aridity 77__tar.gz → 79__tar.gz

{aridity-77/aridity.egg-info → aridity-79}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: aridity
-Version: 77
+Version: 79
 Summary: DRY config and template system, easily extensible with Python
 Home-page: https://pypi.org/project/aridity/
 Author: foyono
@@ -8,11 +8,46 @@ Author-email: shrovis@foyono.com
 Description-Content-Type: text/markdown
 Requires-Dist: importlib-metadata>=2.1.3
 Requires-Dist: importlib-resources>=3.3.1
-Requires-Dist: pyparsing==2.4.7
+Requires-Dist: pyparsing>=3.0.2
 
 # aridity
 DRY config and template system, easily extensible with Python
 
+## Typical usage
+Main module of dbtool app:
+```
+from aridity.config import ConfigCtrl
+
+def main():
+    # Deduce app name from main function and create scope of that name, load dbtool.arid into that scope, load .settings.arid into the root scope, and return app scope:
+    config = ConfigCtrl().loadappconfig(main, 'dbtool.arid')
+    print(config.dbhost)
+```
+Sibling dbtool.arid of main module, self-documenting that `dbhost` should be configured:
+```
+dbhost = $(void)
+```
+Shared config file `.settings.arid` in home directory:
+```
+dbtool dbhost = anaximander.local
+```
+
+### Integration with argparse
+App config file:
+```
+cli v = $(void)
+verbose = $(cli v)
+```
+Main function:
+```
+def main():
+    config = ConfigCtrl().loadappconfig(main, 'dbtool.arid')
+    parser = ArgumentParser()
+    parser.add_argument('-v', action = 'store_true')
+    parser.parse_args(namespace = config.cli)
+    print(config.verbose)
+```
+
 ## The Arid Manifesto
 * Keys are paths to avoid concatenation
 * It's never necessary to repeat a value
@@ -23,7 +58,7 @@ DRY config and template system, easily extensible with Python
 * Templating using same syntax as expressions
 * Easy to correctly quote/escape values in templates
 * Extensibility via user-defined functions
-* Easy tasks are easy, hard tasks are possible
+* Common tasks are easy, rare tasks are possible
 * Many applications can share one user config
 * Principle of least astonishment driven design
 * Don't make users jump through hoops
@@ -39,28 +74,28 @@ DRY config and template system, easily extensible with Python
 * aridity is a general purpose solution for all the above, also see [soak](https://pypi.org/project/soak/)
 
 ## Config API
-* Normally you pass around a Config object, and application code can get data out via attribute access e.g. config.foo.bar
-    * Here config.foo is also a Config object, a child scope of config named foo
+* Normally you pass around a `Config` object, and application code can get data out via attribute access e.g. `config.foo.bar`
+    * Here `config.foo` is also a Config object, a child scope of `config` named foo
     * The passing around can be taken care of by a dependency injection container such as [diapyr](https://pypi.org/project/diapyr/)
-* Every Config has an associated ConfigCtrl on which Python API such as processtemplate is available
-    * Use negation to get ConfigCtrl when you have a Config e.g. (-config).processtemplate(...)
-    * Use the node attribute to get Config when you have a ConfigCtrl, this is a rare situation in practice
+* Every Config has an associated ConfigCtrl on which Python API such as `processtemplate` is available
+    * Use negation to get ConfigCtrl when you have a Config e.g. `(-config).processtemplate(...)`
+    * Use the `node` attribute to get Config when you have a ConfigCtrl, this is a rare situation in practice
 * When unit testing a class or function that expects a Config object, you can use SimpleNamespace to mock one
 
 ## Guidelines
-* Config files have the extension .arid and templates .aridt
+* Config files have the extension `.arid` and templates `.aridt`
 * A template is simply a file-sized aridity expression
-    * Conventionally the template processor sets " to the appropriate quote function for the file format, e.g. jsonquote for JSON/YAML
-* Instead of adding Python objects to the config in main, it's tidier to use aridity's pyref function to achieve this
+    * Conventionally the template processor sets `"` to the appropriate quote function for the file format, e.g. `jsonquote` for JSON/YAML
+* Instead of adding Python objects to the config in `main`, it's tidier to use aridity's `pyref` function to achieve this
 * When some value needs to be constructed using concatenation, consider whether it would be more tasteful to do this in the config
 
