Crate grounddb 

GroundDB

A schema-driven data layer that uses Markdown files as the source of truth.

GroundDB stores data as plain Markdown files with YAML front matter, organized on disk by configurable path templates. A SQLite system database keeps an index for fast queries, and SQL-defined views are maintained automatically.

Core concepts

• Collections — folders of Markdown files, each file is a document
• Schema — a schema.yaml defining collections, fields, types, and views
• Path templates — map field values to filesystem paths (e.g., posts/{status}/{date}-{title}.md)
• Views — SQL queries over collections, maintained incrementally from the document index
• System database — a SQLite file (_system.db) storing the document index, view cache, and schema history

Usage

use grounddb::Store;

let store = Store::open("data").unwrap();

// Access a collection
let users = store.collection("users").unwrap();
let docs = users.list().unwrap();

For compile-time typed access, use grounddb-codegen to generate Rust structs from your schema.

Feature highlights

• Schema validation with enums, defaults, and required fields
• Referential integrity (error, cascade, nullify, archive)
• Auto-generated IDs (ulid, uuid, nanoid)
• Atomic file writes and batch operations with rollback
• Incremental boot with directory-hash change detection

Re-exports

pub use error::GroundDbError;

pub use error::Result;

pub use schema::SchemaDefinition;

pub use store::Store;

pub use store::Collection;

pub use store::Batch;

pub use store::SubscriptionId;

pub use store::ChangeEvent;

pub use document::Document;

pub use view::ViewEngine;

Modules

document

error

migration

path_template

schema

store

system_db

util

validation

view

watcher