spaday-perspective 0.8.3

Perspective for spaday

Overview

spaday-perspective moves the existing PerspectivePanel integration out of spaday core. Its self-contained <perspective-panel> bundle includes Perspective's client, viewer (the multi-panel workspace element since Perspective 5), datagrid, themes, and viewer WASM.

Documentation

• Build a live market workspace — guided first application.
• Stream data and change layouts — task-focused server and UI recipes.
• Add row buttons — indexed row events and server-side deletion.
• API reference — component props, layout shape, and package descriptor.
• Why Perspective uses its own data channel — integration design and tradeoffs.

Quick example

This complete app serves a live Perspective table. Add trade streams a new row into the grid, the select switches between flat and grouped layouts, and the checkbox changes the viewer theme.

import perspective
from perspective.handlers.starlette import PerspectiveStarletteHandler
from starlette.responses import JSONResponse
from starlette.routing import Route, WebSocketRoute
from starlette.websockets import WebSocket

from spaday import CallEndpoint, cond, element, eq, field, obj
from spaday.backends.starlette import serve
from spaday_perspective import PerspectivePanel

server = perspective.Server()
client = server.new_local_client()
trades = client.table(
    [
        {"id": 1, "symbol": "AAPL", "side": "buy", "price": 211.10},
        {"id": 2, "symbol": "MSFT", "side": "sell", "price": 503.02},
    ],
    name="trades",
)
next_id = 2


async def add_trade(_request):
    global next_id
    next_id += 1
    trades.update(
        [
            {
                "id": next_id,
                "symbol": "AAPL" if next_id % 2 else "MSFT",
                "side": "buy" if next_id % 2 else "sell",
                "price": 210 + next_id,
            }
        ]
    )
    return JSONResponse({"id": next_id})


async def perspective_socket(websocket: WebSocket):
    await PerspectiveStarletteHandler(perspective_server=server, websocket=websocket).run()


def layout(*, grouped=False):
    viewer = {"table": "trades", "plugin": "Datagrid", "columns": ["symbol", "side", "price"]}
    if grouped:
        viewer.update({"group_by": ["symbol"], "columns": ["price"], "aggregates": {"price": "avg"}})
    return {
        "layout": {"type": "tab-layout", "tabs": ["trades"]},
        "panels": {"trades": viewer},
    }


panel = (
    PerspectivePanel()
    .compute("theme", cond(field("dark"), "dark", "light"))
    .compute(
        "config",
        obj(
            {
                "ws_url": "/perspective",
                "tables": ["trades"],
                "layout": cond(eq(field("view"), "grouped"), layout(grouped=True), layout()),
            }
        ),
    )
    .style(display="block", height="70vh", margin_top="1rem")
)

view = element("select").bind("value", "view", mode="two-way").child(
    element("option", "Blotter", value="blotter"),
    element("option", "Average by symbol", value="grouped"),
)
dark = element("input", type="checkbox").bind("checked", "dark", mode="two-way")
page = (
    element("main")
    .style(margin="1rem", font_family="system-ui")
    .child(element("button").text("Add trade").on("click", CallEndpoint("POST", "/api/trades")))
    .child(view)
    .child(element("label").child(dark, " Dark theme"))
    .child(panel)
)

app = serve(
    page,
    packages=["perspective"],
    store={"view": "blotter", "dark": False},
    routes=[
        Route("/api/trades", add_trade, methods=["POST"]),
        WebSocketRoute("/perspective", perspective_socket),
    ],
)

Save this as app.py, then run pip install spaday-perspective perspective-python starlette uvicorn and uvicorn app:app. Open http://127.0.0.1:8000.

This keeps Perspective's bulk data on its native websocket. A spaday/transports model only needs to carry the small serializable connection/layout config. Replacing config reconnects when ws_url changes and restores changed layouts; theme accepts light, dark, or a Perspective theme name.

Installing this project registers the perspective entry point with spaday. The equivalent explicit forms are packages=[spaday_perspective.package] and packages=["spaday_perspective:package"].

Browser example

Run the complete example in your browser — Python runs through Pyodide and Perspective uses its in-browser worker, so no install, server, or websocket is required.

Run the example locally

python -m pip install -e ".[examples]"
python -m spaday_perspective.example

Open http://127.0.0.1:8015 to inspect the complete market-monitor example: live native Perspective streaming, flat and grouped layouts, server-authoritative order submission, reactive metrics, and light/dark themes. It passes the local package descriptor directly, so it does not install or resolve the integration from GitHub.