package-auto-assembler
Advanced tools
A tool to automate package creation within ci based on just .py and optionally .ipynb file.
package-auto-assembler is a tool that meant to streamline creation of single module packages.
Its primary goal is to automate as many aspects of python package creation as possible,
thereby shortening the development cycle of reusable components and maintaining a high standard of quality for reusable code.
With package-auto-assembler, you can simplify the package creation process to the point where it can be seamlessly triggered within CI/CD pipelines, requiring minimal setup and preparation for new modules.
.py files without maintaining separate requirements.txt.The following includes additional details to how some features of the packages work with examples that involve internal components. Even though using the package this way is very much possible, cli interface is recomended.
from package_auto_assembler import (VersionHandler, \
ImportMappingHandler, RequirementsHandler, MetadataHandler, \
LocalDependaciesHandler, LongDocHandler, SetupDirHandler, \
ReleaseNotesHandler, MkDocsHandler, PackageAutoAssembler, \
DependenciesAnalyser)
Package versioning within paa is done based on semantic versioning.
major.minor.patch
By default, patch is updated, but the minor and major could also be update based on, for example, commit messages or manually from the log file.
Package auto assembler does try to pull latest version from package storage, but in case of failure uses version logs stored in .paa/tracking.
initialize VersionHandler
pv = VersionHandler(
# required
versions_filepath = '../tests/package_auto_assembler/other/lsts_package_versions.yml',
log_filepath = '../tests/package_auto_assembler/other/version_logs.csv',
# optional
default_version = "0.0.1")
add new package
pv.add_package(
package_name = "new_package",
# optional
version = "0.0.1"
)
update package version
pv.increment_patch(
package_name = "new_package"
)
## for not tracked package
pv.increment_patch(
package_name = "another_new_package",
# optional
default_version = "0.0.1"
)
There are no known versions of 'another_new_package', 0.0.1 will be used!
display current versions and logs
pv.get_versions(
# optional
versions_filepath = '../tests/package_auto_assembler/other/lsts_package_versions.yml'
)
{'another_new_package': '0.0.1', 'new_package': '0.0.2'}
pv.get_version(
package_name='new_package'
)
'0.0.2'
pv.get_logs(
# optional
log_filepath = '../tests/package_auto_assembler/other/version_logs.csv'
)
| Timestamp | Package | Version | |
|---|---|---|---|
| 0 | 2024-07-29 03:26:39 | new_package | 0.0.1 |
| 1 | 2024-07-29 03:26:40 | new_package | 0.0.2 |
| 2 | 2024-07-29 03:26:40 | another_new_package | 0.0.1 |
flush versions and logs
pv.flush_versions()
pv.flush_logs()
get latest available version with pip
pv.get_latest_pip_version(package_name = 'package-auto-assembler')
'0.3.1'
Install and import names of dependencies may vary. The mapping files maps import names to install names so that requirements extraction from .py files is possible. Some of the mapping are packaged and would not need to provided, but in case a dependency used within new package was not inluded, it is possible to augment default mapping through .paa/package_mapping.json.
initialize ImportMappingHandler
im = ImportMappingHandler(
# required
mapping_filepath = "../env_spec/package_mapping.json"
)
load package mappings
im.load_package_mappings(
# optional
mapping_filepath = "../env_spec/package_mapping.json"
)
{'PIL': 'Pillow',
'bs4': 'beautifulsoup4',
'fitz': 'PyMuPDF',
'attr': 'attrs',
'dotenv': 'python-dotenv',
'googleapiclient': 'google-api-python-client',
'google_auth_oauthlib': 'google-auth-oauthlib',
'sentence_transformers': 'sentence-transformers',
'flask': 'Flask',
'stdlib_list': 'stdlib-list',
'sklearn': 'scikit-learn',
'yaml': 'pyyaml',
'package_auto_assembler': 'package-auto-assembler',
'git': 'gitpython'}
Maintaining requirements is much simpler, when done automatically based on the .py files.
The actual requirements files is still constructed. Standard libraries are not added, others are added with their versions, if specified. Local files are also used as dependencies, from which imports are extracted as well.
For example:
import os
import pandas
import attr #>=22.2.0
from .components.local_dep import *
produces
pandas
attrs >=22.2.0
yaml
as requirements file, where yaml is extracted from local_dep.py file.
Checking dependecies for vulnerabilities is usefull and it is done with pip audit which is integrated into the paa package and is used by default.
