
+
|
31
|
+
:param lines: The lines to get the encoding of.
|
32
|
+
:type lines: Iterable of bytes
|
33
|
+
:param name: Name of file or collection of lines
|
34
|
+
:type name: str
|
35
|
+
"""
|
36
|
+
u = UniversalDetector()
|
37
|
+
for line in lines:
|
38
|
+
u.feed(line)
|
39
|
+
u.close()
|
40
|
+
result = u.result
|
41
|
+
if result['encoding']:
|
42
|
+
return '{0}: {1} with confidence {2}'.format(name, result['encoding'],
|
43
|
+
result['confidence'])
|
44
|
+
else:
|
45
|
+
return '{0}: no result'.format(name)
|
46
|
+
|
47
|
+
|
48
|
+
def main(argv=None):
|
49
|
+
'''
|
50
|
+
Handles command line arguments and gets things started.
|
51
|
+
|
52
|
+
:param argv: List of arguments, as if specified on the command-line.
|
53
|
+
If None, ``sys.argv[1:]`` is used instead.
|
54
|
+
:type argv: list of str
|
55
|
+
'''
|
56
|
+
# Get command line arguments
|
57
|
+
parser = argparse.ArgumentParser(
|
58
|
+
description="Takes one or more file paths and reports their detected \
|
59
|
+
encodings",
|
60
|
+
formatter_class=argparse.ArgumentDefaultsHelpFormatter,
|
61
|
+
conflict_handler='resolve')
|
62
|
+
parser.add_argument('input',
|
63
|
+
help='File whose encoding we would like to determine.',
|
64
|
+
type=argparse.FileType('rb'), nargs='*',
|
65
|
+
default=[sys.stdin])
|
66
|
+
parser.add_argument('--version', action='version',
|
67
|
+
version='%(prog)s {0}'.format(__version__))
|
68
|
+
args = parser.parse_args(argv)
|
69
|
+
|
70
|
+
for f in args.input:
|
71
|
+
if f.isatty():
|
72
|
+
print("You are running chardetect interactively. Press " +
|
73
|
+
"CTRL-D twice at the start of a blank line to signal the " +
|
74
|
+
"end of your input. If you want help, run chardetect " +
|
75
|
+
"--help\n", file=sys.stderr)
|
76
|
+
print(description_of(f, f.name))
|
77
|
+
|
78
|
+
|
79
|
+
if __name__ == '__main__':
|
80
|
+
main()
