Cave Utilities

Python utilities for the CAVE App framework: API validation, data builders, and shared helpers.

Overview

cave_utils is a foundational library for building CAVE applications. It provides:

• API validation: validate session_data dicts against the full CAVE API spec, with structured error and warning reporting
• Data builders: helper classes for constructing hierarchical and date-based group structures
• Shared utilities: logging, geographic calculations, coordinate systems, and argument parsing

It is designed to be embedded in a CAVE API server, but can also be used standalone for testing and data development.

Features

• Validates all 10 top-level CAVE API keys (settings, appBar, pages, panes, maps, mapFeatures, globalOutputs, groupedOutputs, draggables, extraKwargs)
• Cross-field validation with structured error paths (e.g. maps.data.myMap.currentProjection)
• Time-series (timeValues) and ordering (order) validation built in
• GroupsBuilder and DateGroupsBuilder for constructing hierarchical group structures from flat data
• GeoUtils for shortest-path calculations over geographic networks
• CustomCoordinateSystem for converting Cartesian 2D/3D coordinates to lat/long
• Runtime type enforcement via type_enforced

Requirements

• Python ≥ 3.11

Installation

pip install cave_utils

Quick Start

Validating Session Data

from cave_utils import Validator

session_data = {
    "settings": {
        "iconUrl": "https://example.com/icon.png",
    },
    "appBar": {
        "data": {
            "btn_home": {
                "icon": "md/MdHome",
                "type": "button",
            }
        }
    },
}

result = Validator(session_data=session_data)

# Inspect errors and warnings
for entry in result.log.log:
    print(entry)
# Empty list means fully valid

Using GroupsBuilder

from cave_utils.builders.groups import GroupsBuilder

builder = GroupsBuilder(
    data=[
        {"id": "region_east", "name": "East", "parent": None},
        {"id": "state_ny", "name": "New York", "parent": "region_east"},
        {"id": "state_ma", "name": "Massachusetts", "parent": "region_east"},
    ],
    id_field="id",
    name_field="name",
    parent_field="parent",
)

# Serialize to the CAVE API groupedOutputs format
groups_data = builder.serialize()

Public API

Full API reference: mit-cave.github.io/cave_utils

LLM Docs

cave_utils can generate plain text documentation for all its modules, suitable for use with AI assistants (Claude, Gemini, etc.).

From Python:

from cave_utils import generate_docs
generate_docs("path/to/output_dir")

From the command line:

cave-utils-docs path/to/output_dir

If no directory is specified, docs are written to ./cave_utils_docs/.

The output directory will contain one .txt file per module plus a README.txt index and a PROJECT_README.md with this project overview. You can point your AI assistant at this directory or reference individual files in a system prompt.

Development

Dev dependencies are declared in [project.optional-dependencies] dev in pyproject.toml. Install them with uv sync --extra dev.

Running Tests

Tests live in test/. The main test file (test_validator.py) imports every example in test/api_examples/, runs it through Validator, and asserts no errors or warnings are produced. Unit tests exist for each module.

uv run pytest

Hot-Reload with a Cave App

To develop cave_utils against a running Cave App, mount the local source as a Docker volume:

cave run --docker-args "--volume {local_path_to_cave_utils}/cave_utils:/cave_utils"

Then in the Cave App's utils/run_server.sh, add pip install -e /cave_utils before the server starts. Changes to cave_utils source will be reflected on the next API call without rebuilding the container.

Set LIVE_API_VALIDATION_PRINT=True in the Cave App's .env to see validation output on every API command.

Release Process

1. Ensure all tests pass and code is formatted:

   uv run nox
   uv run python utils/prettify.py
   

2. Update version in both pyproject.toml and setup.cfg
3. Run tests across all supported Python versions: uv run nox
4. Update VERSION in utils/docs.py and regenerate docs:

   uv run python utils/docs.py