I am using Python 2.7, Webdriver and Chrome. Manually, I can mouse swipe across text containing a hyperlink on a web page and copy it to the clipboard. How do I do this automatically? I have no issue finding the element containing the hyperlink. I am not trying to find the hyperlink. I am trying to paste it into a web page text box which does not process https://www.python.org/ ">Link within an "a" tag but processes it correctly when pasted from elsewhere i.e. "Link" with embedded href.
Even after OP clarifications, it's still hard to understand the exact issue, so I'll try to cover all possible options :)
Suppose we have an anchor element, like Link
We can find this element in such ways
element = driver.find_element_by_xpath('//a[text()="Link"]')
element = driver.find_element_by_xpath('//a[#href=" python.org "]')
depending on what information we currently know about the element and what exactly we want to scrap.
Also, we can use index of anchor element element = driver.find_elements_by_tag_name('a')[0]
1) To get value of href attribute:
value = element.get_attribute('href')
Output: https://python.org
2) To get value of text node:
value = element.text
Output: "Link"
3) To get complete HTML code of element:
value = element.get_attribute('outerHTML')
Output: Link
We have a component that contains a background image. Our front-end guy needs it to be loaded through CSS (i.e. background: url(/*path here*/)...). The following is a possible solution we came up with:
#string src = // Get image path from Sitecore().Field("Picture");
<div style="background: url(#src) left top no-repeat;"> ... </div>
However, there are two problems with this approach:
It makes it very difficult for the content editor to swap out the image. They will have to manually change it through edit item.
It feels like a hack/workaround.
So the question is as follows: Is there a way to edit the CSS of an element through Razor/Sitecore? Specifically, the background: field.
I had a similar case and I used :
<footer class="layout_footer" style="background-color: #Model.BackgroundColor">
on view rendering (cshtml file)
And on the model we have :
public string BackgroundColor
{
get
{
Sitecore.Data.Fields.ImageField imgField =((Sitecore.Data.Fields.ImageField)item.Fields["BackgroundImage"]);
return Sitecore.Resources.Media.MediaManager.GetMediaUrl(imgField.MediaItem);
}
}
For editing this field in page editor you can use Sitecore Field Editor from a command : http://blog.istern.dk/2012/05/21/running-sitecore-field-editor-from-a-command/
Check for edit mode, and display in edit mode a editable field. Also create a Custom Experience Button from the Field Editor Button Type. You can also display. See User friendly developing with the Sitecore Experience Editor
#string src = // Get image path from Sitecore().Field("Picture");
<div style="background: url(#src) left top no-repeat;">
#if (IsInEditingMode)
{
<h3>Backgroiund Picture: #Editable(m => m.Picture)</h3>
}
</div>
There is no Sitecore extension method which will do this out of the box (i.e. #Html.Sitecore().Field("fieldName") will not work here as it would render the entire image tag (also a load of other non-image markup in page editor mode) as you probably know.
The method that #sitecore climber mentions is useful for controller renderings (or view renderings with a custom RenderingModel). If you want to stick with simple view renderings (i.e. not create a RenderingModel) then you could create a Html extension method which can be re-used on any view rendering. This could be something like the following:
public string ImageFieldSrc(this SitecoreHelper sitecoreHelper, string fieldName, Item item = null)
{
if (item == null) {
item = sitecoreHelper.CurrentItem;
}
var imageField = new ImageField(item.Fields[fieldName]);
var mediaItem = imageField.MediaItem;
var mediaUrl = MediaManager.GetMediaUrl(mediaItem);
mediaUrl = HashingUtils.ProtectAssetUrl(mediaUrl); //if you want to use media request protection (adding the hash onto the end of the URL, use this line
return mediaUrl;
}
It's worth noting that if you are using Sitecore 7.5 or above there is a feature to protect media URLs with a hash to prevent malicious DoS type attacks described in this blog post by Adam Najmanowicz.
In summary; if you are using Sitecore 7.5+ and you use media hashing then you will need to call HashingUtils.ProtectAssetUrl on the media URL if it is to respect size parameters.
Using Ctrl-F in most browsers will allow you to search for text, but only in only the text areas. I would like to search for text in what should be accessible areas that are not necessarily text rendered areas such as <map ...><area title="searchable text" /></map> and <img alt="searchable text" />. Is there a browser or addon that will do what I'm asking for? This stuff is here for accessibility, but it doesn't seem to be really all that accessible (except by mouse hover, which again isn't all that accessible).
NOTE
An answer that is required, does not use something that is decoupled from the view. I.e. searching through the source code isn't an option as this is largely difficult to read (esp on complex pages) and doesn't show where the information is located on the rendered page.
Is there a browser or addon that will do what I'm asking for?
Oh yes. Lynx browser does it.
