petekSuite

petekSuite is the umbrella for a subsurface-modelling ecosystem: a small family of Rust libraries (with Python bindings) that take raw subsurface data all the way to static volumes and dynamic forecasts. The suite root itself builds no code — it coordinates the peer libraries, keeps their seams coherent, and holds the shared conventions ("the petek house style") that make them a family rather than a monolith.

Dependencies flow one direction, downward only — no cycles, no sideways sharing of code. Libraries share conventions freely and share code only downward through the graph; each library stays usable standalone.

                          ┌─────────────────────────────────────────┐
   petekIO      DATA      │  ingest · normalize · validate ·         │
      │                   │  interpret → model-ready inputs          │
      ▼                   └─────────────────────────────────────────┘
   petekStatic  GEOMODEL   structural framework · grid · property
      │                    modelling · volumetrics + static uncertainty
      ▼
   petekSim     SIMULATION dynamic / engineering forecast · PVT ·
                           the `peteksim` Python product facade

   petekTools   TOOLKIT    horizontal · shared · domain-agnostic:
                           numeric kernels + units + container + viewer
                           (serves every layer)

The libraries

petekIO — the DATA layer

Ingests, normalizes, validates, and interprets subsurface data (wells, surfaces, logs, seismic) into clean, model-ready inputs. A standalone Rust data-model + IO library with optional PyO3 bindings — data only, no modelling framework. Composes mature crates (las_rs, geo/geozero, ndarray, rstar, giga-segy) behind its own IO traits.

petekStatic — the GEOMODEL layer

Turns model-ready inputs into a populated StaticModel: the structural framework (horizons + faults + zones), grid construction, and property modelling (facies/petrophysics, geostatistics, trend population, log upscaling). It owns volumetrics + static uncertainty — GRV / in-place volumes off the model itself, Monte-Carlo over model realizations, and tornado analysis — producing a StaticModel plus probabilistic (P90/P50/P10) curves.

petekSim — the SIMULATION layer

Dynamic and engineering simulation: recoverable volumes and forecasting (decline, p/z, material balance, and later full dynamic flow), plus PVT. It ships the product — peteksim, the Python-facing appraisal toolkit that presents the whole stack behind one façade.

petekTools — the horizontal TOOLKIT

Domain-agnostic, shared numerics that serve every layer: the scattered-data gridding / kriging / warm-start / geostatistics kernels Rust lacks, a units system, the liftable container format, and a generic bundle viewer. A pure leaf — it depends on none of the others; they build on it.

Install

The whole suite installs with one command:

pip install peteksuite     # the meta-package: pulls all four libraries

The umbrella package also exposes short notebook imports:

from peteksuite import pio, pto, pst, ps

Every library is also directly installable from both crates.io (Rust) and PyPI (Python) — one crate and one wheel per library:

Rust (crates.io)

cargo add petektools@0.2.15    # TOOLKIT — numeric kernels, units, container
cargo add petekio@0.3.15       # DATA    — ingest + model-ready inputs
cargo add petekstatic@0.1.13   # GEOMODEL — StaticModel build + volumetrics + MC
cargo add peteksim@0.1.13      # SIMULATION — the appraisal facade over the stack

Python (PyPI)

pip install petektools==0.2.15
pip install petekio==0.3.15
pip install petekstatic==0.1.13   # static workflow API, StaticModel, volumes + bundles
pip install peteksim==0.1.13      # the full appraisal facade over the whole stack

Dependencies resolve automatically in DAG order (peteksim pulls the stack; petekstatic pulls petekio + petektools).

License

The suite is split by layer:

The horizontal toolkit and the data layer are permissively licensed (Apache-2.0). The geomodel and simulation layers ship under the Business Source License 1.1, each released version converting to Apache-2.0 on its change date. See each library's own LICENSE / NOTICE for the authoritative terms.