resprpy

resprpy is a complete Python port of the respR 2.3.4 R package (Harianto & Carey, Methods in Ecology and Evolution 2019, doi:10.1111/2041-210X.13162) for processing and analysis of respirometry (oxygen uptake / metabolic rate) data.

The goal is numeric parity with R: every exported function is verified against outputs generated by R 4.5.3 (pytest parity tests, all 57 passing).

🇨🇳 中文版说明见 README.zh-CN.md · Full usage guide (EN): docs/respR_resprpy_guide.en.html · 使用手册（中文）: docs/respR_resprpy_guide.zh.html

Installation

pip install -e .            # or: pip install -e ".[plot,test]"

Dependencies: numpy, scipy (optional: matplotlib, pytest, openpyxl).

Feature parity (all 27 respR exports)

respR (R)	resprpy (Python)	Status
calc_rate	calc_rate	✅ ≤1e-13
calc_rate.int	calc_rate_int	✅ ≤1e-6
calc_rate.bg	calc_rate_bg	✅ ≤1e-6
calc_rate.ft	calc_rate_ft	✅ ≤1e-5
auto_rate	auto_rate	✅ bw.SJ/density bit-exact
auto_rate.int	auto_rate_int	✅ ≤1e-5
adjust_rate	adjust_rate	✅ 6 methods
adjust_rate.ft	adjust_rate_ft	✅ ≤1e-6
select_rate	select_rate	✅ all 34 methods
select_rate.ft	select_rate_ft	✅ ≤1e-5
convert_DO	convert_DO	✅ ≤1e-12
convert_MR	convert_MR	✅ ≤1e-12
convert_rate	convert_rate	✅ ≤1e-12
convert_rate.ft	convert_rate_ft	✅ ≤1e-8
convert_val	convert_val	✅ ≤1e-12
format_time	format_time	✅ exact
inspect	inspect	✅ checks/locs match
inspect.ft	inspect_ft	✅
oxy_crit	oxy_crit	✅ bsr + segmented
import_file	import_file	✅ NeoFox exact (14 parsers)
sim_data	sim_data	✅ structure
subsample	subsample	✅
subset_data	subset_data	✅
select	select	✅
%>%	— (Python chaining)	➖ n/a
test_lin	test_lin	🟡 structure matches
unit_args	unit_args	✅
S3 plot()	plot_inspect plot_calc_rate plot_auto_rate plot_oxy_crit	✅ matplotlib

Highlights

• Segmented (Muggeo 2003) breakpoint analysis ported from the segmented 2.2-1 package including R's MT19937 RNG (bit-exact set.seed sequences, bootstrap-restart search) — crit values match R exactly (e.g. 7.719111…).
• select_rate — all 34 methods verified against R, including error behaviour (methods R rejects raise the same errors).
• convert_* — physical anchors exact (sw_dens=1024.64077347883, molvol=24.03038425, mg/L→%Air=108.4419).
• No C/C++ runtime: the rolling regression, kernel density, marelac physics and segmented algorithm used by respR's dependencies are all inlined into pure numpy/scipy — PyInstaller-friendly.

Tests

python -m pytest tests/ -v        # 57 tests, all green

Reference CSVs under tests/reference/ were generated by R 4.5.3 (reference/gen_refs_*.R); pytest compares element-wise with graded tolerances (coeffs 1e-8, rsq 1e-12, time 1e-9, density 1e-3).

Package structure (vs respR)

See the structure comparison section in the usage guides.

Citation

resprpy is a Python port of the R package respR. If you use resprpy in publications or derived software, please cite the original respR work:

Harianto, J., Carey, N., & Byrne, M. (2019). respR — An R package for the manipulation and analysis of respirometry data. Methods in Ecology and Evolution, 10(6), 912–920. doi:10.1111/2041-210X.13162

• respR source: https://github.com/januarharianto/respR
• respR documentation: https://januarharianto.github.io/respR/
• CRAN: https://cran.r-project.org/package=respR

Machine-readable citation metadata: CITATION.cff.

License

GPL-3 (same as respR). This is a clean-room style port of respR's algorithms and its dependencies' algorithms; distributing it requires GPL-3 compliance.