 ## Feature switch
 * Sometimes we want to deploy a change, but something in production isn't ready for that change
 * A feature switch allows deployment to production in this case
-* Add a boolean to the base config (conventionally root.arid) e.g. foo enabled = true
+* Add a boolean to the base config (conventionally root.arid) e.g. `foo enabled = true`
     * This value should be the configuration that we eventually want in all environments
-* In production config, override with foo enabled = false
-* In the code, read config.foo.enabled and enable the change based on this boolean
+* In production config, override with `foo enabled = false`
+* In the code, read `config.foo.enabled` and enable the change based on this boolean
 * The above can now be deployed to all environments, and is not a blocker for other changes
 * Later when production is ready for it, it's a 1 line change to remove the override from production config
 
@@ -604,7 +639,7 @@ Python object in given module with given qualified name. Module may be relative
 ###### pyres
 
 ```python
-def pyres(scope, packageresolvable, nameresolvable)
+def pyres(scope, packageresolvable, nameresolvable, encoding=Text('ascii'))
 ```
 
 Python resource for inclusion with `.` directive.

aridity-77/PKG-INFO → aridity-79/README.md

@@ -1,18 +1,41 @@
-Metadata-Version: 2.1
-Name: aridity
-Version: 77
-Summary: DRY config and template system, easily extensible with Python
-Home-page: https://pypi.org/project/aridity/
-Author: foyono
-Author-email: shrovis@foyono.com
-Description-Content-Type: text/markdown
-Requires-Dist: importlib-metadata>=2.1.3
-Requires-Dist: importlib-resources>=3.3.1
-Requires-Dist: pyparsing==2.4.7
-
 # aridity
 DRY config and template system, easily extensible with Python
 
+## Typical usage
+Main module of dbtool app:
+```
+from aridity.config import ConfigCtrl
+
+def main():
+    # Deduce app name from main function and create scope of that name, load dbtool.arid into that scope, load .settings.arid into the root scope, and return app scope:
+    config = ConfigCtrl().loadappconfig(main, 'dbtool.arid')
+    print(config.dbhost)
+```
+Sibling dbtool.arid of main module, self-documenting that `dbhost` should be configured:
+```
+dbhost = $(void)
+```
+Shared config file `.settings.arid` in home directory:
+```
+dbtool dbhost = anaximander.local
+```
+
+### Integration with argparse
+App config file:
+```
+cli v = $(void)
+verbose = $(cli v)
+```
+Main function:
+```
+def main():
+    config = ConfigCtrl().loadappconfig(main, 'dbtool.arid')
+    parser = ArgumentParser()
+    parser.add_argument('-v', action = 'store_true')
+    parser.parse_args(namespace = config.cli)
+    print(config.verbose)
+```
+
 ## The Arid Manifesto
 * Keys are paths to avoid concatenation
 * It's never necessary to repeat a value
@@ -23,7 +46,7 @@ DRY config and template system, easily extensible with Python
 * Templating using same syntax as expressions
 * Easy to correctly quote/escape values in templates
 * Extensibility via user-defined functions
-* Easy tasks are easy, hard tasks are possible
+* Common tasks are easy, rare tasks are possible
 * Many applications can share one user config
 * Principle of least astonishment driven design
 * Don't make users jump through hoops
@@ -39,28 +62,28 @@ DRY config and template system, easily extensible with Python
 * aridity is a general purpose solution for all the above, also see [soak](https://pypi.org/project/soak/)
 
 ## Config API
-* Normally you pass around a Config object, and application code can get data out via attribute access e.g. config.foo.bar
-    * Here config.foo is also a Config object, a child scope of config named foo
+* Normally you pass around a `Config` object, and application code can get data out via attribute access e.g. `config.foo.bar`
+    * Here `config.foo` is also a Config object, a child scope of `config` named foo
     * The passing around can be taken care of by a dependency injection container such as [diapyr](https://pypi.org/project/diapyr/)
-* Every Config has an associated ConfigCtrl on which Python API such as processtemplate is available
-    * Use negation to get ConfigCtrl when you have a Config e.g. (-config).processtemplate(...)
-    * Use the node attribute to get Config when you have a ConfigCtrl, this is a rare situation in practice
+* Every Config has an associated ConfigCtrl on which Python API such as `processtemplate` is available
+    * Use negation to get ConfigCtrl when you have a Config e.g. `(-config).processtemplate(...)`
+    * Use the `node` attribute to get Config when you have a ConfigCtrl, this is a rare situation in practice
 * When unit testing a class or function that expects a Config object, you can use SimpleNamespace to mock one
 
