Topics

Converting JAWS Dictionaries


Pele West
 

Hi Everyone

Is there a way of converting JAWS Pronunciation dictionaries to NVDA
dictionaries?

Thanks

Pele West


 

Pele,

            None that I know of.  But the structure of the NVDA dictionary files is very, very simple.  If you happen to have a JAWS dictionary file that you can e-mail me as an attachment, I might be able to whip something up, though I won't guarantee it.  I don't know how Freedom Scientific lays out the JAWS dictionary files, and that's where the sticking point might be.  But if they're plain text like the NVDA dictionaries are, it shouldn't be all that difficult to figure out how to convert.

            It's going to be a long, rainy day here, so this might be a fun task to try.  By the way, anyone who has a JAWS dictionary file they're willing to share should feel free to e-mail it to me as an attachment using the Reply to Sender link at the end of this message.

--

Brian - Windows 10 Pro, 64-Bit, Version 1909, Build 18363  

Science has become just another voice in the room; it has lost its platform.  Now, you simply declare your own truth.

      ~ Dr. Paul A. Offit, in New York Times article, How Anti-Vaccine Sentiment Took Hold in the United States, September 23, 2019

 

 


Pele West
 

hi Brian

Thanks for writing.

As you say it is so easy, I will have a go at converting the
dictionaries myself.

Pele


 

Pele,

           If you happen to know Linux/Unix, the awk command will be very easy to use to do the conversion.

A JAWS dictionary line has the fields, in order:  Actual word|Replacement Word|Language (0x09-English)|Synthesizer|Voice|Language (0 is default)|Case sensitive (0 - No, 1 -Yes)
For the JAWS dictionary, the fields themselves are DOT enclosed/separated, UNLESS, the actual word starts with a dot, in which case it's a comma used for the separator.

The NVDA dictionary file layout is simpler, and the fields are tab separated.  The first field does not have a separator before it.  They are arranged as:
Matching word or regular expression|Replacement|Case Sensitivity (0 - no, 1 - yes)|Type (2 - Whole Word, 1 - Regular Expression, 0 - Anywhere)|Comment (All comments are pre-pended with a #

For both dictionaries, there is a single entry per line of the file.

Have fun with it.  It doesn't appear to be particularly complicated.
--

Brian - Windows 10 Pro, 64-Bit, Version 1909, Build 18363  

Science has become just another voice in the room; it has lost its platform.  Now, you simply declare your own truth.

      ~ Dr. Paul A. Offit, in New York Times article, How Anti-Vaccine Sentiment Took Hold in the United States, September 23, 2019

 

 


Bernd Dorer
 

Hi, may someone could write this as an Add-on for NVDA? This would be really cool!


kind regards
Bernd
Am 23.04.2020 um 19:52 schrieb Brian Vogel:

Pele,

           If you happen to know Linux/Unix, the awk command will be very easy to use to do the conversion.

A JAWS dictionary line has the fields, in order:  Actual word|Replacement Word|Language (0x09-English)|Synthesizer|Voice|Language (0 is default)|Case sensitive (0 - No, 1 -Yes)
For the JAWS dictionary, the fields themselves are DOT enclosed/separated, UNLESS, the actual word starts with a dot, in which case it's a comma used for the separator.

The NVDA dictionary file layout is simpler, and the fields are tab separated.  The first field does not have a separator before it.  They are arranged as:
Matching word or regular expression|Replacement|Case Sensitivity (0 - no, 1 - yes)|Type (2 - Whole Word, 1 - Regular Expression, 0 - Anywhere)|Comment (All comments are pre-pended with a #

For both dictionaries, there is a single entry per line of the file.

Have fun with it.  It doesn't appear to be particularly complicated.
--

Brian - Windows 10 Pro, 64-Bit, Version 1909, Build 18363  

Science has become just another voice in the room; it has lost its platform.  Now, you simply declare your own truth.

      ~ Dr. Paul A. Offit, in New York Times article, How Anti-Vaccine Sentiment Took Hold in the United States, September 23, 2019

 

 



 

This needs to go to the Chat Subgroup, or private, but I have written a Windows batch file that can do this.  It does, however, require that you have installed Cygwin, which gives access to a number of common Unix/Linux utilities, and I'm using gawk to accomplish the task.

If anyone wants to discuss this further, please either contact me privately or start a new topic on the Chat Subgroup.  At the moment, what I have is NOT language specific, and since JAWS mixes all active languages (or at least can) in a single dictionary, using a field to differentiate which is which, there would need to be a way to filter entries only for a specific language (e.g. English, French) and those that are not specific to a language.

This discussion is way too "under the hood" to belong on the main NVDA group
--

Brian - Windows 10 Pro, 64-Bit, Version 1909, Build 18363  

Science has become just another voice in the room; it has lost its platform.  Now, you simply declare your own truth.

      ~ Dr. Paul A. Offit, in New York Times article, How Anti-Vaccine Sentiment Took Hold in the United States, September 23, 2019

 

 


Pele West
 

Hi Brian

Thank you so much for this.

I am using VBA to do the conversion, which is quite easy. It seems like
a good project while we are locked down.

I am also converting my Window-Eyes dictionaries (apologies for
mentioning an obsolete screenreader).