Optional requirements for extras_require could be probided the same way normal requirements are, but each like that contains an import like that should be commented out in a special way, starting with #!, for example:
import os
import pandas
import attr #>=22.2.0
#! import hnswlib #==0.8.0
produces
pandas
attrs >=22.2.0
hnswlib==0.8.0; extra == "hnswlib"
Sometimes automatic translation of import names to install names via package_mapping.json, for packages where these names differ, may not be enough. A manual overwrite can be done with exlusion of some dependencies from automatic extraction pipeline with #- comment next to import and #@ prefix before text that is intended to end up in an equivalent requirements file, for example:
import os
import pandas
import attr #>=22.2.0
import tensorflow #-
#@ tensorflow-gpu
produces
pandas
attrs >=22.2.0
tensorflow-gpu
initialize RequirementsHandler
rh = RequirementsHandler(
# optional/required later
module_filepath = "../tests/package_auto_assembler/other/example_module.py",
package_mappings = {'PIL': 'Pillow',
'bs4': 'beautifulsoup4',
'fitz': 'PyMuPDF',
'attr': 'attrs',
'dotenv': 'python-dotenv',
'googleapiclient': 'google-api-python-client',
'sentence_transformers': 'sentence-transformers',
'flask': 'Flask',
'stdlib_list': 'stdlib-list',
'sklearn': 'scikit-learn',
'yaml': 'pyyaml'},
requirements_output_path = "../tests/package_auto_assembler/other/",
output_requirements_prefix = "requirements_",
custom_modules_filepath = "../tests/package_auto_assembler/dependancies",
python_version = '3.8',
add_header = True
)
list custom modules for a given directory
rh.list_custom_modules(
# optional
custom_modules_filepath="../tests/package_auto_assembler/dependancies"
)
['example_local_dependacy_1', 'example_local_dependacy_2']
check if module is a standard python library
rh.is_standard_library(
# required
module_name = 'example_local_dependacy_1',
# optional
python_version = '3.8'
)
False
rh.is_standard_library(
# required
module_name = 'logging',
# optional
python_version = '3.8'
)
True
extract requirements from the module file
rh.extract_requirements(
# optional
module_filepath = "../tests/package_auto_assembler/other/example_module.py",
custom_modules = ['example_local_dependacy_2', 'example_local_dependacy_1'],
package_mappings = {'PIL': 'Pillow',
'bs4': 'beautifulsoup4',
'fitz': 'PyMuPDF',
'attr': 'attrs',
'dotenv': 'python-dotenv',
'googleapiclient': 'google-api-python-client',
'sentence_transformers': 'sentence-transformers',
'flask': 'Flask',
'stdlib_list': 'stdlib-list',
'sklearn': 'scikit-learn',
'yaml': 'pyyaml'},
python_version = '3.8',
add_header=True
)
(['attrs>=22.2.0'],
['torch<=2.4.1', 'fastapi[all]', 'scikit-learn==1.5.1', 'numpy'])
rh.requirements_list
['attrs>=22.2.0']
rh.optional_requirements_list
['torch<=2.4.1', 'fastapi[all]', 'scikit-learn==1.5.1', 'numpy']
audit dependencies
rh.check_vulnerabilities(
# optional if ran extract_requirements() before
requirements_list = None,
raise_error = True
)
No known vulnerabilities found
rh.vulnerabilities
[]
try:
rh.check_vulnerabilities(
# optional if ran extract_requirements() before
requirements_list = ['attrs>=22.2.0', 'pandas', 'hnswlib==0.7.0'],
raise_error = True
)
except Exception as e:
print(f"Error: {e}")
Found 1 known vulnerability in 1 package
Name Version ID Fix Versions
------- ------- ------------------- ------------
hnswlib 0.7.0 GHSA-xwc8-rf6m-xr86
Error: Found vulnerabilities, resolve them or ignore check to move forwards!
rh.vulnerabilities
[{'name': 'hnswlib',
'version': '0.7.0',
'id': 'GHSA-xwc8-rf6m-xr86',
'fix_versions': None}]
save requirements to a file
rh.write_requirements_file(
# optional/required later
module_name = 'example_module',
requirements = ['### example_module.py', 'attrs>=22.2.0'],
output_path = "../tests/package_auto_assembler/other/",
prefix = "requirements_"
)
read requirements
rh.read_requirements_file(
# required
requirements_filepath = "../tests/package_auto_assembler/other/requirements_example_module.txt"
)
['attrs>=22.2.0']
Since all of the necessary information for building a package needs to be contained within main component .py file, basic metadata is provided with the use of __package_metadata__ dictionary object, defined within that .py file. It is also used as a trigger for package building within paa pipeline.
Even though some general information shared between packages could be provided through general config, but package specific info should be provided through __package_metadata__. It should support most text fields from setup file, but for others the following fields are available:
classifiers: adds classifiers to the general ones from config, unless it's Development Status :: then module level definition will overwrite the one from configextras_require: a dictionary of optional package that wouldn't be installed during normal installation. The key could be used during installation and the value would be a list of dependencies.install_requires : adds requirements to the list read from imports* Note that providing dependencies this way does not check them through pip-audit or translate them through package mapping
initializing MetadataHandler
mh = MetadataHandler(
# optional/required later
module_filepath = "../tests/package_auto_assembler/other/example_module.py"
)
check if metadata is available
mh.is_metadata_available(
# optional
module_filepath = "../tests/package_auto_assembler/other/example_module.py"
)
True
extract metadata from module
mh.get_package_metadata(
# optional
module_filepath = "../tests/package_auto_assembler/other/example_module.py"
)
{'author': 'Kyrylo Mordan',
'author_email': 'parachute.repo@gmail.com',
'version': '0.0.1',
'description': 'A mock handler for simulating a vector database.',
'keywords': ['python', 'vector database', 'similarity search']}
Package auto assembler creates single module packages, meaning that once package is built all of the object are imported from a single place. The packaging tool does allow for local dependecies which are .py files imported from specified dependencies directory and its subfolders. Packaging structure may look like the following:
packaging repo/
└src/
├ <package names>.py
└ components
├local_dependecy.py
└subdir_1
└local_dependency_2.py
During packaging process paa merges main module with its local dependies into a single file.
initializing LocalDependaciesHandler
ldh = LocalDependaciesHandler(
# required
main_module_filepath = "../tests/package_auto_assembler/other/example_module.py",
dependencies_dir = "../tests/package_auto_assembler/dependancies/",
# optional
save_filepath = "./combined_example_module.py"
)
combine main module with dependacies
print(ldh.combine_modules(
# optional
main_module_filepath = "../tests/package_auto_assembler/other/example_module.py",
dependencies_dir = "../tests/package_auto_assembler/dependancies/",
add_empty_design_choices = False
)[0:1000])
"""
Mock Vector Db Handler
This class is a mock handler for simulating a vector database, designed primarily for testing and development scenarios.
It offers functionalities such as text embedding, hierarchical navigable small world (HNSW) search,
and basic data management within a simulated environment resembling a vector database.
"""
import logging
import json
import time
import attr #>=22.2.0
import sklearn
__design_choices__ = {}
@attr.s
class Shouter:
"""
A class for managing and displaying formatted log messages.
This class uses the logging module to create and manage a logger
for displaying formatted messages. It provides a method to output
various types of lines and headers, with customizable message and
line lengths.