 ## Guidelines
-* Config files have the extension .arid and templates .aridt
+* Config files have the extension `.arid` and templates `.aridt`
 * A template is simply a file-sized aridity expression
-    * Conventionally the template processor sets " to the appropriate quote function for the file format, e.g. jsonquote for JSON/YAML
-* Instead of adding Python objects to the config in main, it's tidier to use aridity's pyref function to achieve this
+    * Conventionally the template processor sets `"` to the appropriate quote function for the file format, e.g. `jsonquote` for JSON/YAML
+* Instead of adding Python objects to the config in `main`, it's tidier to use aridity's `pyref` function to achieve this
 * When some value needs to be constructed using concatenation, consider whether it would be more tasteful to do this in the config
 
 ## Feature switch
 * Sometimes we want to deploy a change, but something in production isn't ready for that change
 * A feature switch allows deployment to production in this case
-* Add a boolean to the base config (conventionally root.arid) e.g. foo enabled = true
+* Add a boolean to the base config (conventionally root.arid) e.g. `foo enabled = true`
     * This value should be the configuration that we eventually want in all environments
-* In production config, override with foo enabled = false
-* In the code, read config.foo.enabled and enable the change based on this boolean
+* In production config, override with `foo enabled = false`
+* In the code, read `config.foo.enabled` and enable the change based on this boolean
 * The above can now be deployed to all environments, and is not a blocker for other changes
 * Later when production is ready for it, it's a 1 line change to remove the override from production config
 
@@ -604,7 +627,7 @@ Python object in given module with given qualified name. Module may be relative
 ###### pyres
 
 ```python
-def pyres(scope, packageresolvable, nameresolvable)
+def pyres(scope, packageresolvable, nameresolvable, encoding=Text('ascii'))
 ```
 
 Python resource for inclusion with `.` directive.

{aridity-77 → aridity-79}/aridity/functions.py

@@ -221,9 +221,9 @@ class Functions:
             pyobj = getattr(pyobj, name)
         return wrap(pyobj)
 
-    def pyres(scope, packageresolvable, nameresolvable):
+    def pyres(scope, packageresolvable, nameresolvable, encoding = Text('ascii')):
         'Python resource for inclusion with `.` directive.'
-        return Resource(packageresolvable.resolve(scope).cat(), nameresolvable.resolve(scope).cat())
+        return Resource(packageresolvable.resolve(scope).cat(), nameresolvable.resolve(scope).cat(), encoding.resolve(scope).cat())
 
     @realname('\N{NOT SIGN}')
     def not_(scope, resolvable):

{aridity-77 → aridity-79}/aridity/keyring.py

@@ -3,7 +3,7 @@ from .util import null_exc_info
 from base64 import b64decode
 from functools import partial
 from getpass import getpass
-from subprocess import check_output
+from subprocess import CalledProcessError, check_output
 from tempfile import NamedTemporaryFile
 from threading import Semaphore
 import logging, os
@@ -38,9 +38,14 @@ def keyring(scope, serviceres, usernameres):
     password = None if scope.resolved('keyring_force').scalar else get_password(service, username)
     return Scalar(Password(*[getpass(), partial(set_password, service, username)] if password is None else [password, None]))
 
+class DecryptionFailedException(Exception): pass
+
 def gpg(scope, resolvable):
     'Use gpg to decrypt the given base64-encoded blob.'
     with NamedTemporaryFile() as f:
         f.write(b64decode(resolvable.resolve(scope).cat()))
         f.flush()
-        return Scalar(Password(check_output(['gpg', '-d', f.name]).decode('ascii'), None))
+        try:
+            return Scalar(Password(check_output(['gpg', '-d', f.name]).decode('ascii'), None))
+        except CalledProcessError:
+            raise DecryptionFailedException

aridity-77/README.md → aridity-79/aridity.egg-info/PKG-INFO

@@ -1,6 +1,53 @@
+Metadata-Version: 2.1
+Name: aridity
+Version: 79
+Summary: DRY config and template system, easily extensible with Python
+Home-page: https://pypi.org/project/aridity/
+Author: foyono
+Author-email: shrovis@foyono.com
+Description-Content-Type: text/markdown
+Requires-Dist: importlib-metadata>=2.1.3
+Requires-Dist: importlib-resources>=3.3.1
+Requires-Dist: pyparsing>=3.0.2
+
 # aridity
 DRY config and template system, easily extensible with Python
 