|
Binary file
|
@@ -0,0 +1,231 @@
|
|
1
|
+
######################## BEGIN LICENSE BLOCK ########################
|
2
|
+
# The Original Code is Mozilla Communicator client code.
|
3
|
+
#
|
4
|
+
# The Initial Developer of the Original Code is
|
5
|
+
# Netscape Communications Corporation.
|
6
|
+
# Portions created by the Initial Developer are Copyright (C) 1998
|
7
|
+
# the Initial Developer. All Rights Reserved.
|
8
|
+
#
|
9
|
+
# Contributor(s):
|
10
|
+
# Mark Pilgrim - port to Python
|
11
|
+
#
|
12
|
+
# This library is free software; you can redistribute it and/or
|
13
|
+
# modify it under the terms of the GNU Lesser General Public
|
14
|
+
# License as published by the Free Software Foundation; either
|
15
|
+
# version 2.1 of the License, or (at your option) any later version.
|
16
|
+
#
|
17
|
+
# This library is distributed in the hope that it will be useful,
|
18
|
+
# but WITHOUT ANY WARRANTY; without even the implied warranty of
|
19
|
+
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
|
20
|
+
# Lesser General Public License for more details.
|
21
|
+
#
|
22
|
+
# You should have received a copy of the GNU Lesser General Public
|
23
|
+
# License along with this library; if not, write to the Free Software
|
24
|
+
# Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA
|
25
|
+
# 02110-1301 USA
|
26
|
+
######################### END LICENSE BLOCK #########################
|
27
|
+
|
28
|
+
from .euctwfreq import (EUCTWCharToFreqOrder, EUCTW_TABLE_SIZE,
|
29
|
+
EUCTW_TYPICAL_DISTRIBUTION_RATIO)
|
30
|
+
from .euckrfreq import (EUCKRCharToFreqOrder, EUCKR_TABLE_SIZE,
|
31
|
+
EUCKR_TYPICAL_DISTRIBUTION_RATIO)
|
32
|
+
from .gb2312freq import (GB2312CharToFreqOrder, GB2312_TABLE_SIZE,
|
33
|
+
GB2312_TYPICAL_DISTRIBUTION_RATIO)
|
34
|
+
from .big5freq import (Big5CharToFreqOrder, BIG5_TABLE_SIZE,
|
35
|
+
BIG5_TYPICAL_DISTRIBUTION_RATIO)
|
36
|
+
from .jisfreq import (JISCharToFreqOrder, JIS_TABLE_SIZE,
|
37
|
+
JIS_TYPICAL_DISTRIBUTION_RATIO)
|
38
|
+
from .compat import wrap_ord
|
39
|
+
|
40
|
+
ENOUGH_DATA_THRESHOLD = 1024
|
41
|
+
SURE_YES = 0.99
|
42
|
+
SURE_NO = 0.01
|
43
|
+
MINIMUM_DATA_THRESHOLD = 3
|
44
|
+
|
45
|
+
|
46
|
+
class CharDistributionAnalysis:
|
47
|
+
def __init__(self):
|
48
|
+
# Mapping table to get frequency order from char order (get from
|
49
|
+
# GetOrder())
|
50
|
+
self._mCharToFreqOrder = None
|
51
|
+
self._mTableSize = None # Size of above table
|
52
|
+
# This is a constant value which varies from language to language,
|
53
|
+
# used in calculating confidence. See
|
54
|
+
# http://www.mozilla.org/projects/intl/UniversalCharsetDetection.html
|
55
|
+
# for further detail.
|
56
|
+
self._mTypicalDistributionRatio = None
|
57
|
+
self.reset()
|
58
|
+
|
59
|
+
def reset(self):
|
60
|
+
"""reset analyser, clear any state"""
|
61
|
+
# If this flag is set to True, detection is done and conclusion has
|
62
|
+
# been made
|
63
|
+
self._mDone = False
|
64
|
+
self._mTotalChars = 0 # Total characters encountered
|
65
|
+
# The number of characters whose frequency order is less than 512
|
66
|
+
self._mFreqChars = 0
|
67
|
+
|
68
|
+
def feed(self, aBuf, aCharLen):
|
69
|
+
"""feed a character with known length"""
|
70
|
+
if aCharLen == 2:
|
71
|
+
# we only care about 2-bytes character in our distribution analysis
|
72
|
+
order = self.get_order(aBuf)
|
73
|
+
else:
|
74
|
+
order = -1
|
75
|
+
if order >= 0:
|
76
|
+
self._mTotalChars += 1
|
77
|
+
# order is valid
|
78
|
+
if order < self._mTableSize:
|
79
|
+
if 512 > self._mCharToFreqOrder[order]:
|
80
|
+
self._mFreqChars += 1
|
81
|
+
|
82
|
+
def get_confidence(self):
|
83
|
+
"""return confidence based on existing data"""
|
84
|
+
# if we didn't receive any character in our consideration range,
|
85
|
+
# return negative answer
|
86
|
+
if self._mTotalChars <= 0 or self._mFreqChars <= MINIMUM_DATA_THRESHOLD:
|
87
|
+
return SURE_NO
|
88
|
+
|
89
|
+
if self._mTotalChars != self._mFreqChars:
|
90
|
+
r = (self._mFreqChars / ((self._mTotalChars - self._mFreqChars)
|
91
|
+
* self._mTypicalDistributionRatio))
|
92
|
+
if r < SURE_YES:
|
93
|
+
return r
|
94
|
+
|
95
|
+
# normalize confidence (we don't want to be 100% sure)
|
96
|
+
return SURE_YES
|
97
|
+
|
98
|
+
def got_enough_data(self):