"""
# Formatting settings
dotline_length = attr.ib(default=50)
# Logger settings
logger = attr.ib(default=None)
logger_name = attr.ib(default='Shouter')
loggerLvl = attr.ib(default=logging.DEBUG)
log
ldh.dependencies_names_list
['example_local_dependacy_2', 'example_local_dependacy_1', 'dep_from_bundle_1']
save combined module
ldh.save_combined_modules(
# optional
combined_module = ldh.combine_modules(),
save_filepath = "./combined_example_module.py"
)
Package description is based on .ipynb with same name as the .py. By default it is converted to markdown as is, but there is also an option to execute it.
import logging
ldh = LongDocHandler(
# optional/required later
notebook_path = "../tests/package_auto_assembler/other/example_module.ipynb",
markdown_filepath = "../example_module.md",
timeout = 600,
kernel_name = 'python3',
# logger
loggerLvl = logging.DEBUG
)
convert notebook to md without executing
ldh.convert_notebook_to_md(
# optional
notebook_path = "../tests/package_auto_assembler/other/example_module.ipynb",
output_path = "../example_module.md"
)
Converted ../tests/package_auto_assembler/example_module.ipynb to ../example_module.md
convert notebook to md with executing
ldh.convert_and_execute_notebook_to_md(
# optional
notebook_path = "../tests/package_auto_assembler/other/example_module.ipynb",
output_path = "../example_module.md",
timeout = 600,
kernel_name = 'python3'
)
Converted and executed ../tests/package_auto_assembler/example_module.ipynb to ../example_module.md
return long description
long_description = ldh.return_long_description(
# optional
markdown_filepath = "../example_module.md"
)
Packages are created following rather simple sequence of steps. At some point of the process a temporary directory is created to store the following files:
__init__.py is a simple import from a single module<package name>.py is a single module with all of the local dependeciescli.py is optional packaged cli toolroutes.py is optional packaged file with fastapi routesstreamlit.py is optional packaged streamlit appsetup.py is a setup file for making a packageREADME.md is a package description file based on .ipynb fileLICENSE is optional license fileMANIFEST.in is a list of additional files to be included with the packagemkdocs is a folder with built mkdocs site based on optional extra_docs for the module, module docstring and README.mdartifacts contains optional files that would be packaged with the moduletests contains files needed to run tests with pytest.paa.tracking contains tracking files from .paa dir to make each release of the package independent of PPR that released itinitializing SetupDirHandler
sdh = SetupDirHandler(
# required
module_filepath = "../tests/package_auto_assembler/other/example_module.py",
# optional/ required
module_name = "example_module",
metadata = {'author': 'Kyrylo Mordan',
'version': '0.0.1',
'description': 'Example module.',
'long_description' : long_description,
'keywords': ['python']},
license_path = "../LICENSE",
requirements = ['attrs>=22.2.0'],
classifiers = ['Development Status :: 3 - Alpha',
'Intended Audience :: Developers',
'Intended Audience :: Science/Research',
'Programming Language :: Python :: 3',
'Programming Language :: Python :: 3.9',
'Programming Language :: Python :: 3.10',
'Programming Language :: Python :: 3.11',
'License :: OSI Approved :: MIT License',
'Topic :: Scientific/Engineering'],
setup_directory = "./example_setup_dir"
)
create empty setup dir
sdh.flush_n_make_setup_dir(
# optional
setup_directory = "./example_setup_dir"
)
copy module to setup dir
sdh.copy_module_to_setup_dir(
# optional
module_filepath = "./combined_example_module.py",
setup_directory = "./example_setup_dir"
)
copy license to setup dir
sdh.copy_module_to_setup_dir(
# optional
license_path = "../LICENSE",
setup_directory = "./example_setup_dir"
)
create init file
sdh.create_init_file(
# optional
module_name = "example_module",
setup_directory = "./example_setup_dir"
)
create setup file
sdh.write_setup_file(
# optional
module_name = "example_module",
metadata = {'author': 'Kyrylo Mordan',
'version': '0.0.1',
'description': 'Example Module',
'keywords': ['python']},
requirements = ['attrs>=22.2.0'],
classifiers = ['Development Status :: 3 - Alpha',
'Intended Audience :: Developers',
'Intended Audience :: Science/Research',
'Programming Language :: Python :: 3',
'Programming Language :: Python :: 3.9',
'Programming Language :: Python :: 3.10',
'Programming Language :: Python :: 3.11',
'License :: OSI Approved :: MIT License',
'Topic :: Scientific/Engineering'],
setup_directory = "./example_setup_dir"
)
Package versioning could be enhanced with release notes. Since the tool is mainly meant for ci/cd, it takes advantage of commit messages to construct a release note for every version.
Commit history is analysed from the last merge, if nothiong found then the next and the next, until at least one of [<package name>] labels are found within commit messages. They are bundled together to for a note, where each commit message or messages deliminated with ; are turned in a list element. Previos notes are used to establish which part of commit history to use as a starting point.
Commit messages could also be used to increment version by something other then a default patch.
[<package name>][..+] increments patch (default behavior)[<package name>][.+.] increments minor[<package name>][+..] increments major[<package name>][0.1.2] forces specific version 0.1.2* First release within new packaging repo may struggle to extract release note since commit messages are only analysed from merges in the commit history.
rnh = ReleaseNotesHandler(
# path to existing or new release notes file
filepath = '../tests/package_auto_assembler/other/release_notes.md',
# name of label in commit message [example_module] for filter
label_name = 'example_module',
# new version to be used in release notes
version = '0.0.1'
)
No relevant commit messages found!
..trying depth 2 !
No relevant commit messages found!
