Lectern is Overture's Data Dictionary Schema Manager, providing a system for defining Schemas that will validate the structured data collected by an application. The core of Lectern is a web-server application that handles storage and version management of data dictionaries. Lectern data dictionaries are collections of schemas that define the structure of tabular data files (like TSV). This application provides functionality to validate the structure of data dictionaries, maintain a list of dictionary versions, and to compute the difference between dictionary versions.
This repository is organized as a monorepo using pnpm-workspace
and nx
.
Note:
You will need to use
pnpm
instead ofnpm
to manage dependencies in this code base. PNPM will take care of linking all modules together correctly.
The repository is organized with the following directory structure:
.
├── apps/
│ └── server
└── packages/
├── client
├── common
├── dictionary
└── validation
The modules in the monorepo are organized into two categories:
Component | Type | Package Name | Path | Published Location | Description |
---|---|---|---|---|---|
Lectern Server | Application | server | apps/server/ | Lectern Server web application. | |
Lectern Client | Package | @overture-stack/lectern-client | packages/client | TypeScript Client to interact with Lectern Server and perform data validation. | |
common | Package | common | packages/common/ | N/A | Non-specific but commonly reusable utilities. Includes shared Error classes. |
dictionary | Package | dictionary | packages/dictionary/ | N/A | Dictionary meta-schema definition, includes TS types, and Zod schemas. This also exports all utilities for getting the diff of two dictionaries, and for validating data records with a Dictionary. |
validation | Package | @overture-stack/lectern-validation | packages/validation/ | Validate data using Lectern Dictionaries. |
You can install all dependencies for the entire repo from the root (as defined the pnpm-lock.yaml
) with the command:
pnpm install
Using nx
will ensure all local dependencies are built, in the correct sequence, when building, running, or testing any of the applications and packages in the repo. To run a package.json script from any module - after installing dependencies - use a command of the form pnpm nx <script> <package name>
. For example, to build
the module server
can be done with the command:
pnpm nx build server
This will ensure that all dependencies of server
are built in correct order before the server
build is run.
Note that the full name of from the package must be used for this to work, so for the client the command would be:
pnpm nx build @overture-stack/lectern-client
For convenience, scripts have been added to the root level package.json
to run build
and test
scripts for every service using short names. These follow the pattern pnpm <build|test>:<package short name/alias>
. For example, the same build command can be performed by:
pnpm build:client
To work with any module in this repository, follow the instructions in the README provide in that module's directory.
Get started by running the Lecter Server application.
A few commonly reused scripts have been added to the root package.json
. Run them from the root directory, or if you are in a sub directory then use pnpm -w
.
For example, when your current working directory is not in the project root, you can still conveniently test every module in the monorepo with the command:
pnpm -w test:all
pnpm build:all
This will build all modules.
pnpm test:all
This will test everything, building all dependencies needed to fully test.
In addition to the code for Lectern, this repository contains some useful reference material.
Lectern provides a meta-schema definition that describes the structure of Lectern Dictionaries. The generated JSON Schema formatted copy of this schema can be found at ./generated/DictionaryMetaSchema.json
.
This can be used as a programing language agnostic schema for external applications that want to validate, generate, or interact with Lectern Dictionaries.
Note:
Don't manually update any files in the
./generated
path. This content is programatically generated from the source code.
We welcome community contributions! Please follow our code of conduct