PipeWhy documentation#
Version: 0.1.0.dev0
Data changed. Find out where, with the evidence to explain what is known.
PipeWhy captures versioned data, compares fingerprints and follows declared lineage. Its reports keep missing observations and uncertain causes visible. Use the dependency-free reference engine or opt into native CPU and experimental CUDA execution.
Note
This is development-candidate documentation. The first release targets accidental drift and corruption detection. It does not provide an adversarial integrity guarantee. Check the package index for available prereleases; the repository and release downloads require repository access.
Getting started
New to PipeWhy? Install the project, capture a local fixture and read your first comparison. Learn what a snapshot can and cannot establish.
User guide
Explore normalization, fingerprints, Enron data, lineage, saved queries and resumable local workers. Follow practical examples and their limits.
API reference
Find signatures and behavior for the Python entry points used throughout the guides, generated from the installed package.
Developer guide
Set up Poetry, run the quality checks, build distributions and documentation, and consult the architecture contracts and qualification records.
Reading offline#
This complete site works from a local folder. The documentation build also creates
dist/pipewhy-docs.zip. Extract it and open index.html; use a local web server
if your browser restricts search on local files. See Building the documentation.
Download the offline documentation ZIP