No messages to clean were provided
overwritting commit messages from example
# commit messages from last merge
rnh.commit_messages
['fixing paa tests',
'fixing paa tests',
'fixing paa tests',
'[package_auto_assembler] increasing default max search depth for commit history to 5',
'fixing mocker-db release notes',
'Update package version tracking files',
'Update README',
'Update requirements']
example_commit_messages = [
'[example_module] usage example for initial release notes; bugfixes for RNH',
'[BUGFIX] missing parameterframe usage example and reduntant png file',
'[example_module][0.1.2] initial release notes handler',
'Update README',
'Update requirements'
]
rnh.commit_messages = example_commit_messages
internal methods that run on intialiazation of ReleaseNotesHandler
# get messages relevant only for label
rnh._filter_commit_messages_by_package()
print("Example filtered_messaged:")
print(rnh.filtered_messages)
# clean messages
rnh._clean_and_split_commit_messages()
print("Example processed_messages:")
print(rnh.processed_messages)
Example filtered_messaged:
['[example_module] usage example for initial release notes; bugfixes for RNH', '[example_module][0.1.2] initial release notes handler']
Example processed_messages:
['usage example for initial release notes', 'bugfixes for RNH', 'initial release notes handler']
get version update from relevant messages
version_update = rnh.extract_latest_version()
print(f"Example version_update: {version_update}")
Example version_update: 0.1.2
get latest version from relevant release notes
latest_version = rnh.extract_latest_version()
print(f"Example latest_version: {latest_version}")
Example latest_version: 0.1.2
augment existing release note with new entries or create new
# augment existing release note with new entries or create new
rnh.create_release_note_entry(
# optional
existing_contents=rnh.existing_contents,
version=rnh.version,
new_messages=rnh.processed_messages
)
print("Example processed_note_entries:")
print(rnh.processed_note_entries)
Example processed_note_entries:
['# Release notes\n', '\n', '### 0.1.2\n', '\n', ' - usage example for initial release notes\n', '\n', ' - bugfixes for RNH\n', '\n', ' - initial release notes handler\n', '\n', '### 0.0.1\n', '\n', ' - initial version of example_module\n']
saving updated relese notes
rnh.existing_contents
['# Release notes\n',
'\n',
'### 0.1.2\n',
'\n',
' - usage example for initial release notes\n',
' - bugfixes for RNH\n',
' - initial release notes handler\n',
'### 0.1.2\n',
'\n',
' - usage example for initial release notes\n',
'\n',
' - bugfixes for RNH\n',
'\n',
' - initial release notes handler\n',
'\n',
'### 0.0.1\n',
'\n',
' - initial version of example_module\n']
rnh.save_release_notes()
# updated content
rnh.get_release_notes_content()
['# Release notes\n',
'\n',
'### 0.1.2\n',
'\n',
' - usage example for initial release notes\n',
'\n',
' - bugfixes for RNH\n',
'\n',
' - initial release notes handler\n',
'\n',
'### 0.0.1\n',
'\n',
' - initial version of example_module\n']
Extracting info from installed dependencies can provide important insight into inner workings of a package and help avoid some of the licenses.
Licenses are extracted from package metadata and normalized for analysis. Missing labels are marked with - and not recognized licenses with unknown.
Information about unrecognized license labels could be provided through .paa/package_licenses json file that would contain install package name and corresponding license label.
da = DependenciesAnalyser(
# optional
package_name = 'mocker-db',
package_licenses_filepath = '../tests/package_auto_assembler/other/package_licenses.json',
allowed_licenses = ['mit', 'apache-2.0', 'lgpl-3.0', 'bsd-3-clause', 'bsd-2-clause', '-', 'mpl-2.0']
)
finding installed packages with a list of tags
da.filter_packages_by_tags(tags=['aa-paa-tool'])
[('comparisonframe', '0.0.0'),
('mocker-db', '0.0.1'),
('package-auto-assembler', '0.0.0'),
('proompter', '0.0.0')]
extracting some metadata from the installed package
package_metadata = da.get_package_metadata(
package_name = 'mocker-db'
)
package_metadata
{'keywords': ['aa-paa-tool'],
'version': '0.0.1',
'author': 'Kyrylo Mordan',
'author_email': 'parachute.repo@gmail.com',
'classifiers': ['Development Status :: 3 - Alpha',
'Intended Audience :: Developers',
'Intended Audience :: Science/Research',
'Programming Language :: Python :: 3',
'Programming Language :: Python :: 3.9',
'Programming Language :: Python :: 3.10',
'Programming Language :: Python :: 3.11',
'License :: OSI Approved :: MIT License',
'Topic :: Scientific/Engineering',
'PAA-Version :: 0.4.3',
'PAA-CLI :: False'],
'paa_version': '0.4.3',
'paa_cli': 'False',
'license_label': 'MIT'}
extracting package requirements
package_requirements = da.get_package_requirements(
package_name = 'mocker-db'
)
package_requirements
['requests',
'attrs >=22.2.0',
'httpx',
'hnswlib ==0.8.0',
'gridlooper ==0.0.1',
'dill ==0.3.7',
'numpy ==1.26.0',
"sentence-transformers ; extra == 'sentence_transformers'"]
extracting tree of dependencies
extracted_dependencies_tree = da.extract_dependencies_tree(
package_name = 'mocker-db'
)
extracted_dependencies_tree
{'requests': {'charset-normalizer': [],
'idna': [],
'urllib3': [],
'certifi': []},
'attrs': {'importlib-metadata': {'zipp': [], 'typing-extensions': []}},
'httpx': {'anyio': {'idna': [],
'sniffio': [],
'exceptiongroup': [],
'typing-extensions': []},
'certifi': [],
'httpcore': {'certifi': [], 'h11': {'typing-extensions': []}},
'idna': [],
'sniffio': []},
'hnswlib': {'numpy': []},
'gridlooper': {'dill': [],
'attrs': {'importlib-metadata': {'zipp': [], 'typing-extensions': []}},
'tqdm': {'colorama': []}},
'dill': [],
'numpy': []}
addding license labels to tree of dependencies
extracted_dependencies_tree_license = da.add_license_labels_to_dep_tree(
dependencies_tree = extracted_dependencies_tree
)
