Macro "random_XYZ" for own searchable content - is it memorizing media urls?

peterb

If we use e.g.  %random_image% to insert a suitable image for the article, SER uses a user keyword to search for the image online.
Does it memorize the list of scraped image urls for the next use of this keyword with the macro?
Many of the image or video sites have 10 or 20 results. Therefore the searching time for those medias could be reduces at least 10-20 times.
It seems that only the first page of the media site is being processed so when the project runs for a while then less and less time would be needed to search for images/video as they have been saved already.

To get more images per images source one could use the #random<range> macro to access next pages:
e.g.   http://imageshack.us/photos/#random[1..3]/gardening   <= does that work?

If SER would remember all already processed media search terms like  http://imageshack.us/photos/3/gardening  together with the image urls on that page, the searching time for media could be reduced quite a bit.
Maybe it already works like this though

Sven

It keeps the found content (first site of result page only) for 12 hours...than it tries to get some new media and if it has, it deletes all older than 12 hours (for that type and keyword).

This makes sure that fresh content is used and no old none existing once. And I like to keep it that way. The 

AlexR

@Sven - is it possible to set the images that it uses? Like you have done with geeneric urls. 

I have a nice group of images per niche that I'd love to use. 

So if it was:

www.image.com/image1,keyword1,keyword2,keyword3, etc

This means we don't have to waste resources finding images, and it means that the images I use will be different to other SER users. 

peterb

put this line into your article where you need it:       %spinfile-C:/GSADATA/image_spintax.txt%

contents of   "image_spintax.txt" (devide imageurls with "|") 
Important: The whole code below must be in 1 line!
==========================================
<div style ='{float:left;margin-right:10px;'|text-align:center'|float:right;margin-left:10px'}>  <img src='{imageurl1|imageurl2|imageurl3|...}' width='260' alt=''/>  </div>

AlexR

@PeterB - How does that choose an image URL based on a keyword?

peterb

It's interesting that you want to choose an imageurl based on a keyword and not just based on a project.
so if your spinned article is about "animals" and you want to use a random keyword within the article to focus on a certain animal (e.g. elephant, monkey, ..) you maybe could do it like this:

Change the image url part from the previous post:  {imageurl1|imageurl2|imageurl3|...}

• to:  %columnspinfile-C:\GSADATA\animals.txt-2%    <= which takes the 2. column of a random row of a file

In the Article text where you want the keyworld to appear you could then use:

• %columnspinfile-C:\GSADATA\animals.txt-1%  <= which takes the 1. column of the same row of that file
  

A sample row with the two columns of the file "animals.txt" could look like this:  
Monkey,htttp://imageshack.us/fhs5d5s.jpg     <= use spintax {url1|url2|...} for multiple images per keyword

AlexR

@PeterB - I think I haven't explained correctly. 

1) Currently I have an image file per project. It get's called in the project article using a spinfile macro and the image is a related to the article content. (i.e. if article is about animals it picks 1 of my 100 animal images) 

2) BUT each project has the spintax and I have to keep 90 or so different URL files. 1 for each project.

3) It would be much more efficient if it worked like the generic URL's system.

A single file where a number of keywords are assigned to an URL. It finds the appropriate keyword and then grabs the appropriate URL. You use 1 macro in all projects, rather than 1 macro PER project. (90 different macros!)

So it would work:

Image.com/image,keyword1,keyword2,keyword3

Then I could maintain 1 image text file across all VPS's rather than 90 different text files. (1 for each project) Much more scalable and maintainable. 

peterb

haven't done it, but do you use dropbox, so that all VPSs get their files from the same source?  (some threads about it on this forum)
if you really want this format:  Image.com/image,keyword1,keyword2,keyword3
you could just create exactly that file which includes all images with according keywords as you mentioned and then create (or let create) a little e.g. php script (10min) or Winautomation .exe Script that creates automatically the 90 or 900 suitable files with the accordings image urls. the filenames could be the keyword like "animals.txt" and in the article you include  %spinfile-..../animals.txt% .. or even a spintax of %spinfile% commands.