Skip to content

Mountainash Ecosystem

One expression algebra. Every backend. The whole pipeline.

Mountainash is a composable Python data engineering ecosystem where every capability — expressions, rules, validation, schema conformance, pipelines — compiles to a single algebra that runs natively on Polars, pandas, DuckDB, Snowflake, BigQuery, and every backend that Ibis and Narwhals reach. The backend is a deployment decision, not a code decision.

This collection covers 9 library entries, with an inherited catalogue of 925 concepts across 86 chapters. Five manuals are available; four remain deferred. PointBreak and Wearables publish independently outside this catalogue. Learning graphs and API references accompany the available books, while inherited MicroSim availability varies: some diagrams are static and some simulations are not implemented.

Publishing status: Rules, Core, Data, Settings, and TreeSpec all publish independently per repository. Deferred entries link to source repositories, not unpublished textbooks.

Core Framework

The foundation libraries that everything else builds on.

  • Mountainash Core

    Cross-backend expression and relational pipeline library. Write col(), lit(), when() in Polars-compatible syntax — compile to any backend. 18 terminal output formats.

    15 chapters · 200 concepts

    Open manual

  • Mountainash Data

    Unified database and lakehouse connectivity. Ibis (23+ dialects), DuckDB, Apache Iceberg, Delta Lake — one connection interface for every analytical backend.

    9 chapters · 100 concepts

    Open manual

  • Mountainash Settings

    Typed configuration framework. Layered resolution from YAML/TOML/env vars, secret injection from Vault or SSM before validation, auto-derived database connection parameters.

    9 chapters · 110 concepts

    Open manual

Rules & Logic

Business rules engines and data transformation.

  • Mountainash Rules

    Vectorised backend-agnostic business rules engine. Decision tables with 11 match strategies, accumulator for combinatorial problems, per-rule explainability.

    9 chapters · 90 concepts

    Open manual

  • Mountainash Rules Babel

    Format translation for decision tables. Import from CSV, JDM, Drools. Export to DMN 1.3 with auto-generated FEEL expressions for regulatory compliance.

    9 chapters · 70 concepts

    Textbook deferred — unavailable. Source repository

  • Mountainash TreeSpec

    Bidirectional hierarchical data transformation. Flatten XML, JSON, Protobuf, Avro, and Pydantic into relational tables — and rebuild them with round-trip fidelity.

    9 chapters · 100 concepts

    Open manual

Connectivity

API integration and authentication.

  • Mountainash API Client

    Declarative REST API integration. Define endpoints as typed dataclasses — the framework derives auth, pagination, retry, and rate-limiting.

    9 chapters · 80 concepts

    Textbook deferred — unavailable. Source repository (now mountainash-http-client)

  • Mountainash Auth Client

    Type-safe multi-provider authentication. OAuth 2.0, OIDC, API keys, HMAC, and 8 more auth methods via a discriminated-union configuration model.

    8 chapters · 85 concepts

    Textbook deferred — unavailable. Source repository

Utilities

Shared infrastructure libraries.

  • Mountainash Transport

    Cloud-native storage abstraction. Local filesystem, S3, GCS, and Azure Blob with a unified interface for read, write, list, and streaming operations.

    9 chapters · 90 concepts

    Textbook deferred — unavailable. Source repository

Where to Start

If you're new to the ecosystem, start with Mountainash Core — it introduces the expression and relation abstractions that every other library builds on. From there, branch into whichever domain you need:

                    Mountainash Core
                    /       |       \
                   /        |        \
         Data + Settings   Rules    API Client + Auth
              |              |
         Transport      Rules Babel
                         TreeSpec

Each manual is self-contained, but cross-references core concepts where relevant. See the About page for persona-based guidance on which packages to start with.