extracted_dependencies_tree_license
{'requests': 'apache-2.0',
'requests.charset-normalizer': '-',
'requests.idna': '-',
'requests.urllib3': '-',
'requests.certifi': 'mpl-2.0',
'attrs': '-',
'attrs.importlib-metadata': '-',
'attrs.importlib-metadata.zipp': '-',
'attrs.importlib-metadata.typing-extensions': '-',
'httpx': '-',
'httpx.anyio': 'mit',
'httpx.anyio.idna': '-',
'httpx.anyio.sniffio': '-',
'httpx.anyio.exceptiongroup': '-',
'httpx.anyio.typing-extensions': '-',
'httpx.certifi': 'mpl-2.0',
'httpx.httpcore': '-',
'httpx.httpcore.certifi': 'mpl-2.0',
'httpx.httpcore.h11': 'mit',
'httpx.httpcore.h11.typing-extensions': '-',
'httpx.idna': '-',
'httpx.sniffio': '-',
'hnswlib': '-',
'hnswlib.numpy': 'bsd-3-clause',
'gridlooper': '-',
'gridlooper.dill': 'bsd-3-clause',
'gridlooper.attrs': '-',
'gridlooper.attrs.importlib-metadata': '-',
'gridlooper.attrs.importlib-metadata.zipp': '-',
'gridlooper.attrs.importlib-metadata.typing-extensions': '-',
'gridlooper.tqdm': '-',
'gridlooper.tqdm.colorama': '-',
'dill': 'bsd-3-clause',
'numpy': 'bsd-3-clause'}
printing extracted tree of dependencies
da.print_flattened_tree(extracted_dependencies_tree_license)
└── requests : apache-2.0
├── charset-normalizer : -
├── idna : -
├── urllib3 : -
└── certifi : mpl-2.0
└── attrs : -
└── importlib-metadata : -
├── zipp : -
└── typing-extensions : -
└── httpx : -
├── anyio : mit
├── idna : -
├── sniffio : -
├── exceptiongroup : -
└── typing-extensions : -
├── certifi : mpl-2.0
├── httpcore : -
├── certifi : mpl-2.0
└── h11 : mit
└── typing-extensions : -
├── idna : -
└── sniffio : -
└── hnswlib : -
└── numpy : bsd-3-clause
└── gridlooper : -
├── dill : bsd-3-clause
├── attrs : -
└── importlib-metadata : -
├── zipp : -
└── typing-extensions : -
└── tqdm : -
└── colorama : -
└── dill : bsd-3-clause
└── numpy : bsd-3-clause
filtering for unexpected licenses in tree of dependencies
allowed_licenses = ['mit', 'apache-2.0', 'lgpl-3.0', 'mpl-2.0', '-']
da.find_unexpected_licenses_in_deps_tree(
tree_dep_license = extracted_dependencies_tree_license,
# optional
allowed_licenses = allowed_licenses,
raise_error = True
)
{'hnswlib': '', 'gridlooper': ''}
└── dill : bsd-3-clause
└── numpy : bsd-3-clause
└── hnswlib :
└── numpy : bsd-3-clause
└── gridlooper :
└── dill : bsd-3-clause
---------------------------------------------------------------------------
Exception Traceback (most recent call last)
Cell In[9], line 3
1 allowed_licenses = ['mit', 'apache-2.0', 'lgpl-3.0', 'mpl-2.0', '-']
----> 3 da.find_unexpected_licenses_in_deps_tree(
4 tree_dep_license = extracted_dependencies_tree_license,
5 # optional
6 allowed_licenses = allowed_licenses,
7 raise_error = True
8 )
File ~/miniforge3/envs/testenv/lib/python3.10/site-packages/package_auto_assembler/package_auto_assembler.py:2670, in DependenciesAnalyser.find_unexpected_licenses_in_deps_tree(self, tree_dep_license, allowed_licenses, raise_error)
2668 if raise_error and out != {}:
2669 self.print_flattened_tree(flattened_dict = out)
-> 2670 raise Exception("Found unexpected licenses!")
2671 else:
2672 self.logger.info("No unexpected licenses found")
Exception: Found unexpected licenses!
The tool allows to make a package with optional cli interfaces. These could be sometimes preferable when a package contains a standalone tool that would be called from script anyway.
All of the cli logic would need to be included within a .py file which should be stored within cli_dir provided in .paa.config.
Dependencies from these files are extracted in the similar manner to the main module.
Tools from main .py file could still be imported like the following:
from package_name.package_name import ToBeImported
The code is wired in setup.py via the following automatically assuming that appropriate file with the same name as the package exists within cli_dir location.
...,
entry_points = {'console_scripts': [
'<package_alias> = package_name.cli:cli']} ,
...
Alias for name could be provided via the following piece of code, defined after imports, otherwise package name would be used.
__cli_metadata__ = {
"name" : <package_alias>
}
Package-auto-assembler tool itself uses click dependency to build that file, use its cli definition as example.
The tool allows to make a package with optional routes for FastAPI application and run them. Each packages can have one routes file where its logic should be defined. Package-auto-assembler itself can combine multiple routes from packages and filepaths into one application.
A .py file with the same name of the package should be stored within api_routes_dir provided in .paa.config.
Dependencies from these files are extracted in the similar manner to the main module.
Tools from main .py file could still be imported like the following:
from package_name.package_name import ToBeImported
Api description, middleware and run parameters could be provided via optional .paa.api.config file, which for example would look like:
DESCRIPTION : {
'version' : 0.0.0
}
MIDDLEWARE : {
allow_origins : ['*']
}
RUN : {
host : 0.0.0.0
}
where DESCRIPTION contains parameters for FastAPI, MIDDLEWARE for CORSMiddleware and RUN for uvicorn.run
The tools allows to make a package with optional streamlit application as interface to the packaged code. Each package can have one streamlit file. Package-auto-assembler itself would then be used to run packaged applications from the package.
A .py file with the same name of the package should be stored within streamlit_dir provided in .paa.config.
Dependencies from these files are extracted in the similar manner to the main module.
Tools from main .py file could still be imported like the following:
from package_name import ToBeImported
Config file with server, theme and other settings can be provided via optional .paa.streamlit.config.
The tool allows to add files to packages that could be accessed from the package or extracted into selected directory.