But I guess it's not a solution ;-)
If your question is so, there is no way to override what CTRL+F is doing in your browser.
You can design a custom plugin inside your website, or an addon for your browser. This would be quite easy... but will require other shortcut.
If your main problem is to locate tags based on their alt or title attributes content, this is quite easy in javascript:
var search='enter image';
var nodes=document.querySelectorAll("[alt*='"+search+"'],[title*='"+search+"']");
You can then highlight the matching nodes using jquery or what you want.
for (i in nodes) {
nodes[i].className+=' resultHighlighted';
}
and scroll to the first result:
nodes[0].scrollIntoView();
If you intend to create a browser plugin, you can create your custom a bookmarklet or a custom plugin, and associate a shortcut to this bookmark (see https://github.com/iSunilSV/Chrome-Bookmark-Shortcut)
A simple bookmarklet to find the first match by title or alt attribute and scroll to it will be something like that:
javascript:text=prompt("search inside alt or title attribute");
document.querySelector("[alt*='"+text+"'],[title*='"+text+"']").scrollIntoView();
In your browser, use the "View Source" or "Source Code" function, and then within that window that pops up, use the Ctrl-F for Find.
You can also use the "Inspect Element" directly on an element to split the screen into two windows- one for code and one that's rendered.
For more information, here's a sample article for Chrome:
https://support.google.com/adsense/answer/181951?hl=en
Would something like the Web Developer browser plugin work? It's available for Chrome, FF and Opera. There are a few features that toggle the display of various attributes such as title, alt and even ARIA roles. This injects the attribute text inline with the element.
In my opinion, it's not a bug; they were just not designed for this use.
As I'm sure you are aware, the alt attribute replaces the image when it's not available. So how could you scroll to something that is not always displayed? Whereas you seem to be after a permanent description; a figcaption would be more appropriate for this.
As for the title attribute, it was intended to merely clarify the purpose of a link. There should not be any new information to the user in the title; therefore I think it would be redundant to have two lots of the same information highlighted in one place.
The purpose of searching is to find text on screen, seeing as neither title or alt are always displayed I think the user would be more confused by the fact that nothing is highlighted, and that they are just taken to an image or random link/area. If the image has a figcaption that becomes highlighted, then it would make sense to them. Besides, how are they going to search for the title if they don't know what to search for? Title and alt do not come up in text displayed by search engines; the user will never know about it unless they've been to your site before, in which case they'll know where to look.
Also you state the following:
This stuff is here for accessibility, but it doesn't seem to be really all that accessible
Which, understandably, seems true to you as you probably do not need it. However alt and title are read out to those who use screen readers so isn't entirely useless.
Idea 1
I assume you have Windows and Firefox installed
I have my Firefox installed with 2 add-ons.
Install a add-on called Tile Tabs, it make it possible for example left side is web view and the same page on right side with source code.
Install add-on called Web page to source code & viceversa that make it possible to toggle between view and source code by pressing on CTRL+SHIFT+S
Since what you required is not a default thing in all nowadays browses as far as I know.
Screen shot of the solution:
Idea 2
Install FireBug, you can view/edit/debug source codes and view HTML live and what you highlight on the code will be also highlighted on the view.
Screen shot:
Note: Btw idea 1 is not only good for view / source code but it is also good to compare two views or read article to the right and answer question to the left.
You can use the search funktion in Chrome's developer tools "Elements" Tab (Press F12 -> Tab "Elements" -> Press CTRL + F) and use XPath on your searches. Example:
//*[#title="Google"]
Matches will be shown with a yellow background in the code and when you hover it, its position will be hightlited in the view.
Dev Tools "Element" Search with XPath
It is coupled with the view, allows you to see the element's position and it's also an out-of-the-box solution in Chrome (tested in Chromium 45 for Ubuntu).
Hope it helps!
Regards
EDIT
Forgot - If you want to use wildcards on your searches, you can also do it like this:
//*[contains(#title, 'Google')]
EDIT 2
For the posterity! Further research shows that your goal might be possible to achieve using the Firefox-Addon Greasemonkey, which allows you to customize the way a web page displays or behaves, by using small bits of JavaScript.
I performed several tests with this addon and could achieve a nice effect with simple images (display the ALT attribute as a DIV overlapping the image), but with area sections the thing gets a lot more complicated, as area regions can be squares, circles, and polygons with infinite coordinates plus retrieving the exact positioning of the area itself can be a bit tricky but maybe gives you or someone else a start point.
