Page 4 of 9 FirstFirst 123456789 LastLast
Results 61 to 80 of 161
  1.    #61  
    That sounds likely. Now that I think of it, my newline handling is lazy, just chopping off the last character, whereas on Linux newlines are two characters. I'll get this fixed either tonight or tomorrow. In the meantime, you can do a quick fix by changing most of the occurrences of -1 to -2 in the file when it's reading in word lists.
  2. diomark's Avatar
    Posts
    752 Posts
    Global Posts
    770 Global Posts
    #62  
    Quote Originally Posted by dimfeld View Post
    That sounds likely. Now that I think of it, my newline handling is lazy, just chopping off the last character, whereas on Linux newlines are two characters. I'll get this fixed either tonight or tomorrow. In the meantime, you can do a quick fix by changing most of the occurrences of -1 to -2 in the file when it's reading in word lists.
    This fixed it for me -
    sed -e 's/^M//g' new_autoreplace > fixed_autoreplace

    (note - for ^M, it's control-V, Control-M)
    -mark
  3. diomark's Avatar
    Posts
    752 Posts
    Global Posts
    770 Global Posts
    #63  
    working now, but having issues with duplicates...

    i.e. both of these are in there -

    dont|don't
    dont|font


    and when I type dont, it takes the second one (translating to font)

    another example was i - it got translated to u.

    (EDIT - I think this was also because of the line handling.. the original autoreplace file wasn't being read correctly (?) editing the python script and changing -1 to -2 fixed the issue for me)

    -mark
  4. stubbs's Avatar
    Posts
    425 Posts
    Global Posts
    442 Global Posts
    #64  
    great work guys, this is the kind of stuff we need. I'm impressed!
  5. stubbs's Avatar
    Posts
    425 Posts
    Global Posts
    442 Global Posts
    #65  
    Quote Originally Posted by diomark View Post
    working now, but having issues with duplicates...

    i.e. both of these are in there -

    dont|don't
    dont|font


    and when I type dont, it takes the second one (translating to font)

    another example was i - it got translated to u.

    -mark
    Just an idea, but a different usage model to fix that problem would be to have auto-correct symbols typed with a period:

    don.t|don't

    (replace the apostrophe with a period, which has a dedicated key)
    That's how the T-9 works on my Nokia, and I think it's a handy feature. Of course, if you can think of another way to do it, that's welcome too.
  6.    #66  
    Hmm, I'm not seeing duplicates in my files. The script takes some steps to avoid having duplicates, so I'll have to look into this more. I wonder if this is perhaps related to the newline issues screwing up duplicate word detection when the words come from separate files with differing newline styles.

    The period->apostrophe option sounds nice, but unfortunately I think it won't work on the Pre since typing a period triggers the autoreplace.
  7. diomark's Avatar
    Posts
    752 Posts
    Global Posts
    770 Global Posts
    #67  
    My typing is much more accurate with this hack, but using the googwords, browser card takes an extra 15 seconds to open, and emails an extra 25 seconds

    Has anyone found a smaller list with better results?
  8. wprater's Avatar
    Posts
    240 Posts
    Global Posts
    251 Global Posts
    #68  
    Quote Originally Posted by diomark View Post
    My typing is much more accurate with this hack, but using the googwords, browser card takes an extra 15 seconds to open, and emails an extra 25 seconds

    Has anyone found a smaller list with better results?
    Yeah.. my emails are _very_ slow to load up.
  9.    #69  
    That long? How large is the autoreplace file you're using? When I did the first, much larger iteration of the file, I was seeing delays more similar to what you're reporting. But with a file of about 1.2MB, I see a 4-5 second delay in loading emails and I haven't noticed a delay in loading the browser.
  10. diomark's Avatar
    Posts
    752 Posts
    Global Posts
    770 Global Posts
    #70  
    Quote Originally Posted by dimfeld View Post
    That long? How large is the autoreplace file you're using? When I did the first, much larger iteration of the file, I was seeing delays more similar to what you're reporting. But with a file of about 1.2MB, I see a 4-5 second delay in loading emails and I haven't noticed a delay in loading the browser.
    My file (built using the googwords.txt) is 2707177

    The command I used to generate this is

    python generate_autoreplace.py ../text-edit-autoreplace googwords.txt extra_words.txt new_autoreplace

    Are you doing something different?