|
99
|
+
# It is not necessary to receive all data to draw conclusion.
|
100
|
+
# For charset detection, certain amount of data is enough
|
101
|
+
return self._mTotalChars > ENOUGH_DATA_THRESHOLD
|
102
|
+
|
103
|
+
def get_order(self, aBuf):
|
104
|
+
# We do not handle characters based on the original encoding string,
|
105
|
+
# but convert this encoding string to a number, here called order.
|
106
|
+
# This allows multiple encodings of a language to share one frequency
|
107
|
+
# table.
|
108
|
+
return -1
|
109
|
+
|
110
|
+
|
111
|
+
class EUCTWDistributionAnalysis(CharDistributionAnalysis):
|
112
|
+
def __init__(self):
|
113
|
+
CharDistributionAnalysis.__init__(self)
|
114
|
+
self._mCharToFreqOrder = EUCTWCharToFreqOrder
|
115
|
+
self._mTableSize = EUCTW_TABLE_SIZE
|
116
|
+
self._mTypicalDistributionRatio = EUCTW_TYPICAL_DISTRIBUTION_RATIO
|
117
|
+
|
118
|
+
def get_order(self, aBuf):
|
119
|
+
# for euc-TW encoding, we are interested
|
120
|
+
# first byte range: 0xc4 -- 0xfe
|
121
|
+
# second byte range: 0xa1 -- 0xfe
|
122
|
+
# no validation needed here. State machine has done that
|
123
|
+
first_char = wrap_ord(aBuf[0])
|
124
|
+
if first_char >= 0xC4:
|
125
|
+
return 94 * (first_char - 0xC4) + wrap_ord(aBuf[1]) - 0xA1
|
126
|
+
else:
|
127
|
+
return -1
|
128
|
+
|
129
|
+
|
130
|
+
class EUCKRDistributionAnalysis(CharDistributionAnalysis):
|
131
|
+
def __init__(self):
|
132
|
+
CharDistributionAnalysis.__init__(self)
|
133
|
+
self._mCharToFreqOrder = EUCKRCharToFreqOrder
|
134
|
+
self._mTableSize = EUCKR_TABLE_SIZE
|
135
|
+
self._mTypicalDistributionRatio = EUCKR_TYPICAL_DISTRIBUTION_RATIO
|
136
|
+
|
137
|
+
def get_order(self, aBuf):
|
138
|
+
# for euc-KR encoding, we are interested
|
139
|
+
# first byte range: 0xb0 -- 0xfe
|
140
|
+
# second byte range: 0xa1 -- 0xfe
|
141
|
+
# no validation needed here. State machine has done that
|
142
|
+
first_char = wrap_ord(aBuf[0])
|
143
|
+
if first_char >= 0xB0:
|
144
|
+
return 94 * (first_char - 0xB0) + wrap_ord(aBuf[1]) - 0xA1
|
145
|
+
else:
|
146
|
+
return -1
|
147
|
+
|
148
|
+
|
149
|
+
class GB2312DistributionAnalysis(CharDistributionAnalysis):
|
150
|
+
def __init__(self):
|
151
|
+
CharDistributionAnalysis.__init__(self)
|
152
|
+
self._mCharToFreqOrder = GB2312CharToFreqOrder
|
153
|
+
self._mTableSize = GB2312_TABLE_SIZE
|
154
|
+
self._mTypicalDistributionRatio = GB2312_TYPICAL_DISTRIBUTION_RATIO
|
155
|
+
|
156
|
+
def get_order(self, aBuf):
|
157
|
+
# for GB2312 encoding, we are interested
|
158
|
+
# first byte range: 0xb0 -- 0xfe
|
159
|
+
# second byte range: 0xa1 -- 0xfe
|
160
|
+
# no validation needed here. State machine has done that
|
161
|
+
first_char, second_char = wrap_ord(aBuf[0]), wrap_ord(aBuf[1])
|
162
|
+
if (first_char >= 0xB0) and (second_char >= 0xA1):
|
163
|
