Fast JS loading:
https://web.archive.org/web/20200108190917/https://humanwhocodes.com/blog/2009/07/28/the-best-way-to-load-external-javascript/

aiohttp websockets:
https://web.archive.org/web/20200108235147/https://aiohttp.readthedocs.io/en/stable/web_quickstart.html#websockets

aiohttp static files:
https://aiohttp.readthedocs.io/en/stable/web_reference.html#aiohttp.web.UrlDispatcher.add_static

aiohttp client response:
https://aiohttp.readthedocs.io/en/stable/client_reference.html#aiohttp.ClientResponse

WebSocket JS specification:
https://web.archive.org/web/20200109005314/https://html.spec.whatwg.org/multipage/web-sockets.html#dom-websocket

Lazy loading HighCharts:
https://web.archive.org/web/20200109045439/https://www.highcharts.com/stock/demo/lazy-loading

Dynamicly adding data:
https://web.archive.org/web/20200109045531/https://www.highcharts.com/docs/working-with-data/live-data#2-api-methods

uses afterSetExtremes event on xAxis to load data, which then set's it via setData

Stacked area chart:
https://web.archive.org/web/20200109053708/https://www.highcharts.com/demo/area-stacked

Running Python code in browser:
https://github.com/brython-dev/brython

Range series, for showing min/max and avg:
https://www.highcharts.com/docs/chart-and-series-types/range-series

Graphs
------

Note: HighCharts uses milliseconds since unix epoch for time stamps.

(make an option to set the "zero" line, i.e. absolute, or consumption
(i.e. net gen/consumption vs solar)

Proposed graphs:
Daily solar production total
daily consumption total
Per panel stacked production
current production and consumption (5 minute intervals?)

Grid Import/Export
navigation: net import by hour or six hours

Solar Panel
navigation: solar production by hour

Navigation data needs to be generated as part of the index generation.

Maybe heatmap for solar panels

Use removePoint and addPoint: https://web.archive.org/web/20200109083308/https://api.highcharts.com/class-reference/Highcharts.Series.html
to add/remove the end null point and add points as they dynamically arrive

Sample Home power Consumption dataset:
https://archive.ics.uci.edu/ml/datasets/individual+household+electric+power+consumption

List of Solar resources:
https://energydemo.github.io/SolarDatasets/

https://pvoutput.org

WebSocket Definition
--------------------

Messages to the websocket:
<id>
	q	Quit

Messages from the websocket:
<id> <args>

	id	description
	c	Consumption -- current power consumed
		Sent for each update.  units kW.
	ng	Net Grid -- What the current power draw is from the grid
		(negative when sending power).
		Sent for each update.  units kW.
	p	Solar production -- current power produced by the solar panels
		in aggregate.
		Sent for each update.  units kW.

	ov	overview of data, contains gird, production and consumption data.
		following is JSON object

	win	The graph has set a window size.  The two arguments are start
		and end.  If this is the first time, NaN will be sent for both,
		which means to send all data.

	windata	Data in response to a window command.  The format is an object
		w/ the keys, production, grid and consumption.  Each will be a
		list of data points.  Each point is a pair of timestamp in
		miliseconds and the value for the point.  It should include a
		point just before and just after the time range so that lines
		off the graph are rendered properly.

One the first connection, the following messages/data will be sent:
	c, ng, p, ov

The following will be sent as they are updated:
	c, ng, p