+## Typical usage
+Main module of dbtool app:
+```
+from aridity.config import ConfigCtrl
+
+def main():
+    # Deduce app name from main function and create scope of that name, load dbtool.arid into that scope, load .settings.arid into the root scope, and return app scope:
+    config = ConfigCtrl().loadappconfig(main, 'dbtool.arid')
+    print(config.dbhost)
+```
+Sibling dbtool.arid of main module, self-documenting that `dbhost` should be configured:
+```
+dbhost = $(void)
+```
+Shared config file `.settings.arid` in home directory:
+```
+dbtool dbhost = anaximander.local
+```
+
+### Integration with argparse
+App config file:
+```
+cli v = $(void)
+verbose = $(cli v)
+```
+Main function:
+```
+def main():
+    config = ConfigCtrl().loadappconfig(main, 'dbtool.arid')
+    parser = ArgumentParser()
+    parser.add_argument('-v', action = 'store_true')
+    parser.parse_args(namespace = config.cli)
+    print(config.verbose)
+```
+
 ## The Arid Manifesto
 * Keys are paths to avoid concatenation
 * It's never necessary to repeat a value
@@ -11,7 +58,7 @@ DRY config and template system, easily extensible with Python
 * Templating using same syntax as expressions
 * Easy to correctly quote/escape values in templates
 * Extensibility via user-defined functions
-* Easy tasks are easy, hard tasks are possible
+* Common tasks are easy, rare tasks are possible
 * Many applications can share one user config
 * Principle of least astonishment driven design
 * Don't make users jump through hoops
@@ -27,28 +74,28 @@ DRY config and template system, easily extensible with Python
 * aridity is a general purpose solution for all the above, also see [soak](https://pypi.org/project/soak/)
 
 ## Config API
-* Normally you pass around a Config object, and application code can get data out via attribute access e.g. config.foo.bar
-    * Here config.foo is also a Config object, a child scope of config named foo
+* Normally you pass around a `Config` object, and application code can get data out via attribute access e.g. `config.foo.bar`
+    * Here `config.foo` is also a Config object, a child scope of `config` named foo
     * The passing around can be taken care of by a dependency injection container such as [diapyr](https://pypi.org/project/diapyr/)
-* Every Config has an associated ConfigCtrl on which Python API such as processtemplate is available
-    * Use negation to get ConfigCtrl when you have a Config e.g. (-config).processtemplate(...)
-    * Use the node attribute to get Config when you have a ConfigCtrl, this is a rare situation in practice
+* Every Config has an associated ConfigCtrl on which Python API such as `processtemplate` is available
+    * Use negation to get ConfigCtrl when you have a Config e.g. `(-config).processtemplate(...)`
+    * Use the `node` attribute to get Config when you have a ConfigCtrl, this is a rare situation in practice
 * When unit testing a class or function that expects a Config object, you can use SimpleNamespace to mock one
 
 ## Guidelines
-* Config files have the extension .arid and templates .aridt
+* Config files have the extension `.arid` and templates `.aridt`
 * A template is simply a file-sized aridity expression
-    * Conventionally the template processor sets " to the appropriate quote function for the file format, e.g. jsonquote for JSON/YAML
-* Instead of adding Python objects to the config in main, it's tidier to use aridity's pyref function to achieve this
+    * Conventionally the template processor sets `"` to the appropriate quote function for the file format, e.g. `jsonquote` for JSON/YAML
+* Instead of adding Python objects to the config in `main`, it's tidier to use aridity's `pyref` function to achieve this
 * When some value needs to be constructed using concatenation, consider whether it would be more tasteful to do this in the config
 
 ## Feature switch
 * Sometimes we want to deploy a change, but something in production isn't ready for that change
 * A feature switch allows deployment to production in this case
-* Add a boolean to the base config (conventionally root.arid) e.g. foo enabled = true
+* Add a boolean to the base config (conventionally root.arid) e.g. `foo enabled = true`
     * This value should be the configuration that we eventually want in all environments
-* In production config, override with foo enabled = false
-* In the code, read config.foo.enabled and enable the change based on this boolean
+* In production config, override with `foo enabled = false`
+* In the code, read `config.foo.enabled` and enable the change based on this boolean
 * The above can now be deployed to all environments, and is not a blocker for other changes
 * Later when production is ready for it, it's a 1 line change to remove the override from production config
 
@@ -592,7 +639,7 @@ Python object in given module with given qualified name. Module may be relative
 ###### pyres
 
 ```python
-def pyres(scope, packageresolvable, nameresolvable)
+def pyres(scope, packageresolvable, nameresolvable, encoding=Text('ascii'))
 ```
 
 Python resource for inclusion with `.` directive.

