logsuite 0.2.3

Petrophysical well log analysis with depth-weighted statistics, hierarchical filtering, and template-driven visualization

logSuite

Petrophysical well-log analysis in Python. Load LAS files, set a project palette once, fit per-facies regressions, and export deliverable-ready SVGs — without writing matplotlib glue.

Why logSuite

The library aims to stop you from writing the boilerplate that usually surrounds a reservoir-engineering notebook:

• No per-well loops. manager.PHIE.filter("Facies").data(weighted=True) returns one DataFrame across every well, with depth-interval weights for pooled percentiles.
• No monkey-patching. xplot.add_regression_per("Facies", "exponential") fits one line per facies; the line colours come from manager.Facies.colors automatically.
• No matplotlib glue. xplot.add_table_panel(stats); xplot.save("deliverable.svg") produces the SCAL/DG3 figure as one file.
• No equation rewriting. equation_format="petrel" renders fitted exponentials as pow(10, c1*x + c0) — paste straight into Petrel.

Installation

pip install logsuite

Requires Python 3.10+, numpy, pandas, scipy, matplotlib.

Quick start

from logsuite import Crossplot, WellDataManager, set_quiet

set_quiet(True)                                    # silence broadcast prints
manager = WellDataManager()
manager.load_las("12_3-2_B.las").load_las("12_3-2_C.las")
manager.Facies.colors = {0: "#999999", 1: "#3b82f6", 2: "#10b981"}

xplot = Crossplot(manager, x="PHIE", y="PERM", color="Facies", y_log=True,
                  equation_format="petrel", decimals=3)
xplot.add_regression_per("Facies", "exponential", legend_loc="upper left")
xplot.save("poroperm.svg")

Seven lines. One regression line per facies in the manager palette, legend equations in Petrel syntax, ready for Word.

What's in the box

Class	Role
WellDataManager	Load and orchestrate wells; the data substrate every consumer reads from.
ManagerView	Read-only filtered subset of a manager (manager.filter(wells=, where=)).
Property, Well	Domain objects with depth-aligned arithmetic and chained filtering.
Crossplot	2D scatter with .add_regression(...), .add_regression_per(...), .add(...), .add_table_panel(...).
WellView, Template	Single-well log display configured via a track template.
RegressionFit, Artifact	Renderable result objects added to consumers via .add(). Equations available in natural / log10 / Petrel form.

Common workflows

Each is a runnable script in story_tests/ and a how-to guide in the docs:

Task	How-to	Script
One regression per facies / zone / well	fit-regressions-per-group	story_1_regression_subset.py, story_5_per_group_regressions.py
Project palette flowing into every plot	set-discrete-colors	story_2_property_colors.py
Pooled raw-data DataFrame across wells	pool-data-across-wells	story_3_pooled_data.py, story_6_pooled_extraction.py
Petrel-syntax regression equations	export-petrel-equations	story_4_petrel_equations.py
Crossplot + stats table in one SVG	build-scal-deliverable	story_7_scal_deliverable.py
Single-well log display	display-well-logs	story_8_log_plot.py

Documentation

logsuite.readthedocs.io

• Quick start
• How-to guides — task-focused recipes
• User guide — topical reference
• API reference

Architecture

The library follows a strict layered-dependency rule documented in ARCHITECTURE.md. Lower layers do not know consumers exist; data flows up — never down. The design exists to keep the API small and predictable as features are added.

License

MIT — see LICENSE.