Xpath or RSS element of a Tumblr feed's image url

[Openkore]

Hi,

I'm trying to post from a Tumblr feed and so far I've got the other elements in place (tags, post link, etc.) but I'm still missing the image url. What should I put there?



Pic related, that's what I'm trying to find. Someone told me to put this:

Code: Select all

//div[@class="body hasMarkup"]/p/img

but it didn't work.

Thanks!

[martin@rootjazz]

Send me the Feed and I'll tell you.

[Openkore]

Here it is:

Code: Select all

http://mahblog202.tumblr.com/rss

Is this standard for every tumblr feed? or it needs to be modified with every tumblr account I want to post from?

[martin@rootjazz]

ah sorry, you did say tumblr feed, they are standard. Was early this morning when I replied, brain wasn't working....

[martin@rootjazz]

is a difficult one as the image doesn't exist as an RSS tag,it is coming from the encoded description.

So this means it will have to be custom for each blog as we will have to use notation to

1) select the post URL
2) fire a scrape routine to scrape the post URL
3) extract the image using Xpath from the contents of the scraped post URL

If what you are trying to do is post all images as posted to another blog, would be easier to use the IMAGE SCRAPER module, to download all images from a blog to a folder, then use that folder as a watch folder

[Openkore]

May we try with this one: http://mahblog202.tumblr.com/rss to see if it works? I'm trying to pull the tags and the description (<title>) too.

[martin@rootjazz]

The rss feed module, ideally wants the image to be a tag

Code: Select all

<item>
<title>blah</title>
<img>THIS IS WHAT WE WANT</img>

but tumblr doesn't do that, not for the feed, it is encoded HTML used to specify the image in the description, TJ RSS extractor cannot pull the image from encoded HTML description.

So you would need to pull it from the tumblr post page. As this is THEME DEPENDENT, you would need specify code for each page.

[martin@rootjazz]

so if you look at the pinterest image script

Code: Select all

<link>[scrapexpath=//img[@class='pinImage']/@src]

Code: Select all

<link>

this means, use the <link> RSS item

Code: Select all

[scrapexpath=//img[@class='pinImage']/@src]

this means scrape the URL returned from the <link> tag

Code: Select all

img[@class='pinImage']/@src

this bit is the xpath on the page to use to get the image

[martin@rootjazz]

looking at one of your posts

we have this code containing the full image

Code: Select all

<div class="post-view">

    <img alt="" src="http://41.media.tumblr.com/056a708c4fb07688b37c7ad6f1641aa2/tumblr_n8n5ddT8471tpwflxo1_400.jpg"></img>

</div>

so your xpath would be

Code: Select all

//div[@class='post-view']/img/@src

making for this blog / theme the RSS SCRIPT CODE to use

Code: Select all

<link>[scrapexpath=//div[@class='post-view']/img/@src]

[Openkore]

I see it'd be difficult to configure that for every blog, and I guess the function wouldn't work anymore if the theme gets changed.

So, can you please add the "Remove links in description" and "Remove original description" options to the Mass Reblog function? I think I can use that.

Thanks again...