discord-metadata

A Discord toolkit for building interactive metadata browsers with embeds and select-menu navigation.

Define your entities once and get Discord embeds with paginated sections, navigable links, and browsable search results. The Discord counterpart of rich-metadata.

Install

Requires Python 3.12+.

pip install discord-metadata
# or
uv add discord-metadata

Quick start

from discord_metadata import (
    BaseNavigator,
    DisplayEngine,
    EntityDef,
    HeaderField,
    HeaderLink,
    MetadataBot,
    SectionDef,
    SummaryField,
    SyncAPI,
    TableColumn,
)

# 1. Define entities
movie_def = EntityDef(
    type_name="movie",
    summary=[
        SummaryField(key="title", bold=True),
        SummaryField(key="year", prefix="(", transform=lambda v: f"{v})"),
    ],
    header_fields=[
        HeaderField("Director", key="director"),
        HeaderField("Year", key="year"),
        HeaderField("Genre", key="genre"),
    ],
    sections=[
        SectionDef(
            "cast",
            navigable=True,
            columns=[
                TableColumn("Actor", "name"),
                TableColumn("Role", "role"),
            ],
        ),
        SectionDef("synopsis"),
    ],
    header_links=[
        HeaderLink("Director: {director}", "person", ref_key="director_id"),
    ],
    title_key="title",
    color=0xE5A00D,
    footer="Movie Database",
    url_key="url",
)

# 2. Create engine and register definitions
engine = DisplayEngine()
engine.register(movie_def)

# 3. Wire up APIs and create bot
navigator = BaseNavigator(
    engine,
    apis={"movie": SyncAPI(MovieAPI())},  # or pass async APIs directly
)

bot = MetadataBot(navigator)

# 4. Register slash commands
@bot.tree.command(name="movie", description="Search for a movie")
async def cmd_movie(interaction, query: str):
    await bot.navigator.search_and_navigate(interaction, query, ["movie"])

# 5. Run
bot.run_with_args("DISCORD_TOKEN")

See examples/movie_bot.py for a complete working example.

Concepts

EntityDef

Declares how an entity type is rendered as a Discord embed:

SectionDef

Each section can be:

• Text — a plain string value rendered in the embed description
• Table — a list of dicts rendered with columns (list of TableColumn)
• Navigable — table rows become selectable items in the select menu
• Lazy — fetched on demand via a lazy_fetchers={(type, section_key): fn} mapping passed to BaseNavigator (sync or async; sync receives the unwrapped API)
• Numbered — items prefixed with 1., 2., etc. (default: True)

DisplayEngine

Registry of entity definitions. Call engine.register(def1, def2, ...) to add definitions.

BaseNavigator

The interactive browser. Main entry points:

SyncAPI

Wraps a synchronous API (with .get(ref) and .search(query) methods) for safe use in async Discord handlers via asyncio.to_thread.

apis = {"book": SyncAPI(BookAPI(client))}

Async APIs can be passed directly without wrapping.

MetadataBot

A ready-made discord.Client subclass that handles:

• Command tree setup and guild sync
• .env loading via python-dotenv (optional)
• --guild GUILD_ID CLI argument for instant command sync during development
• Cleanup callback via on_close

bot = MetadataBot(navigator, on_close=client.close)
bot.run_with_args("DISCORD_TOKEN")

Navigator options

BaseNavigator(
    engine,
    apis={...},
    entity_ref_key="url",       # key used to fetch entity details (default: "url")
    lazy_fetchers={             # optional: (entity_type, section_key) -> fn(api, entity)
        ("movie", "synopsis"): lambda api, entity: api.fetch_synopsis(entity["id"]),
    },
    ephemeral=False,            # whether responses are ephemeral
    placeholder="Browse...",    # select menu placeholder text
    search_kwargs={},           # extra kwargs passed to search methods
)

Paginated browsing

For APIs that return paginated results (discover, recent, upcoming, etc.), use browse() with a fetch_page callable:

@bot.tree.command(name="recent")
async def cmd_recent(interaction):
    await interaction.response.defer()
    await navigator.browse(
        interaction,
        fetch_page=lambda start, count: api.fetch_page(start, count),
        title="Recent releases",
    )

fetch_page(start, count) must return (results: list[dict], total: int). It can be sync or async.

Data format

Entity data is passed as plain dicts with a _type discriminator:

{
    "_type": "movie",
    "title": "Blade Runner",
    "year": "1982",
    "director": "Ridley Scott",
    "cast": [
        {"_type": "person", "name": "Harrison Ford", "role": "Deckard"},
    ],
}

Keys prefixed with _ (except _type) are internal and stripped from public output.

Development

git clone https://github.com/gabriel-jung/discord-metadata.git
cd discord-metadata
uv sync

License

MIT