To get text from a web page

[koma]

I'm writing a script that takes different actions depending on the contents of the current web page in my browser.

Since WinGetText doesn't work well with Firefox I'm using a workaround:

I send ctrl-A and ctrl-C to the browser and then uses whats in the clipboard.

However, the problem is that sometimes it is almost like the ctrl-C doesn't work. Nothing ends up in the clipboard even though it should have...

Anyone knows what's going on?

[phillip123adams]

I'm writing a script that takes different actions depending on the contents of the current web page in my browser.

Since WinGetText doesn't work well with Firefox I'm using a workaround:

I send ctrl-A and ctrl-C to the browser and then uses whats in the clipboard.

However, the problem is that sometimes it is almost like the ctrl-C doesn't work. Nothing ends up in the clipboard even though it should have...

Anyone knows what's going on?

<{POST_SNAPBACK}>

Just a guess, but I have had similar problems with ctri-C because of the case of the "C". Here's a quote from the AutoIt help file for the Send function:

N.B. Some programs are very choosy about capital letters and CTRL keys, i.e. "^A" is different to "^a". The first says CTRL+SHIFT+A, the second is CTRL+a. If in doubt, use lowercase!

[koma]

Just a guess, but I have had similar problems with ctri-C because of the case of the "C".  Here's a quote from the AutoIt help file for the Send function:

<{POST_SNAPBACK}>

I don't think that is the problem, because it works _sometimes_...

[Jos]

Just a couple of thoughts:

- are you sure the page is loaded before doing ctrl+a & crtl+c?

- are you sure the Browser/Section has the focus when doing ctrl+a & crtl+c?

[layer]

JdeB is probably right, but about the Ctrl+c.... This has happened to me a few times, like im unable to use my windows key, or copy stuff... This is because i used to have a program where it's hotkey was "Alt", it was the program Ventrilo where you could talk to people with your voice online... And anyways, pressing "Alt" usually disabled my windows key and copy functions for some reason... but pressing "Alt" again, fixed it! So I'd give JdeB's idea a try first, if that doesn't work, try pressing "Alt"!

EDIT: fixed bold

Edited March 28, 2005 by layer

[koma]

Just a couple of thoughts:

- are you sure the page is loaded before doing ctrl+a & crtl+c?

- are you sure the Browser/Section has the focus when doing ctrl+a & crtl+c?

<{POST_SNAPBACK}>

I am sure the page is loaded, but I am not sure that the Browser/Section has the focus...

Is there a best practice for setting focus for a specific part of a window?

[steveR]

Try the ControlFocus() command

Also remember alot of times you can use CTRL-INSERT as Copy and SHIFT-INSERT as Paste as alternatives. Not all software still support those hotkeys tho.

EDIT: If Firefox doesn't like the ControlFocus() command, pick somewhere on the page where you know there won't be any links or anything and just do MouseMove() and MouseClick() to set the focus.

Edited March 28, 2005 by steveR

[Blue_Drache]

What about InetGet() and pulling the page out of the cache?

[koma]

What about InetGet() and pulling the page out of the cache?

<{POST_SNAPBACK}>

InetGet may be good for static pages, but this is really dynamic stuff...

(Or did I understand InetGet wrong?)

[koma]

Try the ControlFocus() command

Also remember alot of times you can use CTRL-INSERT as Copy and SHIFT-INSERT as Paste as alternatives. Not all software still support those hotkeys tho.

EDIT: If Firefox doesn't like the ControlFocus() command, pick somewhere on the page where you know there won't be any links or anything and just do MouseMove() and MouseClick() to set the focus.

<{POST_SNAPBACK}>

I did try the mouseClick without luck. However, I didn't try the mouseMove... Maybe that will do the trick...

[koma]

Here is a bit of code that works - sometimes:

ClipPut("!")

WinActivate("CNN")
Send("^a")
Send("^c")

MsgBox(0, "Text found was:", ClipGet())

To run it, open either Firefox or IE at www.cnn.com and run the script.

The funny thing is that most of the time, the script returns the text from cnn.com.

But when it fails it returns 1. Not "!" as I would expect!! Anyone knows why??

[steveR]

I did try the mouseClick without luck. However, I didn't try the mouseMove... Maybe that will do the trick...

<{POST_SNAPBACK}>

Well I meant that you use the MouseMove() in conjuction with MouseClick().

Use MouseMove() to move the mouse to an area of the page (off to the side maybe), and then use MouseClick() to let the control (in this case, a fancy edit box) gain focus.

Someplace where there are no links or anything. It only takes a pixel.

Edited March 28, 2005 by steveR

[layer]

try activating it, or "focusing" it..

WinActivate ( "title" [, "text"] )

although i dont use firefox..

EDIT: the "Text" paremeter in WinActivate, is optional...

Edited March 28, 2005 by layer

[koma]

try activating it, or "focusing" it..

WinActivate ( "title" [, "text"] )

although i dont use firefox..

EDIT: the "Text" paremeter in WinActivate, is optional...

<{POST_SNAPBACK}>

I have a WinActivate in my script... But with no text parameter. But that doesn't matter, does it?

[steveR]

Missing text parameter just means it will match any text.

[phillip123adams]

Here is a bit of code that works - sometimes:

ClipPut("!")

WinActivate("CNN")
Send("^a")
Send("^c")

MsgBox(0, "Text found was:", ClipGet())

To run it, open either Firefox or IE at www.cnn.com and run the script.

The funny thing is that most of the time, the script returns the text from cnn.com.

But when it fails it returns 1. Not "!" as I would expect!! Anyone knows why??

<{POST_SNAPBACK}>

After allowing the CNN page to fully load in Firefox, I tried your code, and without a Sleep period after ctrl-c, I get "!". With a Sleep period greater than 150, I got the page each of the 10 times I tried it on 3 different pages.

Perhaps a larger value would be even safer.

[Alterego]

Use Lynx. You can either use Cygwin or get one of the several flavors compiled for windows. After Lynx is set up as an environmental path variable your run command will look something like this.

RunWait(@ComSpec & ' /c ' & 'lynx -dump -accept_all_cookies C:\website.html http://www.website.com')

Note that -dump instructs Lynx to parse the html of the page. So you're text file will end up with the text that you would normally see in your browser and not the source code. I use this method to scrape alexa for their rankings since they are not always directly in the browser code and are obfuscated. (and I use cygwin, it's worth the day it takes to set up and get used to it. Windows compiled versions of Lynx never seem to work quite right for me)

[DaleHohm]

Or just get surgical about it:

Opt("WinTitleMatchMode", 4); allow ClassName lookup to avoid window confusion
$appWindow = WinGetHandle("classname=MozillaWindowClass")

;Give Focus to the main window in FireFox
;ControlId for main FireFox window is 1 (use the Active Window Info tool)
ControlFocus($appWindow,"",1)
Send("^{a}^{c}")

Dale

[Poconnor2018]

I know this issue is likely solved at this point, but when sends are randomly not working, a simple fix is often to put a brief Sleep() between them.

[BrewManNH]

1 hour ago, Poconnor2018 said:

I know this issue is likely solved at this point

After 13 years,  I would hope so.