+
return 94 * (first_char - 0xB0) + second_char - 0xA1
|
164
|
+
else:
|
165
|
+
return -1
|
166
|
+
|
167
|
+
|
168
|
+
class Big5DistributionAnalysis(CharDistributionAnalysis):
|
169
|
+
def __init__(self):
|
170
|
+
CharDistributionAnalysis.__init__(self)
|
171
|
+
self._mCharToFreqOrder = Big5CharToFreqOrder
|
172
|
+
self._mTableSize = BIG5_TABLE_SIZE
|
173
|
+
self._mTypicalDistributionRatio = BIG5_TYPICAL_DISTRIBUTION_RATIO
|
174
|
+
|
175
|
+
def get_order(self, aBuf):
|
176
|
+
# for big5 encoding, we are interested
|
177
|
+
# first byte range: 0xa4 -- 0xfe
|
178
|
+
# second byte range: 0x40 -- 0x7e , 0xa1 -- 0xfe
|
179
|
+
# no validation needed here. State machine has done that
|
180
|
+
first_char, second_char = wrap_ord(aBuf[0]), wrap_ord(aBuf[1])
|
181
|
+
if first_char >= 0xA4:
|
182
|
+
if second_char >= 0xA1:
|
183
|
+
return 157 * (first_char - 0xA4) + second_char - 0xA1 + 63
|
184
|
+
else:
|
185
|
+
return 157 * (first_char - 0xA4) + second_char - 0x40
|
186
|
+
else:
|
187
|
+
return -1
|
188
|
+
|
189
|
+
|
190
|
+
class SJISDistributionAnalysis(CharDistributionAnalysis):
|
191
|
+
def __init__(self):
|
192
|
+
CharDistributionAnalysis.__init__(self)
|
193
|
+
self._mCharToFreqOrder = JISCharToFreqOrder
|
194
|
+
self._mTableSize = JIS_TABLE_SIZE
|
195
|
+
self._mTypicalDistributionRatio = JIS_TYPICAL_DISTRIBUTION_RATIO
|
196
|
+
|
197
|
+
def get_order(self, aBuf):
|
198
|
+
# for sjis encoding, we are interested
|
199
|
+
# first byte range: 0x81 -- 0x9f , 0xe0 -- 0xfe
|
200
|
+
# second byte range: 0x40 -- 0x7e, 0x81 -- oxfe
|
201
|
+
# no validation needed here. State machine has done that
|
202
|
+
first_char, second_char = wrap_ord(aBuf[0]), wrap_ord(aBuf[1])
|
203
|
+
if (first_char >= 0x81) and (first_char <= 0x9F):
|
204
|
+
order = 188 * (first_char - 0x81)
|
205
|
+
elif (first_char >= 0xE0) and (first_char <= 0xEF):
|
206
|
+
order = 188 * (first_char - 0xE0 + 31)
|
207
|
+
else:
|
208
|
+
return -1
|
209
|
+
order = order + second_char - 0x40
|
210
|
+
if second_char > 0x7F:
|
211
|
+
order = -1
|
212
|
+
return order
|
213
|
+
|
214
|
+
|
215
|
+
class EUCJPDistributionAnalysis(CharDistributionAnalysis):
|
216
|
+
def __init__(self):
|
217
|
+
CharDistributionAnalysis.__init__(self)
|
218
|
+
self._mCharToFreqOrder = JISCharToFreqOrder
|
219
|
+
self._mTableSize = JIS_TABLE_SIZE
|
220
|
+
self._mTypicalDistributionRatio = JIS_TYPICAL_DISTRIBUTION_RATIO
|
221
|
+
|
222
|
+
def get_order(self, aBuf):
|
223
|
+
# for euc-JP encoding, we are interested
|
224
|
+
# first byte range: 0xa0 -- 0xfe
|
225
|
+
# second byte range: 0xa1 -- 0xfe
|
226
|
+
# no validation needed here. State machine has done that
|
227
|
+
char = wrap_ord(aBuf[0])
|
228
|
+
if char >= 0xA0:
|
229
|
+
return 94 * (char - 0xA1) + wrap_ord(aBuf[1]) - 0xa1
|
230
|
+
else:
|
231
|
+
return -1
|
Binary file
|
@@ -0,0 +1,106 @@
|
|
1
|
+
######################## BEGIN LICENSE BLOCK ########################
|
2
|
+
# The Original Code is Mozilla Communicator client code.
|
3
|
+
#
|
4
|
+
# The Initial Developer of the Original Code is
|
5
|
+
# Netscape Communications Corporation.
|
6
|
+
# Portions created by the Initial Developer are Copyright (C) 1998
|
7
|
+
# the Initial Developer. All Rights Reserved.
|
8
|
+
#
|
9
|
+
# Contributor(s):
|
10
|
+
# Mark Pilgrim - port to Python
|
11
|
+
#
|
12
|
+
# This library is free software; you can redistribute it and/or
|
13
|
+
# modify it under the terms of the GNU Lesser General Public
|
14
|
+
# License as published by the Free Software Foundation; either
|
15
|
+
# version 2.1 of the License, or (at your option) any later version.
|
16
|
+
#
|
17
|
+
# This library is distributed in the hope that it will be useful,
|
18
|
+
# but WITHOUT ANY WARRANTY; without even the implied warranty of
|
19
|
+
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
|
20
|
+
# Lesser General Public License for more details.
|
21
|
+
#
|
22
|
+
# You should have received a copy of the GNU Lesser General Public
|
23
|
+
# License along with this library; if not, write to the Free Software
|
24
|
+
# Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA
|
25
|
+
# 02110-1301 USA
|
26
|
+
######################### END LICENSE BLOCK #########################
|
27
|
+
|
28
|
+
from . import constants
|
29
|
+
import sys
|
30
|
+
from .charsetprober import CharSetProber
|
31
|
+
|
32
|
+
|
33
|
+
class CharSetGroupProber(CharSetProber):
|
34
|
+
def __init__(self):
|
35
|
+
CharSetProber.__init__(self)
|
36
|
+
self._mActiveNum = 0
|
37
|
+
self._mProbers = []
|
38
|
+
self._mBestGuessProber = None
|
39
|
+
|
40
|
+
def reset(self):
|
41
|
+
CharSetProber.reset(self)
|
42
|
+
self._mActiveNum = 0
|
43
|
+
for prober in self._mProbers:
|
44
|
+
if prober:
|
45
|
+
prober.reset()
|
46
|
+
prober.active = True
|
47
|
+
self._mActiveNum += 1
|
48
|
+
self._mBestGuessProber = None
|
49
|
+
|
50
|
+
def get_charset_name(self):
|
51
|
+
if not self._mBestGuessProber:
|
52
|
+
self.get_confidence()
|
53
|
+
if not self._mBestGuessProber:
|
54
|
+
return None
|
55
|
+
# self._mBestGuessProber = self._mProbers[0]
|
56
|
+
return self._mBestGuessProber.get_charset_name()
|
57
|
+
|
58
|
+
def feed(self, aBuf):
|
59
|
+
for prober in self._mProbers:
|
60
|
+
if not prober:
|
61
|
+
continue
|
62
|
+
if not prober.active:
|
63
|
+
continue
|
64
|
+
st = prober.feed(aBuf)
|
65
|
+
if not st:
|
66
|
+
continue
|
67
|
+
if st == constants.eFoundIt:
|
68
|
+
self._mBestGuessProber = prober
|
69
|
+
return self.get_state()
|
70
|
+
elif st == constants.eNotMe:
|
71
|
+
prober.active = False
|
72
|
+
self._mActiveNum -= 1
|
73
|
+
if self._mActiveNum <= 0:
|
74
|
+
self._mState = constants.eNotMe
|
75
|
+
return self.get_state()
|
76
|
+
return self.get_state()
|
77
|
+
|
78
|
+
def get_confidence(self):
|
79
|
+
st = self.get_state()
|
80
|
+
if st == constants.eFoundIt:
|
81
|
+
return 0.99
|
82
|
+
elif st == constants.eNotMe:
|
83
|
+
return 0.01
|
84
|
+
bestConf = 0.0
|
85
|
+
self._mBestGuessProber = None
|
86
|
+
for prober in self._mProbers:
|
87
|
+
if not prober:
|
88
|
+
continue
|
89
|
+
if not prober.active:
|
90
|
+
if constants._debug:
|
91
|
+
sys.stderr.write(prober.get_charset_name()
|
92
|
+
+ ' not active\n')
|
93
|
+
continue
|
94
|
+
cf = prober.get_confidence()
|
95
|
+
if constants._debug:
|
96
|
+
sys.stderr.write('%s confidence = %s\n' %
|
97
|
+
(prober.get_charset_name(), cf))
|
98
|
+
if bestConf < cf:
|
99
|
+
bestConf = cf
|
100
|
+
self._mBestGuessProber = prober
|
101
|
+
if not self._mBestGuessProber:
