luisy
luisy is a Python framework that extends luigi and further simplifies building data science pipelines by reducing development complexity. Thus, it makes LUIgi more eaSY.
An example pipeline
This is how an end-to-end luisy pipeline may look like:
import luisy
import pandas as pd
@luisy.raw
@luisy.csv_output(delimiter=',')
class InputFile(luisy.ExternalTask):
label = luisy.Parameter()
def get_file_name(self):
return f"file_{self.label}"
@luisy.interim
@luisy.requires(InputFile)
class ProcessedFile(luisy.Task):
def run(self):
df = self.input().read()
# Some more preprocessings
# ...
# Write to disk
self.write(df)
@luisy.final
class MergedFile(luisy.ConcatenationTask):
def requires(self):
for label in ['a', 'b', 'c', 'd']:
yield ProcessedFile(label=label)
Learn more about luisy in our Tutorials.
- Full reference
- luisy package
- Subpackages
- Submodules
- luisy.cli module
- luisy.code_inspection module
ImportReturner
RequirementFileNotFound
VariablesReturner
create_deps_map()
create_hashes()
get_all_deps()
get_all_deps_with_versions()
get_import_data()
get_irregular_pypi_names()
get_node()
get_requirements_dict()
get_requirements_path()
get_varnames_delta()
is_builtin()
is_standard()
make_pypi_mapper()
map_imports_to_requirements()
produce_node_hash()
walk_nodes()
- luisy.config module
- luisy.decorators module
- luisy.default_params module
- luisy.file_system module
- luisy.hashes module
- luisy.helpers module
- luisy.luigi_interface module
- luisy.targets module
- luisy.testing module
- luisy.visualize module
- Module contents
- luisy package