Add-on:Parsedom for xbmc plugins: Difference between revisions
>NedBot m (Addon-Bot repo category update) |
No edit summary |
||
| Line 9: | Line 9: | ||
|Description= | |Description= | ||
|Platform=all | |Platform=all | ||
|Repo= | |||
|icon url=}} | |icon url=}} | ||
= Testing/Status = | = Testing/Status = | ||
Revision as of 08:23, 17 December 2011
| Parsedom for xbmc plugins. | ||||||||||
| ||||||||||
|
 Add-ons |
Parsedom for xbmc plugins.
|
Testing/Status
Integration and unittests are run continously by Jenkins
http://tc.tobiasussing.dk/jenkins/view/Common%20Functions/
Developers
This DOM parser is a fast replacement for Beautiful Soup.
And also a few other useful functions.
Setup
To use the parsedom functions edit your addon.xml like this.
<requires> <import addon="xbmc.python" version="2.0"/> <import addon="script.module.parsedom" version="0.9.0"/> # Add this </requires>
And add the following to your py file.
import xbmc, xbmcvfs, xbmcaddon import CommonFunctions common = CommonFunctions.CommonFunctions() common.plugin = plugin
Debugging
To enable debugging set the following values in default.py
common.dbg = True # Default common.dbglevel = 3 # Default
Whenever you debug your own code you should also debug in the cache. Otherwise you should remember to DISABLE it.
parseDOM(self, html, name = "", attrs = {}, ret = False)
- html(string or list) - String to parse, or list of strings to parse.
- name(string) - Element to match ( for instance "span" )
- attrs(dictionary) - Dictionary with attributes you want matched in the elment ( for instance { "id": "span3", "class": "oneclass.*anotherclass", "attribute": "a random tag" } )
- ret(string or False) - Attribute in element to return value of. If not set(or False), returns content of DOM element.
returns list
Getting element content.
link_html = "<a href='bla.html'>Link Test</a>" ret = common.parseDOM(link_html, "a") print repr(ret) # Prints ['Link Test']
Getting an element attribute.
link_html = "<a href='bla.html'>Link Test</a>" ret = common.parseDOM(link_html, "a", ret = "href") print repr(ret) # Prints ['bla.html']
Get element with matching attribute.
link_html = "<a href='bla1.html' id='link1'>Link Test1</a><a href='bla2.html' id='link2'>Link Test2</a><a href='bla3.html' id='link3'>Link Test3</a>"
ret1 = common.parseDOM(link_html, "a", attrs = { "id": "link1" }, ret = "href")
ret2 = common.parseDOM(link_html, "a", attrs = { "id": "link2" })
ret3 = common.parseDOM(link_html, "a", attrs = { "id": "link3" }, ret = "id")
print repr(ret1) # Prints ['bla1.html']
print repr(ret2) # Prints ['Link Test2']
print repr(ret3) # Prints ['link3']
When scraping sites it is prudent to scrape in steps, since real websites are often complicated.
Take this example where you want to get all the user uploads.
<div id="content"> <div id="sidebar"> <div id="latest"> <a href="/video?8wxOVn99FTE">Miley Cyrus - When I Look At You</a>>br /< <a href="/video?46">Puppet theater</a><br /> <a href="/video?98">VBLOG #42</a><br /> <a href="/video?11">Fourth upload</a><br /> </div> </div> <div id="user"> <div id="uploads"> <a href="/video?12">First upload</a><br /> <a href="/video?23">Second upload</a><br /> <a href="/video?34">Third upload</a><br /> <a href="/video?41">Fourth upload</a><br /> </div> </div> </div>
The first step is to limit your search to the correct area.
One should always find the inner most DOM element that contains the needed data.
ret = common.parseDOM(html, "div", attrs = { "id": "uploads" })
The variable ret now contains
['<a href="/video?12">First upload</a><br /> <a href="/video?23">Second upload</a><br /> <a href="/video?34">Third upload</a><br /> <a href="/video?41">Fourth upload</a><br />']
And now we get the video url.
videos = common.parseDOM(ret, "a", ret = "href") print repr(videos) # Prints [ "video?12", "video?23", "video?34", "video?41" ]
XML with parseDOM
The parseDOM function supports the XML syntax
We have used parseDOM for some simple XML structures.