{aridity-77 → aridity-79}/aridity.egg-info/requires.txt

@@ -1,3 +1,3 @@
 importlib-metadata>=2.1.3
 importlib-resources>=3.3.1
-pyparsing==2.4.7
+pyparsing>=3.0.2

{aridity-77 → aridity-79}/setup.cfg

@@ -1,5 +1,5 @@
 [bdist_wheel]
-universal = 1
+universal = 0
 
 [egg_info]
 tag_build = 

aridity-79/setup.py

@@ -0,0 +1,62 @@
+from setuptools import setup
+try:
+    from Cython.Build import cythonize
+except ModuleNotFoundError:
+    pass
+from pathlib import Path
+from setuptools import find_packages
+import os
+
+class SourceInfo:
+
+    class PYXPath:
+
+        def __init__(self, module, path):
+            self.module = module
+            self.path = path
+
+        def make_ext(self):
+            g = {}
+            with open(f"{self.path}bld") as f: # Assume project root.
+                exec(f.read(), g)
+            return g['make_ext'](self.module, self.path)
+
+    def __init__(self, rootdir):
+        def addextpaths(dirpath, moduleprefix, suffix = '.pyx'):
+            for name in sorted(os.listdir(os.path.join(rootdir, dirpath))):
+                if name.endswith(suffix):
+                    module = f"{moduleprefix}{name[:-len(suffix)]}"
+                    if module not in extpaths:
+                        extpaths[module] = self.PYXPath(module, os.path.join(dirpath, name))
+        self.packages = find_packages(rootdir)
+        extpaths = {}
+        addextpaths('.', '')
+        for package in self.packages:
+            addextpaths(package.replace('.', os.sep), f"{package}.")
+        self.extpaths = extpaths.values()
+
+    def setup_kwargs(self):
+        kwargs = dict(packages = self.packages)
+        try:
+            kwargs['long_description'] = Path('README.md').read_text()
+            kwargs['long_description_content_type'] = 'text/markdown'
+        except FileNotFoundError:
+            pass
+        if self.extpaths:
+            kwargs['ext_modules'] = cythonize([path.make_ext() for path in self.extpaths])
+        return kwargs
+
+sourceinfo = SourceInfo('.')
+setup(
+    name = 'aridity',
+    version = '79',
+    description = 'DRY config and template system, easily extensible with Python',
+    url = 'https://pypi.org/project/aridity/',
+    author = 'foyono',
+    author_email = 'shrovis@foyono.com',
+    py_modules = ['parabject'],
+    install_requires = ['importlib-metadata>=2.1.3', 'importlib-resources>=3.3.1', 'pyparsing>=3.0.2'],
+    package_data = {'': ['*.pxd', '*.pyx', '*.pyxbld', '*.arid', '*.aridt']},
+    entry_points = {'console_scripts': ['aridity=aridity.__init__:main', 'arid-config=aridity.arid_config:main', 'processtemplate=aridity.processtemplate:main']},
+    **sourceinfo.setup_kwargs(),
+)

aridity-77/setup.py

@@ -1,21 +0,0 @@
-from setuptools import find_packages, setup
-
-def long_description():
-    with open('README.md') as f:
-        return f.read()
-
-setup(
-    name = 'aridity',
-    version = '77',
-    description = 'DRY config and template system, easily extensible with Python',
-    long_description = long_description(),
-    long_description_content_type = 'text/markdown',
-    url = 'https://pypi.org/project/aridity/',
-    author = 'foyono',
-    author_email = 'shrovis@foyono.com',
-    packages = find_packages(),
-    py_modules = ['parabject'],
-    install_requires = ['importlib-metadata>=2.1.3', 'importlib-resources>=3.3.1', 'pyparsing==2.4.7'],
-    package_data = {'': ['*.pxd', '*.pyx', '*.pyxbld', '*.arid', '*.aridt']},
-    entry_points = {'console_scripts': ['aridity=aridity.__init__:main', 'arid-config=aridity.arid_config:main', 'processtemplate=aridity.processtemplate:main']},
-)