There are different types of artifacts with a package like this:
.paa.tracking : includes some tracking files for the purposes of the tool, added to every packagemkdocs : optional static mkdocs siteartifacts contains directories, files and links to files copied from artifacts_dir/<package_name> (from .paa.config)tests contains optional directory with tests with pytest, copied from tests_dir/<package_name> (from .paa.config)Tracking files are added automatically of artifacts adding was not turned off. At the moment contains:
.paa.config : config file that specifies how paa show work.paa.version: version of package-auto-assembler that was used for packagingrelease_notes.md : latest release notes for the packageversion_logs.csv : logs for version updates for all packages in the packaging repolsts_package_versions.yml : latests versions of all packages in the packaging repopackage_mapping.json : additional user-provided remapping of package import names to install namespackage_licenses.json : additional user-provided license labels to overwrite detected onesnotebook.ipynb : optional jupyter notebook that was used for package description`User provided artifacts could be provided in two ways:
adding directory, file or link to the file under artifacts_dir/<package_name>
These files would be packaged with the packages, and files from links would be downloaded and packaged as well. To add a link, create a file with .link extension in the name.
adding artifact_urls dictionary to __package_metadata__ within module .py file
Example of __package_metadata__ with these additional dictionary would be:
__package_metadata__ = {
...,
"artifact_urls" : {
'downloaded.md' : 'https://raw.githubusercontent.com/Kiril-Mordan/reusables/refs/heads/main/docs/module_from_raw_file.md',
'downloaded.png' : 'https://raw.githubusercontent.com/Kiril-Mordan/reusables/refs/heads/main/docs/reuse_logo.png'
}
}
where key would contain name of the artifact and value its link.
These files would not be downloaded and only links would be packaged. After package installation both kinds of links could be refreshed/donwloaded using cli interface from package-auto-assembler.
Artifacts packaged within the package could be accessed in two ways:
using package-auto-assembler cli tools to copy a file or files into a selected directory
reading packaged artifacts from the package, by the use of similar code that finds path to installed package in your system:
import importlib
import importlib.metadata
import importlib.resources as pkg_resources
installed_package_path = pkg_resources.files('<package_name>')
file_name_path = None
if 'artifacts' in os.listdir(installed_package_path):
with pkg_resources.path('<package_name>.artifacts', '<file_name>') as path:
file_name_path = path
To add/edit a python package with package-auto-assembler, its building blocks would need to follow requirements mentioned below.
Note: For basic package only main module would be required. More about role of optional files could be seen here.
Main module [required]
Main module is considered to be a <package_name>.py file stored in module_dir (from .paa.config), which can also me called packaging layer.
This is where most code of the package would be, unless local dependencies are used.
Requirements to this file:
__package_metadata__ (a way to provide package metadata)local dependencies stored only in dependencies_dirLocal dependencies
Local dependencies are optional modules that other modules from packaging layer would import from, and are not packaged on their own. These are useful for segmenting codebase into smaller independent components.
They would be stored in subdirectory of module_dir, specified as dependencies_dir in .paa.config.
Requirements for these files:
dependencies_dir or its subdirectoriesCli interface
Command line interface could be optionally defined for a package by placing <package_name>.py file into cli_dir (from .paa.config). More about adding cli could be found here.
Requirements for this file:
click to define cli tools__cli_metadata__ (a way to provide alias for cli tools)from <package_name>.<package_name> import ToBeImportedAPI routes
Routes for FastAPI application could be optionally defined for a package by placing <package_name>.py file into api_routes_dir (from .paa.config). More about adding api routes could be found here.
Requirements for this file:
from fastapi import APIRouterrouter = APIRouter(prefix = "/<package-name>")@router.* to define endpointsfrom <package_name>.<package_name> import ToBeImportedStreamlit app
Streamlit application could be optionally defined for a package by placing <package_name>.py file into streamlit_dir (from .paa.config). More about adding streamlit app could be found here.
Requirements for this file:
streamlitfrom <package_name>.<package_name> import ToBeImportedPackage description
Providing package description is a useful way to show what the package is about to the intended audience. By default, main_module docstring would be used for this purpose and stored alongside the package in a README.md file, as shown here.
Optionally, one could provide additional information that would be appended to the package docstring via <package_name>.ipynb file, placed into example_notebooks_path (from .paa.config). In the early stages of development, this file could both serve as a place where code is developed/tested before a release and a way to show usage examples.
Additional documentation
Package description may not be enough to store documentation. A simple mkdocs static site is built by default for every package with for followig tabs:
intro contains module docstring and pypi link for github + pypi type of packaging repositorydescription contains optional content from <package_name>.ipynb file, placed into example_notebooks_path (from .paa.config)release notes contains release notes assembled based on commit messagesAdditional documentation can be provided via the following files:
<package_name>.drawio file, placed into drawio_dir (from .paa.config). Each tab would be turned into <package_name>-<tab_name>.png, which if not referenced would be presented as a separate tab;.png, .ipynb or .md files placed in extra_docs_dir/<package_name> (from .paa.config). Each file would be turned into a separate tab, but just like .png from drawio, if referenced, would not be presented as a separate tab.Note: During packaging process files derived from <package_name>.drawio conversion and files from extra_docs_dir/<package_name> would both be placed into .paa/docs with <package_name>-* prefix. Make sure not to name drawio tabs and extra docs with the same names.
Packaging files alongside code and documentation could be achieved by placing files into artifacts_dir/<package_name> directory (from .paa.config). These files could be referenced from within the package and there is an option to provide links instead of physical files. More about adding artifacts to the package could be found here.
Tests that would be used in ci/cd pipeline are expected to be placed into tests_dir<package_name> directory (from .paa.config). These would be copied into the package as well.
Main wrapper for the package integrates described above components into a class that could be used to build package building pipelines within python scripts.