There is no guarantee that parseDOM for XML will be better than minidom.
xml = '<?xml version="1.0" encoding="utf-8"?><rsp generated_in="0.1760" stat="ok" />' stat = self.common.parseDOM(xml, "rsp", ret = "stat") print stat # Prints ok
Here is a slightly more complex example.
xml = '<?xml version="1.0" encoding="utf-8"?>
<rsp generated_in="0.0132" stat="ok">
<contacts on_this_page="1" page="1" perpage="25" total="1">
<contact display_name="MileyCyrus" id="458471" is_plus="0" is_pro="0" is_staff="0" mutual="1" profileurl="http://somesite/user458471" realname="MileyCyrus" username="user458471" videosurl="http://somesite/user458471/videos">
<portraits>
<portrait height="30" width="30">http://somesite/portraits/defaults/d.30.jpg</portrait>
<portrait height="75" width="75">http://somesite/portraits/defaults/d.75.jpg</portrait>
<portrait height="100" width="100">http://somesite/portraits/defaults/d.100.jpg</portrait>
<portrait height="300" width="300">http://somesite/portraits/defaults/d.300.jpg</portrait>
</portraits>
</contact>
</contacts>
</rsp>'
ids = self.common.parseDOM(result, "contact", ret = "id")
titles = self.common.parseDOM(result, "contact", ret = "display_name")
portraits = self.common.parseDOM(result, "portraits")
next = "false"
result = [];
for i in range(0, len(ids)):
group = {}
group['contact'] = ids[i]
group['Title'] = titles[i]
thumbs_width = self.common.parseDOM(portraits, "portrait", ret = "width")
thumbs_url = self.common.parseDOM(portraits, "portrait")
for j in range(0, len(thumbs_width)):
if (int(thumbs_width[j]) <= 300):
group['thumbnail'] = thumbs_url[j]
result.append(group)
print repr(result) # Prints [{'contact': '458471', 'thumbnail': 'http://somesite/portraits/defaults/d.300.jpg', 'Title': 'MileyCyrus'}
Again, this function has not been tested with XML in any way.
_fetchPage(self, dict)
WARNING: This call will be changed from "_fetchPage" to "fetchPage" in the future.
Fetches a page from the internet.
returns dict.
Input dictinary variables:
- link(string): The URL.
Input dictionary variables to come:
- cookie(string): The cookies that need to be set.
- refering(string): The refering URL.
- post_data(dict): Data to POST to the link.
The dict returned contains
- content: HTML content
- new_url: Redirect url
- header: Header information
- status: http return status
result = common._fetchPage({"link": "http://www.example.com/index.html"})
if result["status"] == 200:
print "content: %s" %result["content"]
result = common._fetchPage({"link": "http://www.example.com/doesnotexist.html"})
if result["status"] == 500:
print "redirect url: %s" %result["new_url"]
print "header: %s" %result["header"]
print "content: %s" %result["content"]
log(self, str, level = 0)
Sends the string to the xbmc.log function if the level provided is less than the level set in dbglevel.
Returns None
import CommonFunctions common = CommonFunctions.CommonFunctions() common.plugin = "PluginName" common.dbg = True common.dbglevel = 3
def helloWorld():
common.log("Ran this")
common.log("Ignored this", 4)
common.log("Ran this as well", 2)
helloWorld()
Will give the following output in the xbmc.log
[PluginName] helloWorld : 'Ran This' [PluginName] helloWorld : 'Ran This as well'
openFile(self, filepath, options = "w")
Opens a binary or text file handle
Returns filehandle
file = common.openFile("myfile.txt", "wb")
file.write("my data")
file.close()
This function raises a keyboard for user input
Returns string
search = common.getUserInput("Artist", "") # Will ask the user to write an Artist to search for
def_search = common.getUserInput("Artist", "Miley Cyrus") # Will default to Miley Cyrus if the user doesn't enter another artist.
Warning: This will return string in the future.
This function raises a keyboard numpad for user input
Returns int
pin = common.getUserInput("Userpin", "") # Will ask the user to write a pin.
def_pin = common.getUserInput("Userpin", "1234") # Will default to 1234 if the user doesn't enter another pin.
getParameters(self, dict)
Converts the request url passed on by xbmc to the plugin into a dict of key-value pairs
returns dict
params = common.getParameters(sys.argv[2]) # sys.argv[2] would be something like "?path=/root/favorites&login=true"
print repr(params) # Prints '{ "path": "/root/favorites", "login": "true" }'
replaceHtmlCodes(self, str)
WARNING: This call will be changed from "replaceHtmlCodes" to "replaceHTMLCodes" in the future.
Replaces html codes with ascii.
Returns string
clean_string = common.replaceHtmlCodes("&"…><'")
print clean_string # Prints &"...><'
stripTags(self, str)
Removes all DOM elements.
Returns string
clean_string = common.stripTags("I want this text <img src= alt='without this'>")
print clean_string # Prints "I want this text"
makeAscii(self, str)
This function implements a horrible hack related to python 2.4's terrible unicode handling.
Returns string
clean_string = common.makeAscii("test נלה מהי test")
print clean_string # Prints "test test"