API

Only the below API should be considered public.

Enthought egg features

EggMetadata class

This models the metadata of an Enthought egg:

metadata = EggMetadata.from_egg("numpy-1.7.1-1.egg")
print(metadata.platform_tag)

class okonomiyaki.file_formats.EggMetadata(raw_name, version, platform, python, abi_tag, dependencies, pkg_info, summary, metadata_version=None)

Enthought egg metadata for format 1.x.

abi_tag = None

The ABI tag, following the PEP425 format, except that no ABI is sorted as None.

build

The build number.

egg_basename

The egg “base name”, i.e. the name part of the egg filename.

egg_name

The egg filename.

classmethod from_egg(path_or_file, strict=True)

Create a EggMetadata instance from an existing Enthought egg.

path: str or file-like object.

If a string, understood as the path to the egg. Otherwise, understood as a zipfile-like object.

strict: bool

If True, will fail if metadata cannot be decoded correctly (e.g. unicode errors in EGG-INFO/PKG-INFO). If false, will ignore those errors, at the risk of data loss.

metadata_version = None

The version format of the underlying metadata.

name

The package name.

pkg_info = None

A PackageInfo instance modeling the underlying PKG-INFO. May be None for eggs without an PKG-INFO file.

platform = None

The platform, as a Platform instance.

platform_tag

Platform tag following PEP425, except that no platform is represented as None and not ‘any’.

python = None

The python implementation.

runtime_dependencies = None

List of runtime dependencies (as strings).

summary = None

The summary string.

version = None

The version, as an EnpkgVersion instance.

EggBuilder class

This is a class to build Enthought eggs from an install tree

class okonomiyaki.file_formats.EggBuilder(egg_metadata, compress=True, cwd=None)

Class to build eggs from an install tree. This is mostly useful to build Enthought eggs for non-python packages (C/C++ libraries, etc...)

add_data(data, archive_name)

Write the given data as the given archive name.

add_file(path, archive_prefix='')

Add the given file to the egg, under the given archive prefix.

add_file_as(path, archive_name)

Add the given file to the egg, under the given archive name.

add_iterator(iterator)

Add the files specified by the given iterator.

iterator: iterator

An iterator yielding (path, arcname) pairs.

add_tree(directory, archive_prefix='')

Add the given directory to the egg, under the given archive_prefix.

directory: path

A path to a directory. Every file in this directory will be included, recursively.

commit()

Commit the metadata, and close the file.

EggBuilder class

This is a class to build Enthought eggs from an existing setuptools egg.

class okonomiyaki.file_formats.EggRewriter(egg_metadata, egg, compress=True, cwd=None, rename=None, accept=None, allow_overwrite=False)

Class to create Enthought eggs from existing setuptools eggs.

add_data(data, archive_name)

Write the given data as the given archive name.

add_file(path, archive_prefix='')

Add the given file to the egg, under the given archive prefix.

add_file_as(path, archive_name)

Add the given file to the egg, under the given archive name.

add_iterator(iterator)

Add the files specified by the given iterator.

iterator: iterator

An iterator yielding (path, arcname) pairs.

add_tree(directory, archive_prefix='')

Add the given directory to the egg, under the given archive_prefix.

directory: path

A path to a directory. Every file in this directory will be included, recursively.

Runtimes

Runtime metadata factory

This class allows you to parse jaguar runtime format, through its factory class methods.

class okonomiyaki.runtimes.IRuntimeMetadata(metadata_version)

The metadata of a runtime package (i.e. the actual zipfile containing the runtime code).

classmethod factory_from_path(path)

Creates a metadata instance from the given path.

The created instance’s class will be detected dynamically from the metadata file content.

filename

The filename a runtime with this set of metadata.

Platforms representations

The main API is EPDPlatform. You can either create an explicit platform, or try to guess the platform from the running system:

platform = EPDPlatform.from_epd_string("rh5-32")
platform = EPDPlatform.from_running_system()

One can access details through EPDPlatform instances’ attributes.

class okonomiyaki.platforms.EPDPlatform(platform, **kw)

An sane Canopy/EPD platform representation.

Example:

epd_platform = EPDPlatform.from_epd_string("rh5-32")
assert epd.name == "rh5"
assert epd.arch_bits == "32"
assert epd.arch == "x86"

arch_bits

The number of bits (as a string) of this epd platform.

classmethod from_epd_string(s)

Create a new instance from an epd platform string (e.g. ‘win-32’)

Note: new, more explicit architecture names are also supported, e.g. ‘win-x86’ or ‘win-x86_64’ are supported.

classmethod from_running_python()

Attempt to create an EPDPlatform instance by guessing the running python. May raise an OkonomiyakiError exception

classmethod from_running_system(arch_name=None)

Attempt to create an EPDPlatform instance by guessing the running platform. May raise an OkonomiyakiError exception

arch_name: str, None

If given, must be a valid architecture string (e.g. ‘x86’). If None, will be guessed from the running platform.

platform

Main name of the platform (e.g. ‘rh5’)

Generic platform representations are available through the Platform class.

class okonomiyaki.platforms.Platform(os, name, family, arch, machine=None, release='')

An sane generic platform representation.

arch

Actual architecture. The architecture is guessed from the running python.

family

The ‘kind’ of platforms. For example, both debian and ubuntu distributions share the same kind, ‘debian’.

classmethod from_running_python()

Guess the platform, using the running python to guess the architecture.

classmethod from_running_system(arch_string=None)

Guess the platform, with an optional architecture string.

arch_string: str, None

If given, should be a valid architecture name (e.g. ‘x86’)

machine

The machine (e.g. ‘x86’). This is the CPU architecture (e.g. a 32 bits python running on 64 bits Intel OS will be ‘amd64’, whereas arch will be ‘x86’)

name

The most specific platform description

os

The most generic OS description

release

The release string. May be empty

Repositories format

Classes in this module model our different index entries.

GritsEggEntry can be used to automatically create the grits key, tags and metadata from an egg package.

class okonomiyaki.repositories.GritsEggEntry(*args, **kw)

This class models an egg entry metadata as required by Grits.

Example

>>> entry = GritsEggEntry.from_egg("numpy-1.7.1-1.egg", "rh5-32", "free")
# Grits key, metadata and tags are available as simple properties
>>> entry.grits_key
>>> entry.grits_metadata
>>> entry.grits_tags

egg_name

The egg filename.

classmethod from_egg(path, platform, repository_type='commercial', qa_level='stable')

Create a GritsEggEntry from an egg package.

Parameters: str

Path to the egg

platform: str

The consolidated platform (e.g. ‘rh5-32’)

repository_type: str

Type of repository: ‘commercial’ or ‘free’

version

The upstream version (not including the build number).