To simplify usage cli interface is recomended instead.
initializing PackageAutoAssembler
paa = PackageAutoAssembler(
# required
module_name = "example_module",
module_filepath = "../tests/package_auto_assembler/other/example_module.py",
# optional
mapping_filepath = "../env_spec/package_mapping.json",
licenses_filepath = "../tests/package_auto_assembler/other/package_licenses.json",
allowed_licenses = ['mit', 'apache-2.0', 'lgpl-3.0', 'mpl-2.0', '-'],
dependencies_dir = "../tests/package_auto_assembler/dependancies/",
example_notebook_path = "./mock_vector_database.ipynb",
versions_filepath = '../tests/package_auto_assembler/other/lsts_package_versions.yml',
log_filepath = '../tests/package_auto_assembler/other/version_logs.csv',
setup_directory = "./example_module",
release_notes_filepath = "../tests/package_auto_assembler/other/release_notes.md",
license_path = "../LICENSE",
license_label = "mit",
classifiers = ['Development Status :: 3 - Alpha',
'Intended Audience :: Developers',
'Intended Audience :: Science/Research',
'Programming Language :: Python :: 3',
'Programming Language :: Python :: 3.9',
'Programming Language :: Python :: 3.10',
'Programming Language :: Python :: 3.11',
'License :: OSI Approved :: MIT License',
'Topic :: Scientific/Engineering'],
requirements_list = [],
execute_readme_notebook = True,
python_version = "3.8",
version_increment_type = "patch",
default_version = "0.0.1",
check_vulnerabilities = True,
check_dependencies_licenses = False,
add_requirements_header = True
)
add metadata from module
paa.add_metadata_from_module(
# optional
module_filepath = "../tests/package_auto_assembler/other/example_module.py"
)
Adding metadata ...
add or update version
paa.add_or_update_version(
# overwrites auto mode (not suggested)
version_increment_type = "patch",
version = "1.2.6",
# optional
module_name = "example_module",
versions_filepath = '../tests/package_auto_assembler/lsts_package_versions.yml',
log_filepath = '../tests/package_auto_assembler/version_logs.csv'
)
Incrementing version ...
No relevant commit messages found!
..trying depth 2 !
No relevant commit messages found!
..trying depth 3 !
No relevant commit messages found!
..trying depth 4 !
No relevant commit messages found!
..trying depth 5 !
No relevant commit messages found!
No messages to clean were provided
add release notes from commit messages
paa.add_or_update_release_notes(
# optional
filepath="../tests/package_auto_assembler/release_notes.md",
version=paa.metadata['version']
)
Updating release notes ...
prepare setup directory
paa.prep_setup_dir()
Preparing setup directory ...
merge local dependacies
paa.merge_local_dependacies(
# optional
main_module_filepath = "../tests/package_auto_assembler/other/example_module.py",
dependencies_dir= "../tests/package_auto_assembler/dependancies/",
save_filepath = "./example_module/example_module.py"
)
Merging ../tests/package_auto_assembler/other/example_module.py with dependecies from ../tests/package_auto_assembler/dependancies/ into ./example_module/example_module.py
add requirements from module
paa.add_requirements_from_module(
# optional
module_filepath = "../tests/package_auto_assembler/other/example_module.py",
import_mappings = {'PIL': 'Pillow',
'bs4': 'beautifulsoup4',
'fitz': 'PyMuPDF',
'attr': 'attrs',
'dotenv': 'python-dotenv',
'googleapiclient': 'google-api-python-client',
'sentence_transformers': 'sentence-transformers',
'flask': 'Flask',
'stdlib_list': 'stdlib-list',
'sklearn': 'scikit-learn',
'yaml': 'pyyaml',
'git' : 'gitpython'}
)
Adding requirements from ../tests/package_auto_assembler/other/example_module.py
No known vulnerabilities found
paa.requirements_list
['### example_module.py', 'attrs>=22.2.0']
make README out of example notebook
paa.add_readme(
# optional
example_notebook_path = "../tests/package_auto_assembler/other/example_module.ipynb",
output_path = "./example_module/README.md",
execute_notebook=False,
)
Adding README from ../tests/package_auto_assembler/other/example_module.ipynb to ./example_module/README.md
prepare setup file
paa.prep_setup_file(
# optional
metadata = {'author': 'Kyrylo Mordan',
'version': '0.0.1',
'description': 'Example module',
'keywords': ['python'],
'license' : 'mit'},
requirements = ['### example_module.py',
'attrs>=22.2.0'],
classifiers = ['Development Status :: 3 - Alpha',
'Intended Audience :: Developers',
'Intended Audience :: Science/Research',
'Programming Language :: Python :: 3',
'Programming Language :: Python :: 3.9',
'Programming Language :: Python :: 3.10',
'Programming Language :: Python :: 3.11',
'License :: OSI Approved :: MIT License',
'Topic :: Scientific/Engineering'],
cli_module_filepath = "../tests/package_auto_assembler/other/cli.py"
)
Preparing setup file for example-module package ...
make package
paa.make_package(
# optional
setup_directory = "./example_module"
)
Making package from ./example_module ...
