Applescript: GUI Programming for recursively printing to PDF for Appleworks - pdf

Question

The Goal: thousands of old Clarisworks and Appleworks documents need to become PDFs.
The Problem: Scripting Pages is not an option; it doesn't preserve formatting reliably. It must be Appleworks. This script works, when it works; but it skips files for reasons I cannot identify.
set appleworksFolder to choose folder
tell application "Finder"
set folderItems to (files of entire contents of appleworksFolder)
repeat with I from 1 to number of items in folderItems
set the_doc to item I of folderItems
set doc_name to name of the_doc as text
(* Some files are missing metatags and try to save as "Untitled Document",
this block ensures a file name is unique, later *)
tell application "Finder"
set the clipboard to doc_name & ".pdf"
end tell
(* Each file exists in a folder with a path.txt file that will later
be used to put the file back where it was originally stored prior
to this conversion process *)
if name of the_doc is not "path.txt" then
try
tell application "Finder"
(* Many files no longer have name extensions and appear as UNIX
executables if not repaired *)
try
set nmex to name extension of the_doc as text
on error
set nmex to "ok"
end try
if nmex is not "cwk" or "CWK" then
set the_doc_str to the_doc as text
set doc_path to POSIX path of the_doc_str
do shell script "mv '" & doc_path & "' " & "'" & doc_path & ".cwk'"
end if
delay 1
(* In case Appleworks hangs or otherwise bungs up, I force-quit
it at the end of the script; this ensures its closed before
it tries to open the next file *)
if (application process "Appleworks 6" of application "System Events" exists) then
do shell script "killall 'LaunchCFMApp'"
delay 1
end if
tell application "AppleWorks 6"
open the_doc
end tell
(* Some of the documents are huge, this delay gives the app time to load
since this is all GUI-scripted *)
delay 5
tell application process "Appleworks 6" of application "System Events"
(* This is where I think I am encountering problems; there are two
possible warnings that may or may not appear on opening the doc;
that Appleworks needs to append a version number to the file (if
its old) or that the file may be damaged and thus would need to be
skipped. I get system beeps sometimes in this part, but I don't know
why! *)
if (button "OK" of window 1 of application process "AppleWorks 6" of application "System Events" exists) then
delay 0.5
keystroke return
delay 0.5
end if
delay 2
if (button "OK" of window 1 of application process "AppleWorks 6" of application "System Events" exists) then
delay 0.5
keystroke return
delay 0.5
end if
delay 2
(* If the document loads, the Appleworks welcome pane won't be there;
this part of the script works flawlessly, when it happens. Sometimes
documents are outside of print margins, hence the press-ok-button
conditional *)
if not (window "Starting Points" of application process "AppleWorks 6" of application "System Events" exists) then
tell application process "Appleworks 6" of application "System Events"
keystroke "p" using command down
delay 1
click menu button "PDF" of window "Print" of application process "AppleWorks 6" of application "System Events"
delay 1
click menu item "Save as PDF…" of menu "PDF" of menu button "PDF" of window "Print" of application process "AppleWorks 6" of application "System Events"
delay 1
keystroke "v" using command down
click button "Save" of window "Save" of application process "AppleWorks 6" of application "System Events"
delay 8
keystroke "w" using command down
delay 0.5
if (button 1 of window 1 of application process "AppleWorks 6" of application "System Events" exists) then
delay 0.5
keystroke "d" using command down
delay 0.5
end if
delay 0.5
end tell
end if
do shell script "killall 'LaunchCFMApp'"
delay 3
end tell
end tell
end try
end if
end repeat
end tell
I want to this baby to just run through a couple of thousands files over a weekend and create PDFs, but each time I run it overnight I find a few hundred correctly processed documents, hundreds or thousands of skipped documents, and often a print dialogue for the script itself, which obviously comes from using Command+P outside of an Appleworks context. I'm an Applescript noob, to be sure, this has been driving me nuts for weeks!

Answer 1

This is not really a solution to your problem, as it is highly specific and tailored, but I have some possibly useful tips:
You don't need to close the "Starting Points" window. It will be dismissed when you tell AppleWorks to open a document.
Try to use as few simulated keystrokes as possible. For instance, instead of simulating cmd+P to open the Print dialog, it is better to simulate a click on the application's Print… command inside its File menu:
click menu item "Print…" of menu "File" of menu bar 1 of application process "AppleWorks 6"
and instead of performing a cmd+W to close the document, the proper way is to do
tell application "AppleWorks 6" to close front document saving no
If you do need to simulate keystrokes, ensure that they are received by the correct application, by activating it just before the keystroke command with a tell app "AppleWorks 6" to activate, just in case it wasn't the frontmost app at that point.
When checking for additional dialog windows, it's a bad idea to use if (button 1 of window 1 of (*...*) exists), because on my machine, for instance, checking this while the "Print" dialog was open caused the script to hang for minutes (possibly for good), and because it's always advisable to address a window by name rather than number (since window 1 is simply the frontmost window, and that can quickly become another). So, it's better to check if a window named such and such exists.
In order to process thousands of files where you expect errors to happen, you need to restructure your code and rethink the error handling. For instance, you could log error information to a text file in order to see which files have been skipped and why. To separate program logic from the error handling and details, you could use handlers, so your main loop could look like this:
set logfile to alias "Macintosh HD:Users:user:Desktop:errors.log"
open for access logfile with write permission
repeat with the_doc in folderItems
try
open_with_appleworks(the_doc)
print_to_pdf()
close_document_and_quit_appleworks()
on error error_message
write error_message & "\n" to logfile
close_document_and_quit_appleworks()
end try
end repeat
close access logfile
In a similar way, you can log success messages, too. In the single handlers, then, you are supposed to specify details about how exactly to do every step. They can also provide some error information to be logged:
to open_with_appleworks(the_doc)
tell app "AppleWorks 6"
try
(* do your stuff here *)
on error err_msg
(* re-signal the error to main loop! *)
error "Failed to open " & (the_doc as text) & err_msg
end try
end tell
end open_with_appleworks
Hard-coded delays slow down your script considerably: while only some of your documents might take a few seconds to open, your script will always wait 5 seconds – even if the document opens instantaneously. It's better to have a loop check whether the document has opened yet:
repeat while not (exists front document)
-- just wait
-- or, delay 0.1 -- if you really want to
end repeat
The same holds for nearly all your other delays, most of them can be avoided, as they will be much shorter or longer in practice. Instead of delaying half a second, hoping that the warning or dialog will have appeared by then, it's better to check whether it's there or not.
(one last tip and then I'll be quiet) Unfortunately, you seem to already have gone through the trouble of copying all the files to be processed into one folder, remembering in some "path.txt" where they came from... This can be avoided if the files are in sub-folders of some folder and can be found by a filter: You just create a "smsart folder" with that specific search (example), then say in your AppleScript set theFiles to choose file with multiple selections allowed, then select all files in that particular smart folder, and you're good. The script can then easily find out where a file f is by asking application "Finder" to get parent of f (that will return a Finder file reference, which can be easily converted to an alias and to a POSIX file address). Then, since you know the name and location of the PDF file generated by "printing" the document, you can move that PDF to the folder where its source document is.