    mhr@jukebox:~/webos$ wc -l googwords.txt
    14762 googwords.txt
    mhr@jukebox:~/webos$ wc -l extra_words.txt
    69 extra_words.txt
    mhr@jukebox:~/webos$ wc -l new_autoreplace
    168177 new_autoreplace
    mhr@jukebox:~/webos$


    cheers,
    -mark
  11. diomark's Avatar
    Posts
    752 Posts
    Global Posts
    770 Global Posts
    #71  
    Quote Originally Posted by diomark View Post
    My typing is much more accurate with this hack, but using the googwords, browser card takes an extra 15 seconds to open, and emails an extra 25 seconds

    Has anyone found a smaller list with better results?
    btw - to be specific -

    browser card opens immediately. actually using a site (bookmark or search) takes an extra ~15 seconds. (confirmed by replacing w/ the original autoreplace file)

    Email header comes up asap. But the body takes an extra 25 seconds unusable as it is..
    -mark
  12. diomark's Avatar
    Posts
    752 Posts
    Global Posts
    770 Global Posts
    #72  
    I'm going to try a quick test based on this list (500 words) - will post the result here.

    THE MOST COMMON WORDS IN ENGLISH

    -mark
  13.    #73  
    Hmm, so our command lines look the same, but your output file has twice as many lines as mine. I'm going to have to try this on Linux and see what happens...
  14. diomark's Avatar
    Posts
    752 Posts
    Global Posts
    770 Global Posts
    #74  
    Quote Originally Posted by dimfeld View Post
    Hmm, so our command lines look the same, but your output file has twice as many lines as mine. I'm going to have to try this on Linux and see what happens...
    line handling issues still?

    While looking at the top500 words, I'm finding some weirdness in here.. specifically, words in the extra_words.txt are showing up weird like this -
    yesterday's in here -


    testerdatesterday|yesterday
    testerday|yesterday
    uesterdauesterday|yesterday
    uesterday|yesterday
    yeaterday|yesterday
    yeaterdayeaterday|yesterday
    yedterday|yesterday
    yedterdayedterday|yesterday
    yesrerday|yesterday
    yesrerdayesrerday|yesterday
    yesteeday|yesterday
    yesteedayesteeday|yesterday
    yesterdat|yesterday
    yesterdau|yesterday
    yesterdayesterdat|yesterday
    yesterdayesterdau|yesterday
    yesterdsy|yesterday
    yesterdsyesterdsy|yesterday
    yesterfay|yesterday
    yesterfayesterfay|yesterday
    yestersay|yesterday
    yestersayestersay|yesterday
    yestetday|yesterday
    yestetdayestetday|yesterday
    yestrrday|yesterday
    yestrrdayestrrday|yesterday
    yestwrday|yesterday
    yestwrdayestwrday|yesterday
    yesyerday|yesterday
    yesyerdayesyerday|yesterday
    yrsterday|yesterday
    yrsterdayrsterday|yesterday
    ywsterday|yesterday
    ywsterdaywsterday|yesterday


    Any ideas?

    cheers,
    -mark
  15.    #75  
    Ok, so I tried it on Linux and the output file is still the same size as on Windows for me. I'm not sure offhand what the problem might be, but here's an updated script that uses the rstrip function to get rid of newlines. Maybe that will help some?
    New Autoreplace Script

    Also, when you pass in the existing text-edit-autoreplace to the script, is that the original Pre autoreplace file that's around 46KB, or is it the new large one? I'm using the original 46KB file there.

    EDIT: Just saw your last post after posting this. Does the same thing happen with the new script? That should solve all newline handling issues.
  16. diomark's Avatar
    Posts
    752 Posts
    Global Posts
    770 Global Posts
    #76  
    Your new script seems to work better in linux.. I can confirm that the problem I reported above (with yesterday and other words in extra_words.txt) doesn't exist, and the file is only 1183928 (vs 2.7mb's..)

    will test this on the pre.
    -mark
  17. diomark's Avatar
    Posts
    752 Posts
    Global Posts
    770 Global Posts
    #77  
    Quote Originally Posted by dimfeld View Post
    Also, when you pass in the existing text-edit-autoreplace to the script, is that the original Pre autoreplace file that's around 46KB, or is it the new large one? I'm using the original 46KB file there.
    46kb as well here.
    -m
  18. diomark's Avatar
    Posts
    752 Posts
    Global Posts
    770 Global Posts
    #78  
    Is there a way to test this on the device without rebooting the pre?
    -m
  19.    #79  
    Quote Originally Posted by diomark View Post
    Is there a way to test this on the device without rebooting the pre?
    -m
    There might be, but as far as I know you have to reboot.
  20. diomark's Avatar
    Posts
    752 Posts
    Global Posts
    770 Global Posts
    #80  
    (My gmail 24hour sync decided to syncup while I was testing this.. will wait for it to finish (I get ~500 messages /day in that box) before doing anything else)

Tags for this Thread

Posting Permissions