CompletedProcess(args=['python', './example_module/setup.py', 'sdist', 'bdist_wheel'], returncode=0, stdout="running sdist\nrunning egg_info\nwriting example_module.egg-info/PKG-INFO\nwriting dependency_links to example_module.egg-info/dependency_links.txt\nwriting entry points to example_module.egg-info/entry_points.txt\nwriting requirements to example_module.egg-info/requires.txt\nwriting top-level names to example_module.egg-info/top_level.txt\nreading manifest file 'example_module.egg-info/SOURCES.txt'\nwriting manifest file 'example_module.egg-info/SOURCES.txt'\nrunning check\ncreating example_module-0.0.0\ncreating example_module-0.0.0/example_module\ncreating example_module-0.0.0/example_module.egg-info\ncopying files to example_module-0.0.0...\ncopying example_module/__init__.py -> example_module-0.0.0/example_module\ncopying example_module/cli.py -> example_module-0.0.0/example_module\ncopying example_module/example_module.py -> example_module-0.0.0/example_module\ncopying example_module/setup.py -> example_module-0.0.0/example_module\ncopying example_module.egg-info/PKG-INFO -> example_module-0.0.0/example_module.egg-info\ncopying example_module.egg-info/SOURCES.txt -> example_module-0.0.0/example_module.egg-info\ncopying example_module.egg-info/dependency_links.txt -> example_module-0.0.0/example_module.egg-info\ncopying example_module.egg-info/entry_points.txt -> example_module-0.0.0/example_module.egg-info\ncopying example_module.egg-info/requires.txt -> example_module-0.0.0/example_module.egg-info\ncopying example_module.egg-info/top_level.txt -> example_module-0.0.0/example_module.egg-info\ncopying example_module.egg-info/SOURCES.txt -> example_module-0.0.0/example_module.egg-info\nWriting example_module-0.0.0/setup.cfg\nCreating tar archive\nremoving 'example_module-0.0.0' (and everything under it)\nrunning bdist_wheel\nrunning build\nrunning build_py\ncopying example_module/example_module.py -> build/lib/example_module\ncopying example_module/__init__.py -> build/lib/example_module\ncopying example_module/setup.py -> build/lib/example_module\ncopying example_module/cli.py -> build/lib/example_module\ninstalling to build/bdist.linux-x86_64/wheel\nrunning install\nrunning install_lib\ncreating build/bdist.linux-x86_64/wheel\ncreating build/bdist.linux-x86_64/wheel/example_module\ncopying build/lib/example_module/example_module.py -> build/bdist.linux-x86_64/wheel/example_module\ncopying build/lib/example_module/__init__.py -> build/bdist.linux-x86_64/wheel/example_module\ncopying build/lib/example_module/setup.py -> build/bdist.linux-x86_64/wheel/example_module\ncopying build/lib/example_module/cli.py -> build/bdist.linux-x86_64/wheel/example_module\nrunning install_egg_info\nCopying example_module.egg-info to build/bdist.linux-x86_64/wheel/example_module-0.0.0-py3.10.egg-info\nrunning install_scripts\ncreating build/bdist.linux-x86_64/wheel/example_module-0.0.0.dist-info/WHEEL\ncreating 'dist/example_module-0.0.0-py3-none-any.whl' and adding 'build/bdist.linux-x86_64/wheel' to it\nadding 'example_module/__init__.py'\nadding 'example_module/cli.py'\nadding 'example_module/example_module.py'\nadding 'example_module/setup.py'\nadding 'example_module-0.0.0.dist-info/METADATA'\nadding 'example_module-0.0.0.dist-info/WHEEL'\nadding 'example_module-0.0.0.dist-info/entry_points.txt'\nadding 'example_module-0.0.0.dist-info/top_level.txt'\nadding 'example_module-0.0.0.dist-info/RECORD'\nremoving build/bdist.linux-x86_64/wheel\n", stderr='warning: sdist: standard file not found: should have one of README, README.rst, README.txt, README.md\n\n/home/kyriosskia/miniconda3/envs/testenv/lib/python3.10/site-packages/setuptools/_distutils/cmd.py:66: SetuptoolsDeprecationWarning: setup.py install is deprecated.\n!!\n\n ********************************************************************************\n Please avoid running ``setup.py`` directly.\n Instead, use pypa/build, pypa/installer or other\n standards-based tools.\n\n See https://blog.ganssle.io/articles/2021/10/setup-py-deprecated.html for details.\n ********************************************************************************\n\n!!\n self.initialize_options()\n')
Package documentation can be presented in a form of mkdocs static site, which could be either served or deployed to something like github packages.
Main module docstring is used as intro package that contains something like optional pypi and license badges. Package description and realease notes are turned into separate tabs. Png with diagrams for example could be provided and displayed as their own separate tabs as well.
The one for this package can be seen here
It can be packaged with the package and be displayed in web browser like documentation for api via <package-name>\docs when using included api handling capabilities.
package_name = "example_module"
module_content = LongDocHandler().read_module_content(filepath=f"../tests/package_auto_assembler/{package_name}.py")
docstring = LongDocHandler().extract_module_docstring(module_content=module_content)
pypi_link = LongDocHandler().get_pypi_badge(module_name=package_name)
docs_file_paths = {
"../example_module.md" : "usage-examples.md",
'../tests/package_auto_assembler/release_notes.md' : 'release_notes.md'
}
mdh = MkDocsHandler(
# required
## name of the package to be displayed
package_name = package_name,
## dictionary of markdown files, with path as keys
docs_file_paths = docs_file_paths,
# optional
## module docstring to be displayed in the index
module_docstring = docstring,
## pypi badge to be displayed in the index
pypi_badge = pypi_link,
## license badge to be displayed in the index
license_badge="[](https://github.com/Kiril-Mordan/reusables/blob/main/LICENSE)",
## name of the project directory
project_name = "temp_project")
mdh.create_mkdocs_dir()
mdh.move_files_to_docs()
mdh.generate_markdown_for_images()
mdh.create_index()
mdh.create_mkdocs_yml()
mdh.build_mkdocs_site()
Created new MkDocs dir: temp_project
Copied ../example_module.md to temp_project/docs/usage-examples.md
Copied ../tests/package_auto_assembler/release_notes.md to temp_project/docs/release_notes.md
index.md has been created with site_name: example-module
mkdocs.yml has been created with site_name: Example module
Custom CSS created at temp_project/docs/css/extra.css
INFO - Cleaning site directory
INFO - Building documentation to directory: /home/kyriosskia/Documents/nlp/reusables/example_notebooks/temp_project/site
INFO - Documentation built in 0.12 seconds
mdh.serve_mkdocs_site()
FAQs
A tool to automate package creation within ci based on just .py and optionally .ipynb file.
We found that package-auto-assembler demonstrated a healthy version release cadence and project activity because the last version was released less than a year ago. It has 1 open source maintainer collaborating on the project.
Did you know?
Socket for GitHub automatically highlights issues in each pull request and monitors the health of all your open source dependencies. Discover the contents of your packages and block harmful activity before you install or update your dependencies.
Security News
PyPI confirms no security flaws were exploited in the Ultralytics supply chain attack and highlights improvements for safer package publishing.
Security News
Research
The Socket Research Team breaks down a malicious wrapper package that uses obfuscation to harvest credentials and exfiltrate sensitive data.
Research
Security News
Attackers used a malicious npm package typosquatting a popular ESLint plugin to steal sensitive data, execute commands, and exploit developer systems.