Reference

A Mappia map is defined by a query: one JavaScript expression that lists the layers to show, the widgets the user gets and the arithmetic that runs over the pixels. These pages explain how the platform reads that query. They are meant to be read in order inside each section, rather than looked up alphabetically — the generated API reference is the alphabetical list, and this is the explanation of it.

Writing a query

1. Where to find it: layers, settings and links Everything you change in a Mappia map is written in a layer, before the list of layers, or in the map's link. Find the place first: it tells you which part of the API reference to open.
2. How a query is evaluated A query is plain JavaScript - one expression whose value is the list of layers to show. This page is the contract around it: how the platform evaluates it, where state lives, and which parts of it run in a worker instead of on the page.
3. Execution model What the platform does between reading a query and painting a tile — when each callback runs, what waits for what, and which code crosses into the calculation worker.

Model

1. The layer and group model The vocabulary a query is written in — the kinds of layer, how groups nest them, how a calculated layer reads several maps at once, and where the full property list lives.
2. Layer and group property catalogue Every property a layer, a group or a layer-row button accepts, generated from the platform source: the declared API of a query's layer definitions.

Tools

1. The widget markup language The double-brace markup that puts sliders, buttons, pickers and loaders inside a layer's panel - every tag, the parameter grammar, how callbacks are resolved and what each widget's value holds.
2. Widget catalogue Every widget the markup accepts, with its parameters, its runtime methods and the shape of the value it contributes to a calculation - generated from the platform source.

Integration

1. Embedding a map and talking to it How a page embeds a Mappia map in an iframe with mappia_io.js, applies a query into it at runtime, and exchanges messages with the query running inside — the envelope, the handshake, the queueing and the failure modes.

Practical guidance

1. Pitfalls, and properties that do nothing The mistakes that make a query fail silently, the keys that have been copied between queries for years without ever being read, and the handful of parameters the platform accepts but does not honour.

For machines

Everything above is also published as plain text, so an AI assistant or a script can read it without parsing this page:

• /llms.txt The index: every chapter with a one-line summary and a link to its markdown.
• /llms-full.txt The whole reference concatenated into one file.
• /assets/api.json Every documented member of the query API, generated from the platform source.

Each chapter is also served as markdown at /assets/reference/<chapter>.md.