ua-parser · mattrobenolt · Dec 22, 2015 · Dec 21, 2015 · Dec 21, 2015 · Dec 21, 2015
diff --git a/.gitignore b/.gitignore
@@ -1,8 +1,10 @@
 *.pyc
 *.egg-info/
 .eggs/
+.cache/
+.tox/
 build/
 dist/
 tmp/
 regexes.yaml
-regexes.json
+_regexes.py
diff --git a/MANIFEST.in b/MANIFEST.in
@@ -1,2 +1 @@
 include README.md
-include ua_parser/regexes.json
diff --git a/Makefile b/Makefile
@@ -1,24 +1,18 @@
-all: prep test
+all: test
 
-prep:
-	#git submodule update --init
-	#sudo apt-get install python-yaml
-
-test:
-	@#test ! -d tmp && mkdir tmp
-	@export PYTHONPATH=tmp && python setup.py develop -d tmp
+test: clean
+	@mkdir -p tmp
+	@PYTHONPATH=tmp python setup.py develop -d tmp
 	@# run all tests
-	@python ua_parser/user_agent_parser_test.py
+	@PYTHONPATH=tmp python ua_parser/user_agent_parser_test.py
 	@# run a single test
-	@#python ua_parser/user_agent_parser_test.py ParseTest.testStringsDeviceBrandModel
+	@#PYTHONPATH=tmp python ua_parser/user_agent_parser_test.py ParseTest.testStringsDeviceBrandModel
 
 clean:
-	@rm -f ua_parser/user_agent_parser.pyc\
-	   ua_parser/regexes.yaml\
-	   ua_parser/regexes.json
+	@find . -name '*.pyc' -delete
 	@rm -rf tmp\
 	   ua_parser.egg-info\
 	   dist\
 	   build
 
-.PHONY: all prep test clean
+.PHONY: all test clean
diff --git a/setup.py b/setup.py
@@ -1,38 +1,123 @@
 #!/usr/bin/env python
+import os
+from distutils import log
+from distutils.core import Command
 from setuptools import setup
 from setuptools.command.develop import develop as _develop
 from setuptools.command.sdist import sdist as _sdist
 
 
-def install_regexes():
-    print('Copying regexes.yaml to package directory...')
-    import os
-    cwd = os.path.abspath(os.path.dirname(__file__))
-    yaml_src = os.path.join(cwd, 'uap-core', 'regexes.yaml')
-    if not os.path.exists(yaml_src):
-        raise RuntimeError(
-            'Unable to find regexes.yaml, should be at %r' % yaml_src)
+def check_output(*args, **kwargs):
+    from subprocess import Popen
+    proc = Popen(*args, **kwargs)
+    output, _ = proc.communicate()
+    rv = proc.poll()
+    assert rv == 0, output
 
