📈 @smartplant/history

Analyzes stored readings for trends, cycles and slow decline.

A plugin for smartplant — a bridge between AI and plants.

Install

npm install smartplant @smartplant/history

Use

import { createPlant } from 'smartplant'
import history from '@smartplant/history'

const plant = await createPlant( {
  name    : 'Ivy',
  species : 'Monstera deliciosa',
  sensor  : 'mock',                  // no hardware needed
  ai      : { provider : 'ollama' }, // or gemini / openai / claude / grok
} )

await plant.use( history )
await plant.read()

const result = await plant.plugin( 'history' ).analyzeTrends( { hours : 72 } )
console.log( result.advice, result.emoji )

API

analyzeTrends( input? )

Returns { advice, emoji, direction, findings }.

input is optional; anything you pass is handed to the AI as extra context alongside the plant's readings, trends and care history.

If the AI is unreachable or unconfigured, the call returns the deterministic offline result rather than throwing.

Honest about insufficient data

With fewer than three readings in the window it reports direction: 'unknown' instead of inventing a trend.

Raw numbers, no AI

plant.plugin( 'history' ).stats( 72 )    // { soil: { min, max, avg, trend, n }, ... }
plant.plugin( 'history' ).series( 168 )  // the full stored series, for charting

Test

npm test

Runs against the mock sensor and mock AI provider — no hardware, no API key, no network.

📜 License

MIT — see LICENSE.