Comparisons

De-Time is designed to sit beside specialist libraries, not erase them. The package is strongest when a workflow needs one decomposition contract, one saved-output story, and one place to switch between Python, CLI, and selected native-backed runs.

Reviewer-grade software matrix

Axis	De-Time	statsmodels	PyEMD	PyWavelets	PySDKit	SSALib	sktime
Core positioning	workflow-oriented decomposition layer	classical decomposition and modeling	EMD-family toolkit	wavelet toolkit	broader signal decomposition toolkit	SSA-focused toolkit	broad time-series ecosystem
Common config object	yes	no	no	no	partial	SSA-specific	no
Common result object	yes	partial	no	no	partial	SSA-specific	no
Batch CLI	yes	no	no	no	limited	no	no
Profiling path	yes	no	no	no	no	no	no
Multivariate under one surface	yes	limited	family-specific	transform-specific	yes	no	partial
Native-backed retained methods	yes	upstream internals	no	mixed	mixed	mixed	mixed
Maturity labeling	explicit	not applicable	family-specific	family-specific	less explicit	focused	ecosystem-level

Where specialist packages are deeper

Package	Where it is deeper	How De-Time positions itself
statsmodels	mature classical decomposition and statistical modeling	De-Time wraps STL and MSTL rather than replacing statsmodels
PyEMD	deeper EMD-family tooling	De-Time exposes EMD and CEEMDAN through the same workflow contract used for other families
PyWavelets	deeper wavelet transforms and transform-specific APIs	De-Time uses wavelet decomposition as one workflow option, not as a claim of wavelet leadership
PySDKit	broader signal-decomposition toolkit, including optional multivariate backends	De-Time uses PySDKit selectively for MVMD and MEMD while keeping a time-series-centered config/result layer
SSALib	deeper SSA-only environment and SSA-specific tooling	De-Time offers SSA inside a broader cross-family package, not as a deeper SSA-only library
sktime	current maintained VMD reality plus a larger time-series transformation ecosystem	De-Time treats VMD as one integrated workflow option and compares against the maintained sktime path rather than the old standalone vmdpy story

Runtime snapshot

The release evidence includes a reproducible runtime snapshot generated by scripts/generate_performance_snapshot.py. The current committed snapshot was generated on Windows 10, Python 3.11.9, AMD64, with native SSA and STD capabilities available.

Method	Python mean runtime (ms)	Native mean runtime (ms)	Speedup
SSA	13.815	1.910	7.232x
STD	0.178	0.036	4.962x
STDR	0.183	0.019	9.599x

These numbers are software-validation evidence, not a universal benchmark claim. The raw evidence lives in docs/assets/generated/evidence/.

Packaging and quality evidence

• Release 0.1.0 is published as de-time and tagged as de-time-v0.1.0.
• The canonical coverage gate applies to the detime core-plus-flagship surface, not to the entire repository tree.
• The current gated coverage snapshot reached 93.20%.
• Wheel and sdist smoke installs, documentation consistency checks, mkdocs build --strict, and twine check are part of the validation story.

What De-Time does not claim

• It does not claim to outperform specialist packages across every task.
• It does not use benchmark leaderboards as the main evidence for the package.
• It does not claim that every wrapped method is equally mature.