Skip to content
This repository has been archived by the owner on Sep 16, 2021. It is now read-only.

API to fetch weather maps from the Met Office using their DataPoint web services. Written in PHP.

License

Notifications You must be signed in to change notification settings

waynedgrant/met-office-weather-maps

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

74 Commits
 
 
 
 
 
 
 
 
 
 

Repository files navigation

met-office-weather-maps

Copyright © 2015 Wayne D Grant

Licensed under the MIT License

API to fetch weather maps from the Met Office using their DataPoint web services. Written in PHP.

Overview

The Met Office provide the DataPoint API which exposes web services that, among other capabilities, allows clients to request a wide selection of time series weather maps for the UK as images.

These maps make a great addition to any UK-centric weather website. However, there are several common issues the API's users may have to overcome:

  1. Most of the map images are simple layers with no UK map. Users have to supply their own UK base and/or overlay map images that conform to the required boundary box.
  2. Most of the map images have no visual indication as to what time they are relevant for.
  3. DataPoint has fair use call limits and map types can have very different update schedules. Users have to create call schedules that keep their website's maps up-to-date but do not exceed the fair use call limit.

met-office-weather-maps solves each of these issues:

  1. Map images available only as simple layers are processed by met-office-weather-maps which adds appropriate UK base and/or overlay images for the required boundary box.
  2. Where map images are not timestamped met-office-weather-maps automatically adds the relevant UK date/time to the top-left corner.
  3. met-office-weather-maps queries DataPoint's capabilities API and caches the timestamps of all web service calls so that it only fetches map images when new versions are available. This greatly reduces the number of DataPoint web service calls even if met-office-weather-maps is called on a frequent schedule.

Examples

Surface Pressure Forecast Map

Raw layer fetched from DataPoint:

Surface Pressure Forecast Map - Raw Layer

met-office-weather-maps processed image:

Surface Pressure Forecast Map - Processed Image

Visible Satellite Observation Map

Raw layer fetched from DataPoint:

Visible Satellite Observation Map - Raw Layer

met-office-weather-maps processed image:

Visible Satellite Observation Map - Processed Image

Temperature Forecast Map

Raw layer fetched from DataPoint:

Temperature Forecast Map - Raw Layer

met-office-weather-maps processed image:

Temperature Forecast Map - Processed Image

Cloud Cover and Rainfall Forecast Map

Raw layer fetched from DataPoint:

Cloud Cover and Rainfall Forecast Map - Raw Layer

met-office-weather-maps processed image:

Cloud Cover and Rainfall Forecast Map - Processed Image

More Examples

For more examples of met-office-weather-maps in action see http://waynedgrant.com/weather/maps.html.

Requirements

  1. DataPoint API key (available for free)
  2. PHP version 5 or above installed on the web server

Installation

These instructions can be used to set up met-office-weather-maps on a regular schedule on a generic LAMP stack web server.

  • Download the source code for the latest release and unzip it
  • Write code in PHP to fetch the maps you require (see API and Example Code below and pay heed to the advice in DataPoint Fair Use Notes)
  • Upload all files in met-office-weather-maps/src and your code to a directory on your web server
  • Set up a cron schedule to kick off your code regularly (e.g every 15 minutes)

API

The following PHP classes are available with corresponding to a map available from DataPoint:

PHP Class DataPoint Reference
CloudCoverAndRainfallForecastMap N/A
CloudCoverForecastMap Total cloud cover forecast
InfraredSatelliteObservationMap Infrared satellite
LightningObservationMap Lightning strikes
RainfallForecastMap Precipitation forecast
RainfallObservationMap Rainfall radar
SurfacePressureExtendedForecastMap Surface pressure charts
SurfacePressureForecastMap Pressure forecast
TemperatureForecastMap Temperature forecast
VisibleSatelliteObservationMap Visible satellite

All class constructors take the same two mandatory parameters:

  • apiKey Your DataPoint API Key
  • workingFolder The folder to output map images to. Note each class instance must use a different working folder or their outputs will overwrite each other

After construction of any of the classes simply call the fetch() method on it to fetch all images for the given map. After a successful fetch the workingFolder (which will be created automatically if necessary) will contain the following files:

  • 0.png ... n.png or 0.gif ... n.gif - time series images for map in time order
  • info.json - JSON formatted description of the fetched map (see below)

info.json

The info.json for a map can be used to dynamically build an HTML web page to host it using, for example, JQuery. The info.json file contains the following fields:

Field Description
name Map name
datapoint_timestamp Map creation timestamp originally returned by Datapoint
base_time Map base time: YYYY-MM-DD hh:mm UTC
images Array of time series image details (in increasing time order for forecast maps and decreasing time order for observation maps)
images.file Image file name
images.width Image width in pixels
images.height Image height in pixels
images.time Image time: YYYY-MM-DD hh:mm UTC
images.timestep_mins Image timestep relative to base_time in minutes (0 or positive value for forecast maps, 0 or negative value for observation maps)

For example, this snippet of a info.json file:

{
    "name": "Rainfall Observation Map",
    "datapoint_timestamp" : "2015-11-14T08:30:00",
    "base_time": "2015-11-14 08:30 UTC",
    "images":
    [
        {
            "file": "0.png",
            "width": 500,
            "height": 500,
            "time": "2015-11-14 08:30 UTC",
            "timestep_mins": 0
        },
        {
            "file": "1.png",
            "width": 500,
            "height": 500,
            "time": "2015-11-14 08:15 UTC",
            "timestep_mins": -15
        },
        ... etc ...
    ]
}

Example Code

This example code will fetch a selection of maps into separate folders under fetched (replace with your own DataPoint API Key).

<?php

require_once('CloudCoverAndRainfallForecastMap.php');
require_once('InfraredSatelliteObservationMap.php');
require_once('RainfallObservationMap.php');
require_once('SurfacePressureExtendedForecastMap.php');
require_once('TemperatureForecastMap.php');
require_once('VisibleSatelliteObservationMap.php');

define(API_KEY, 'aaaaaaaa-bbbb-cccc-dddd-eeeeeeeeeeee');

$fetchFolderName = dirname(__FILE__ ) . '/fetched/';

$maps = array(
    new CloudCoverAndRainfallForecastMap(API_KEY, $fetchFolderName . 'cloud-rain-fcast'),
    new InfraredSatelliteObservationMap(API_KEY, $fetchFolderName . 'ir-sat-obs'),
    new RainfallObservationMap(API_KEY, $fetchFolderName . 'rain-obs'),
    new SurfacePressureExtendedForecastMap(API_KEY, $fetchFolderName . 'ext-pressure-fcast'),
    new TemperatureForecastMap(API_KEY, $fetchFolderName . 'temp-fcast'),
    new VisibleSatelliteObservationMap(API_KEY, $fetchFolderName . 'vis-sat-obs'));

foreach ($maps as $map) {
    $map->fetch();
}

?>

DataPoint Fair Use Notes

At the time of writing the DataPoint Terms and Conditions state fair use for a single API Key is up to 5,000 web service calls a day and up to 100 calls in a single minute.

Calling code that requests all map types every 15 minutes will result in less than 4,000 daily calls (a smaller number than it would be otherwise because of met-office-weather-map's caching). However, it is possible that the same code could make more than 100 calls in a minute at times when new versions of all maps become available simultaneously and if your web server is especially fast.

Given this I would advise met-office-weather-map's users not to request all maps as above but instead only what they need. This will prevent their DataPoint account from being banned. This should not pose an issue for most use cases given the similarity of some of the maps (e.g. CloudCoverAndRainfallForecastMap vs CloudCoverForecastMap & RainfallForecastMap, SurfacePressureExtendedForecastMap vs SurfacePressureForecastMap).

Image Reference

Base and Overlay Map Images

PHP Class Base Map Image Overlay Map Image
CloudCoverAndRainfallForecastMap uk base colour uk overlay black full outline
CloudCoverForecastMap uk base colour uk overlay black full outline
InfraredSatelliteObservationMap N/A uk overlay yellow cutout outline
LightningObservationMap uk base greyscale N/A
RainfallForecastMap uk base colour uk overlay black full outline
RainfallObservationMap uk base colour uk overlay black full outline
SurfacePressureExtendedForecastMap N/A N/A
SurfacePressureForecastMap uk base colour N/A
TemperatureForecastMap uk base colour uk overlay black full outline
VisibleSatelliteObservationMap blank black uk overlay yellow cutout outline

Output Map Image Formats and Sizes

PHP Class Format Width/Height
CloudCoverAndRainfallForecastMap png 500, 500
CloudCoverForecastMap png 500, 500
InfraredSatelliteObservationMap png 500, 500
LightningObservationMap png 500, 500
RainfallForecastMap png 500, 500
RainfallObservationMap png 500, 500
SurfacePressureExtendedForecastMap gif 891, 601
SurfacePressureForecastMap png 500, 500
TemperatureForecastMap png 500, 500
VisibleSatelliteObservationMap png 500, 500