-    print('Converting regexes.yaml to regexes.json...')
-    import json
-    import yaml
-    json_dest = os.path.join(cwd, 'ua_parser', 'regexes.json')
-    with open(yaml_src, 'rb') as fp:
-        regexes = yaml.safe_load(fp)
-    with open(json_dest, "w") as f:
-        json.dump(regexes, f)
+
+class build_regexes(Command):
+    description = 'build supporting regular expressions from uap-core'
+    user_options = [
+        ('work-path=', 'w',
+         'The working directory for source files. Defaults to .'),
+    ]
+
+    def initialize_options(self):
+        self.work_path = None
+
+    def finalize_options(self):
+        if self.work_path is None:
+            self.work_path = os.path.realpath(os.path.join(os.path.dirname(__file__)))
+
+    def run(self):
+        work_path = self.work_path
+        if os.path.exists(os.path.join(work_path, '.git')):
+            log.info('initializing git submodules')
+            check_output(['git', 'submodule', 'init'], cwd=work_path)
+            check_output(['git', 'submodule', 'update'], cwd=work_path)
+
+        yaml_src = os.path.join(work_path, 'uap-core', 'regexes.yaml')
+        if not os.path.exists(yaml_src):
+            raise RuntimeError(
+                'Unable to find regexes.yaml, should be at %r' % yaml_src)
+
+        def force_bytes(text):
+            if text is None:
+                return text
+            return text.encode('utf8')
+
+        import yaml
+        py_dest = os.path.join(work_path, 'ua_parser', '_regexes.py')
+
+        log.info('Compiling regexes.yaml -> _regexes.py')
+        with open(yaml_src, 'rb') as fp:
+            regexes = yaml.safe_load(fp)
+        with open(py_dest, 'wb') as fp:
+            fp.write(b'############################################\n')
+            fp.write(b'# NOTICE: This file is autogenerated from  #\n')
+            fp.write(b'# regexes.yaml. Do not edit by hand,       #\n')
+            fp.write(b'# instead, re-run `setup.py build_regexes` #\n')
+            fp.write(b'############################################\n')
+            fp.write(b'\n')
+            fp.write(b'from __future__ import absolute_import\n')
+            fp.write(b'from .user_agent_parser import (\n')
+            fp.write(b'    UserAgentParser, DeviceParser, OSParser,\n')
+            fp.write(b')\n')
+            fp.write(b'\n')
+            fp.write(b'__all__ = (\n')
+            fp.write(b'    \'USER_AGENT_PARSERS\', \'DEVICE_PARSERS\', \'OS_PARSERS\',\n')
+            fp.write(b')\n')
+            fp.write(b'\n')
+            fp.write(b'USER_AGENT_PARSERS = [\n')
+            for device_parser in regexes['user_agent_parsers']:
+                fp.write(b'    UserAgentParser(\n')
+                fp.write(force_bytes('        %r,\n' % device_parser['regex']))
+                fp.write(force_bytes('        %r,\n' % device_parser.get('family_replacement')))
+                fp.write(force_bytes('        %r,\n' % device_parser.get('v1_replacement')))
+                fp.write(force_bytes('        %r,\n' % device_parser.get('v2_replacement')))
+                fp.write(b'    ),\n')
+            fp.write(b']\n')
+            fp.write(b'\n')
+            fp.write(b'DEVICE_PARSERS = [\n')
+            for device_parser in regexes['device_parsers']:
+                fp.write(b'    DeviceParser(\n')
+                fp.write(force_bytes('        %r,\n' % device_parser['regex']))
+                fp.write(force_bytes('        %r,\n' % device_parser.get('regex_flag')))
+                fp.write(force_bytes('        %r,\n' % device_parser.get('device_replacement')))
+                fp.write(force_bytes('        %r,\n' % device_parser.get('brand_replacement')))
+                fp.write(force_bytes('        %r,\n' % device_parser.get('model_replacement')))
+                fp.write(b'    ),\n')
+            fp.write(b']\n')
+            fp.write(b'\n')
+            fp.write(b'OS_PARSERS = [\n')
+            for device_parser in regexes['os_parsers']:
+                fp.write(b'    OSParser(\n')
+                fp.write(force_bytes('        %r,\n' % device_parser['regex']))
+                fp.write(force_bytes('        %r,\n' % device_parser.get('os_replacement')))
+                fp.write(force_bytes('        %r,\n' % device_parser.get('os_v1_replacement')))
+                fp.write(force_bytes('        %r,\n' % device_parser.get('os_v2_replacement')))
+                fp.write(force_bytes('        %r,\n' % device_parser.get('os_v3_replacement')))
+                fp.write(force_bytes('        %r,\n' % device_parser.get('os_v4_replacement')))
+                fp.write(b'    ),\n')
+            fp.write(b']\n')
+            fp.write(b'\n')
 
 
 class develop(_develop):
     def run(self):
-        install_regexes()
+        self.run_command('build_regexes')
         _develop.run(self)
 
 
 class sdist(_sdist):
-    def run(self):
-        install_regexes()
-        _sdist.run(self)
+    sub_commands = _sdist.sub_commands + [('build_regexes', None)]
+
+
+cmdclass = {
+    'develop': develop,
+    'sdist': sdist,
+    'build_regexes': build_regexes,
+}
 
 
 setup(
@@ -47,13 +132,9 @@ def run(self):
     zip_safe=False,
     url='https://github.com/ua-parser/uap-python',
     include_package_data=True,
-    package_data={'ua_parser': ['regexes.json']},
     setup_requires=['pyyaml'],
     install_requires=[],
-    cmdclass={
-        'develop': develop,
-        'sdist': sdist,
-    },
+    cmdclass=cmdclass,
     classifiers=[
         'Development Status :: 4 - Beta',
         'Environment :: Web Environment',

diff --git a/tox.ini b/tox.ini
@@ -5,4 +5,5 @@ envlist = py26, py27, pypy, py31, py32, py33, py34, py35, docs, pep8, py3pep8
 deps =
     pyyaml
 commands =
+    python setup.py develop
     python ua_parser/user_agent_parser_test.py
Original file line number	Diff line number	Diff line change
		@@ -1,2 +1 @@
		include README.md
		include ua_parser/regexes.json