|
102
|
+
return 0.0
|
103
|
+
return bestConf
|
104
|
+
# else:
|
105
|
+
# self._mBestGuessProber = self._mProbers[0]
|
106
|
+
# return self._mBestGuessProber.get_confidence()
|
Binary file
|
@@ -0,0 +1,62 @@
|
|
1
|
+
######################## BEGIN LICENSE BLOCK ########################
|
2
|
+
# The Original Code is Mozilla Universal charset detector code.
|
3
|
+
#
|
4
|
+
# The Initial Developer of the Original Code is
|
5
|
+
# Netscape Communications Corporation.
|
6
|
+
# Portions created by the Initial Developer are Copyright (C) 2001
|
7
|
+
# the Initial Developer. All Rights Reserved.
|
8
|
+
#
|
9
|
+
# Contributor(s):
|
10
|
+
# Mark Pilgrim - port to Python
|
11
|
+
# Shy Shalom - original C code
|
12
|
+
#
|
13
|
+
# This library is free software; you can redistribute it and/or
|
14
|
+
# modify it under the terms of the GNU Lesser General Public
|
15
|
+
# License as published by the Free Software Foundation; either
|
16
|
+
# version 2.1 of the License, or (at your option) any later version.
|
17
|
+
#
|
18
|
+
# This library is distributed in the hope that it will be useful,
|
19
|
+
# but WITHOUT ANY WARRANTY; without even the implied warranty of
|
20
|
+
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
|
21
|
+
# Lesser General Public License for more details.
|
22
|
+
#
|
23
|
+
# You should have received a copy of the GNU Lesser General Public
|
24
|
+
# License along with this library; if not, write to the Free Software
|
25
|
+
# Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA
|
26
|
+
# 02110-1301 USA
|
27
|
+
######################### END LICENSE BLOCK #########################
|
28
|
+
|
29
|
+
from . import constants
|
30
|
+
import re
|
31
|
+
|
32
|
+
|
33
|
+
class CharSetProber:
|
34
|
+
def __init__(self):
|
35
|
+
pass
|
36
|
+
|
37
|
+
def reset(self):
|
38
|
+
self._mState = constants.eDetecting
|
39
|
+
|
40
|
+
def get_charset_name(self):
|
41
|
+
return None
|
42
|
+
|
43
|
+
def feed(self, aBuf):
|
44
|
+
pass
|
45
|
+
|
46
|
+
def get_state(self):
|
47
|
+
return self._mState
|
48
|
+
|
49
|
+
def get_confidence(self):
|
50
|
+
return 0.0
|
51
|
+
|
52
|
+
def filter_high_bit_only(self, aBuf):
|
53
|
+
aBuf = re.sub(b'([\x00-\x7F])+', b' ', aBuf)
|
54
|
+
return aBuf
|
55
|
+
|
56
|
+
def filter_without_english_letters(self, aBuf):
|
57
|
+
aBuf = re.sub(b'([A-Za-z])+', b' ', aBuf)
|
58
|
+
return aBuf
|
59
|
+
|
60
|
+
def filter_with_english_letters(self, aBuf):
|
61
|
+
# TODO
|
62
|
+
return aBuf
|
Binary file
|
@@ -0,0 +1,61 @@
|
|
1
|
+
######################## BEGIN LICENSE BLOCK ########################
