Implementation Guide

The repository now includes F# implementation projects in addition to the markdown specification. The implementation is intentionally small and follows the schema/spec work rather than replacing it.

Projects

ARC Core User Documentation

The F# object model, YAML codec, graph traversal helpers, fragment selector providers, and table views are documented in the ARC Core user guide. The public model is unified: Dataset carries ARC Core, datamap, and administrative properties rather than splitting them into separate runtime profile models.

SQL Profile

The SQL profile artifacts live in schemas/sql/:

• 001_core.sql contains the current executable SQLite DDL for the process graph profile. Datamap and administrative SQL storage are intentionally out of scope for this transition step.
• seed_example.sql contains a small seeded process graph.
• seeded_core.sqlite is the generated SQLite database.
• design.md explains the relational design.

ProcessCore.SQL inside the consolidated ProcessCore project mirrors the SQL profile:

• Tables.fs defines row types for tabular representation of the underlying processes.
• RowCodecs.fs converts between SqlRow values and row types.
• Repository.fs defines table metadata and CRUD facades.
• Sql.fs defines the portable SQL value shape and ISqliteDriver.
• Platform/Driver.fs records runtime adapter choices.

Runtime Adapters

Runtime-specific SQL drivers live in target-specific files under src/ProcessCore/SQL/:

• .NET uses Microsoft.Data.Sqlite.
• JavaScript uses better-sqlite3 after Fable transpilation.
• Python uses the standard-library sqlite3 module after Fable transpilation.

The JavaScript and Python adapter projects compile as .NET stubs outside their Fable runtime so the solution can build while still exposing useful runtime code after transpilation.

Commands

.\build.cmd BuildSolution
.\build.cmd RunTests

Repo-level Node tooling is already present:

npm run test:js

Python test execution uses uv through the FAKE RunTests target.