Based on the ALT Tooltips Script (http://greasemonkey-user-scripts.arantius.com/alt-tooltips-for-firefox), I created the following script and defined it in Greasemonkey:
// ==UserScript==
// #name Alt Tooltips 2
// #namespace http://www.biterion.com
// #description Alt Tooltips 2
// #include *
// #grant all
// ==/UserScript==
function getPosition(element) {
var xPosition = 0;
var yPosition = 0;
while(element) {
xPosition += (element.offsetLeft - element.scrollLeft + element.clientLeft);
yPosition += (element.offsetTop - element.scrollTop + element.clientTop);
element = element.offsetParent;
}
return { x: xPosition, y: yPosition };
}
function getAreaPosition(element) {
var position = element.coords.split(',');
xPosition = position[0];
yPosition = position[1];
return { x: xPosition, y: yPosition}
}
var res = document.evaluate("//img",document, null, XPathResult.UNORDERED_NODE_SNAPSHOT_TYPE, null);
var i, el;
for (i=0; el=res.snapshotItem(i); i++) {
if(el.alt) {
alternate = el.alt
} else {
alternate = "No alt text";
}
position = getPosition(el);
var newDIV = document.createElement ('div');
newDIV.innerHTML = "<div style='position:absolute;background:yellow;color:black;top:" + position["y"] + ";left:" + position["x"] + "' id=" + i + ">" + alternate + "</div>";
document.body.appendChild(newDIV);
}
var res2 = document.evaluate("//area",document, null, XPathResult.UNORDERED_NODE_SNAPSHOT_TYPE, null);
var i2, el2;
for (i2=0; el2=res2.snapshotItem(i2); i2++) {
if(el2.alt) {
alternate2 = el2.alt
} else {
alternate2 = "No alt text";
}
position2 = getAreaPosition(el2);
var newDIV2 = document.createElement('div');
newDIV2.innerHTML = "<div style='position:absolute;background:yellow;color:black;top:" + position2["y"] + ";left:" + position2["x"] + "' id=" + i2 + ">" + alternate2 + "</div>";
document.body.appendChild(newDIV2);
}
As you can see, the script firstly detects all "img" and "area" elements, extracts its positioning and creates a new DIV element containing the "alt" attribute, which is then positioned on the upper left corner of the image.
As stated, the problem with areas is, that the positioning should be relative to the parent image and not absolute like in the script, plus the coordinates should be extracted accordingly to the type of area shape (currently only extracting the two first coordinates of each area, which will work for squares but will surely fail for other shapes).
Hope this will help someone :-D
Regards
I have a base .docx for which I need to change the page header / footer image on a case by case basis. I read that python-docx does not yet handle headers/footers but it does handle Pictures.
What I cannot work around is how to replace them.
I found the Pictures in the documents ._package.parts objects as ImagePart, I could even try to identify the image by its partname attribute.
What I could not find in any way is how to replace the image. I tried replacing the ImagePart ._blob and ._image attributes but it makes no difference after saving.
So, what would be the "good" way to replace one Image blob with another one using python-docx? (it is the only change I need to do).
Current code is:
d = Document(docx='basefile.docx')
parts = d._package
for p in parts:
if isinstance(p, docx.parts.image.ImagePart) and p.partname.find('image1.png'):
img = p
break
img._blob = open('newfile.png', 'r').read()
d.save('newfile.docx')
Thanks,
marc
There is no requirement to use python-docx. I found another Python library for messing with docx files called "paradocx" altought it seems a bit abandoned it works for what I need.
python-docx would be preferable as the project seems more healthy so a solution based on it is still desired.
Anyway, here is the paradocx based solution:
from paradocx import Document
from paradocx.headerfooter import HeaderPart
template = 'template.docx'
newimg = open('new_file.png', 'r')
doc = Document.from_file(template)
header = doc.get_parts_by_class(HeaderPart).next()
img = header.related('http://schemas.openxmlformats.org/officeDocument/2006/relationships/image')[0]
img.data = newimg.read()
newimg.close()
doc.save('prueba.docx')
I am writing a simple plugin that changes the whole content of a Sublime editor and replaces it with something else. When doing this
viewRegion = sublime.Region(0, self.view.size())
everything = self.view.substr(viewRegion)
self.view.replace(edit, viewRegion, everything)
the view scrolls to the top. How do I prevent this or restore the original viewport?
It turns out, there isn't really a good way. You can try to get around the replace as much as possible (e.g. by using the excellent merge_utils), but this will always change the position after the command is finished.
As a hack, you can save the position and restore it in a timeout like this:
def format(self, edit):
self.curpos = self.view.viewport_position()
self.selection = self.view.sel()[0]
viewRegion = sublime.Region(0, self.view.size())
everything = self.view.substr(viewRegion)
# do something useful
self.view.replace(edit, viewRegion, everything)
# reestablish selection
sublime.set_timeout(self.reset_pos, 0)
def reset_pos(self):
self.view.sel().clear()
self.view.sel().add(self.selection)
self.view.set_viewport_position(self.curpos, False)