|
2
|
+
# The Original Code is mozilla.org code.
|
3
|
+
#
|
4
|
+
# The Initial Developer of the Original Code is
|
5
|
+
# Netscape Communications Corporation.
|
6
|
+
# Portions created by the Initial Developer are Copyright (C) 1998
|
7
|
+
# the Initial Developer. All Rights Reserved.
|
8
|
+
#
|
9
|
+
# Contributor(s):
|
10
|
+
# Mark Pilgrim - port to Python
|
11
|
+
#
|
12
|
+
# This library is free software; you can redistribute it and/or
|
13
|
+
# modify it under the terms of the GNU Lesser General Public
|
14
|
+
# License as published by the Free Software Foundation; either
|
15
|
+
# version 2.1 of the License, or (at your option) any later version.
|
16
|
+
#
|
17
|
+
# This library is distributed in the hope that it will be useful,
|
18
|
+
# but WITHOUT ANY WARRANTY; without even the implied warranty of
|
19
|
+
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
|
20
|
+
# Lesser General Public License for more details.
|
21
|
+
#
|
22
|
+
# You should have received a copy of the GNU Lesser General Public
|
23
|
+
# License along with this library; if not, write to the Free Software
|
24
|
+
# Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA
|
25
|
+
# 02110-1301 USA
|
26
|
+
######################### END LICENSE BLOCK #########################
|
27
|
+
|
28
|
+
from .constants import eStart
|
29
|
+
from .compat import wrap_ord
|
30
|
+
|
31
|
+
|
32
|
+
class CodingStateMachine:
|
33
|
+
def __init__(self, sm):
|
34
|
+
self._mModel = sm
|
35
|
+
self._mCurrentBytePos = 0
|
36
|
+
self._mCurrentCharLen = 0
|
37
|
+
self.reset()
|
38
|
+
|
39
|
+
def reset(self):
|
40
|
+
self._mCurrentState = eStart
|
41
|
+
|
42
|
+
def next_state(self, c):
|
43
|
+
# for each byte we get its class
|
44
|
+
# if it is first byte, we also get byte length
|
45
|
+
# PY3K: aBuf is a byte stream, so c is an int, not a byte
|
46
|
+
byteCls = self._mModel['classTable'][wrap_ord(c)]
|
47
|
+
if self._mCurrentState == eStart:
|
48
|
+
self._mCurrentBytePos = 0
|
49
|
+
self._mCurrentCharLen = self._mModel['charLenTable'][byteCls]
|
50
|
+
# from byte's class and stateTable, we get its next state
|
51
|
+
curr_state = (self._mCurrentState * self._mModel['classFactor']
|
52
|
+
+ byteCls)
|
53
|
+
self._mCurrentState = self._mModel['stateTable'][curr_state]
|
54
|
+
self._mCurrentBytePos += 1
|
55
|
+
return self._mCurrentState
|
56
|
+
|
57
|
+
def get_current_charlen(self):
|
58
|
+
return self._mCurrentCharLen
|
59
|
+
|
60
|
+
def get_coding_state_machine(self):
|
61
|
+
return self._mModel['name']
