How to find and click 100% or 99% match images in sikuli?

all is in the docs:
http://sikulix-2014.readthedocs.org/en/latest/index.html

version 1.0.1+

Jython:
find(Pattern(image).exact())
click()
... and in the Sikuli IDE use the Preview feature

Java:
scr = new Screen();
scr.find(find(new Pattern(image).exact())
scr.click()

Hi, I'm not able to find the 99% match with that function. Any alternatives for this?

please show me your relevant coding.

what Sikuli version?

@M

Are you interested in finding objects with 99% match specifically?

If so you can try this to check found object's similarity score like that:

res = find(pattern)
print res.getScore()

In case of an exact match, you will get:
1.0

In case of non 100% match, you will get something like this:
0.997599840164

Then you can work with these numbers.

Eugene

Hi,
I have another doubt. When I'm doing form validation like entering a password in the password box by using the screenshot of the password text box, the next time in the loop, the exact match of the password text box is not found because earlier it was an empty text box and now it has some text. So the match fails. How to overcome this problem? Is there a way to compare or click images at run time?

You have to use the part of the password box, that does not contain the entered password (usually some label text left or above, together with part of the field border) and use a targetOffset for the click point.
Or use some other image, that is visible together with the password box and again use targetOffset to define the click point inside the box relative to this image.

Hi RaiMan, it would be helpful if you could give the code snippet for the same.
From the explanation above my understanding is that I have to be using the text box with some text as the initial image for comparison and then use target offset to get the exact click point. Is that right?

easiest and most effective approach:
send me a screenshot of your password box situation (whole application window, initial or filled does not matter) and I will return a fitting snippet

... silently to my mail at https://launchpad.net/~raimund-hocke

you have to send from you normal mail program