content from XML

Developer's Corner General Dev Discussions
#1

So i’m trying to get specific values from a xml file but after days of searching I’m none the wiser

This is the source XML

[code]

UPS Status

UPS Status

      <h2 class="expandable ">- System Status</h2>
      <div class="label-value-group ui-corner-all">
    
    
    <div class="label-value-pair">
      <label>Device Status</label>
      
        <div id="value_DeviceStatus" class="value">On line (Green Mode)</div>
        
    </div>
    
    
    
    <div class="label-value-pair">
      <label>UPS Load</label>
      
        <div id="value_RealPowerPct" class="value">20.8</div>
        <div class="unit">%</div>
    </div>
    
    
    
    <div class="label-value-pair">
      <label>Runtime Remaining</label>
      
        <div id="value_RuntimeRemaining" class="value">49</div>
        <div class="unit">Minutes</div>
    </div>
    
    </div>

</div>
There is a problem retrieving information for this dialog. Please close this dialog and try again. Close
[/code]

I first thought LXP and xpatch would be the cleanest way to do it so I stole this code

[code]local socket = require(“socket”)
local ltn12 = require(‘ltn12’)
local https = require(“ssl.https”)
local lom = require’lxp.lom’
local xpath = require’xpath’

function https.get(u)
local response_body = { }
local rBody, code, response_headers, rStatus = https.request{
url = u,
method = “GET”,
verify = “none”,
mode = “client”,
options = “all”,
protocol = “sslv3”,
headers = {
[“Accept”] = “text/html”,
[“Accept-Encoding”] = “sdch”,
},
sink = ltn12.sink.table(response_body),
}
return table.concat(response_body), rBody, code, response_headers, rStatus
end

local xml_tree = lom.parse(https.get((‘https://192.168.148.3:6547/quickstatus’)))
print(xml_tree)
[/code]

The output was blank and this is the furthest I’ve have gotten without any errors :-[

Then I thought regex could solve it but this code won’t work neither probably because of the " symbols
I’m trying to get this

<div id="value_DeviceStatus" class="value">On line (Green Mode)</div>

[code]
local socket = require(“socket”)
local ltn12 = require(‘ltn12’)
local https = require(“ssl.https”)
local lom = require’lxp.lom’
local xpath = require’xpath’

function https.get(u)
local response_body = { }
local rBody, code, response_headers, rStatus = https.request{
url = u,
method = “GET”,
verify = “none”,
mode = “client”,
options = “all”,
protocol = “sslv3”,
headers = {
[“Accept”] = “text/html”,
[“Accept-Encoding”] = “sdch”,
},
sink = ltn12.sink.table(response_body),
}
return table.concat(response_body), rBody, code, response_headers, rStatus
end

local xml_tree = (https.get((‘https://192.168.148.3:6547/quickstatus’)))
local DeviceStatus = xml_tree: match “

([^<]*)
”
print(DeviceStatus)[/code]

returns a “attempt to call global ‘value_DeviceStatus’ (a nil value)”

The XML file is not that big and i’ve read that regex is faster then LXP. So what is the best approach in my case and what I’m I doing wrong?

My reading material for anyone who’s on the same path:
[url=http://luaxpath.luaforge.net/manual.html]http://luaxpath.luaforge.net/manual.html[/url] LUA XPath
[url=http://forum.micasaverde.com/index.php?topic=12408.15]http://forum.micasaverde.com/index.php?topic=12408.15[/url] user: Hek, wrote a nice script that uses LXP and Xpath
[url=http://forum.micasaverde.com/index.php/topic,15176.0.html]http://forum.micasaverde.com/index.php/topic,15176.0.html[/url] user: mda, wrote a script that uses regex

Belkin WeMo plugin and the weatherparsing script also uses LXP but that script was way to complex for me to understand

#2

At least one of your problems is this line

local DeviceStatus = xml_tree: match "<div id="value_DeviceStatus" class="value">([^<]*)</div>"

You could fix it like this

local DeviceStatus = xml_tree: match '<div id="value_DeviceStatus" class="value">([^<]*)</div>'

…although there are others ways too.

#3

Heres a simple solution for you…


function getParams(html)
  local params = {}
  html = html:gsub("\r",""):gsub("\n",""):gsub('<head>(.+)</head>',''):gsub('(.+)<div id="mainbox">','')
  for vID,vValue in html:gmatch('<div id="(.-)" class="value">(.-)</div>') do
    params[vID] = vValue
  end
  return params
end

This will extract all the parameters on the page into a lua array.

You can then reference the returned values by name…

eg:

local html = https.get('https://192.168.148.3:6547/quickstatus')
local rValues = getParams(html)
local upsStatus = rValues["value_DeviceStatus"]

Basically, the getParams function strips out line feed and carriage return characters from the input html, then extracts the contents of the “mainbox”

, then uses the string.gmatch function to extract the elements of the individual value
's…

#4

Thank you both for your suggested solution.

I tested both scripts to see which one had the fastest run time (tested in LuaTest 1.6)
Akbooer’s solution had a run time of 125.3 ms
Cybrmage solution had a run time of 454.1 ms

Since I only need three values I think I’m going for Akboer’s suggestion since it has the fastest run time. But I’m definitely keeping Cybrmage his suggestion in my bookmarks. I think his suggestion is better when you need the extract a lot of values out an XML file and probably makes up the time then.