|
Binary file
|
@@ -0,0 +1,34 @@
|
|
1
|
+
######################## BEGIN LICENSE BLOCK ########################
|
2
|
+
# Contributor(s):
|
3
|
+
# Ian Cordasco - port to Python
|
4
|
+
#
|
5
|
+
# This library is free software; you can redistribute it and/or
|
6
|
+
# modify it under the terms of the GNU Lesser General Public
|
7
|
+
# License as published by the Free Software Foundation; either
|
8
|
+
# version 2.1 of the License, or (at your option) any later version.
|
9
|
+
#
|
10
|
+
# This library is distributed in the hope that it will be useful,
|
11
|
+
# but WITHOUT ANY WARRANTY; without even the implied warranty of
|
12
|
+
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
|
13
|
+
# Lesser General Public License for more details.
|
14
|
+
#
|
15
|
+
# You should have received a copy of the GNU Lesser General Public
|
16
|
+
# License along with this library; if not, write to the Free Software
|
17
|
+
# Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA
|
18
|
+
# 02110-1301 USA
|
19
|
+
######################### END LICENSE BLOCK #########################
|
20
|
+
|
21
|
+
import sys
|
22
|
+
|
23
|
+
|
24
|
+
if sys.version_info < (3, 0):
|
25
|
+
base_str = (str, unicode)
|
26
|
+
else:
|
27
|
+
base_str = (bytes, str)
|
28
|
+
|
29
|
+
|
30
|
+
def wrap_ord(a):
|
31
|
+
if sys.version_info < (3, 0) and isinstance(a, base_str):
|
32
|
+
return ord(a)
|
33
|
+
else:
|
34
|
+
return a
|
Binary file
|
@@ -0,0 +1,39 @@
|
|
1
|
+
######################## BEGIN LICENSE BLOCK ########################
|
2
|
+
# The Original Code is Mozilla Universal charset detector code.
|
3
|
+
#
|
4
|
+
# The Initial Developer of the Original Code is
|
5
|
+
# Netscape Communications Corporation.
|
6
|
+
# Portions created by the Initial Developer are Copyright (C) 2001
|
7
|
+
# the Initial Developer. All Rights Reserved.
|
8
|
+
#
|
9
|
+
# Contributor(s):
|
10
|
+
# Mark Pilgrim - port to Python
|
11
|
+
# Shy Shalom - original C code
|
12
|
+
#
|
13
|
+
# This library is free software; you can redistribute it and/or
|
14
|
+
# modify it under the terms of the GNU Lesser General Public
|
15
|
+
# License as published by the Free Software Foundation; either
|
16
|
+
# version 2.1 of the License, or (at your option) any later version.
|
17
|
+
#
|
18
|
+
# This library is distributed in the hope that it will be useful,
|
19
|
+
# but WITHOUT ANY WARRANTY; without even the implied warranty of
|
20
|
+
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
|
21
|
+
# Lesser General Public License for more details.
|
22
|
+
#
|
23
|
+
# You should have received a copy of the GNU Lesser General Public
|
24
|
+
# License along with this library; if not, write to the Free Software
|
25
|
+
# Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA
|
26
|
+
# 02110-1301 USA
|
27
|
+
######################### END LICENSE BLOCK #########################
|
28
|
+
|
29
|
+
_debug = 0
|
30
|
+
|
31
|
+
eDetecting = 0
|
32
|
+
eFoundIt = 1
|
33
|
+
eNotMe = 2
|
34
|
+
|
35
|
+
eStart = 0
|
36
|
+
eError = 1
|
37
|
+
eItsMe = 2
|
38
|
+
|
39
|
+
SHORTCUT_THRESHOLD = 0.95
|
Binary file
|