Pele


Bernd Dorer
 

Hi Pele,

well, this in an addon would help many JAWS and Windows Eyes users to
switch to NVDA. Alternatively could s.o. write a guid to doing this and
attach this to the NVDA comunity wiki.

kind regards
Bernd

Am 24.04.2020 um 10:23 schrieb Pele West:

Hi Brian

Thank you so much for this.

I am using VBA to do the conversion, which is quite easy. It seems like
a good project while we are locked down.

I am also converting my Window-Eyes dictionaries (apologies for
mentioning an obsolete screenreader).

Pele




 

I could still use some additional test cases - JAWS dictionaries that I can convert.   If anyone has a JDF file, preferably one that is either in English or where the JAWS language is "any" that they'd like converted, send me the JDF file using the Reply to Sender link at the bottom of this message.

--

Brian - Windows 10 Pro, 64-Bit, Version 1909, Build 18363  

Science has become just another voice in the room; it has lost its platform.  Now, you simply declare your own truth.

      ~ Dr. Paul A. Offit, in New York Times article, How Anti-Vaccine Sentiment Took Hold in the United States, September 23, 2019

 

 


Ron Kolesar
 

I too agree.
I have several different JAWS dictionary folders, and since I’m now learning NVDA, it would be nice to have NVDA pronounce words correctly.
Many Thanks.
Ron KR3DOG
 

From: Bernd Dorer
Sent: Thursday, April 23, 2020 14:17
To: nvda@nvda.groups.io
Subject: Re: [nvda] Converting JAWS Dictionaries
 
Hi, may someone could write this as an Add-on for NVDA? This would be really cool!


kind regards
Bernd
Am 23.04.2020 um 19:52 schrieb Brian Vogel:
Pele,

           If you happen to know Linux/Unix, the awk command will be very easy to use to do the conversion.

A JAWS dictionary line has the fields, in order:  Actual word|Replacement Word|Language (0x09-English)|Synthesizer|Voice|Language (0 is default)|Case sensitive (0 - No, 1 -Yes)
For the JAWS dictionary, the fields themselves are DOT enclosed/separated, UNLESS, the actual word starts with a dot, in which case it's a comma used for the separator.

The NVDA dictionary file layout is simpler, and the fields are tab separated.  The first field does not have a separator before it.  They are arranged as:
Matching word or regular expression|Replacement|Case Sensitivity (0 - no, 1 - yes)|Type (2 - Whole Word, 1 - Regular Expression, 0 - Anywhere)|Comment (All comments are pre-pended with a #

For both dictionaries, there is a single entry per line of the file.

Have fun with it.  It doesn't appear to be particularly complicated.
--

Brian - Windows 10 Pro, 64-Bit, Version 1909, Build 18363  

Science has become just another voice in the room; it has lost its platform.  Now, you simply declare your own truth.

      ~ Dr. Paul A. Offit, in New York Times article, How Anti-Vaccine Sentiment Took Hold in the United States, September 23, 2019

 

 


In the good old days of Morse code Shorthand, 73's AKA Best Regards and or Best Whishes,From
Ron Kolesar
Volunteer Certified Licensed Emergency Communications Station
And
Volunteer Certified Licensed Ham Radio Station
With the Call Sign of KR3DOG


 

Again, from the one person I've done a conversion for, the report is that it worked for some words/phrases but for a few it didn't.  Until or unless I can get enough test cases, and those who get the conversions back and put them in can report what's not working so I can attempt to figure out why, there's not much more I can do.

I do not code in Python, so I won't be doing this as an NVDA add-on.  Given the nature of what's needed, a stand-alone utility (which is what I've tried to create) should be fine, anyway, as this is generally a "one and done" event for a JAWS dictionary to be converted for appending to the appropriate NVDA dictionary.  As the utility currently stands, it will only handle language agnostic JAWS dictionaries reliably.  That means the language for the given word/phrase is not specified.  It's a simple pattern match.

I will not distribute the results to others as far as the code goes until I'm satisfied that I have something that "is bulletproof" and I don't have that yet and can't get there without test cases.

So, if you have dictionaries that are language agnostic that you'd like converted get in touch with me privately and we can work together.
--

Brian - Windows 10 Pro, 64-Bit, Version 1909, Build 18363  

Science has become just another voice in the room; it has lost its platform.  Now, you simply declare your own truth.

      ~ Dr. Paul A. Offit, in New York Times article, How Anti-Vaccine Sentiment Took Hold in the United States, September 23, 2019

 

 


Luke Davis
 

On Sun, 3 May 2020, Brian Vogel wrote:

much more I can do.I do not code in Python, so I won't be doing this as an NVDA add-on.  Given the nature of what's needed, a stand-alone utility (which is
what I've tried to create) should be fine, anyway, as this is generally a "one and done" event for a JAWS dictionary to be converted for appending to the
appropriate NVDA dictionary.
Is that a reasonable assumption though?
How many people use both Jaws and NVDA?
Haven't you yourself advised having more than one screen reader, on the multiple tools theory?

Wouldn't it therefore make more sense as an add-on that scans Jaws dictionaries on the system, and updates NVDA's dictionary on startup?
Or vice-versa, if that's possible?

Luke


 

On Wed, May 6, 2020 at 01:52 AM, Luke Davis wrote:
Is that a reasonable assumption though?
Sure it is.  But even if it isn't, I don't work for NVAccess, and I don't code in python.  When someone's offered to do something, for free, that you want done then you can either accept or decline.  If someone else wants to step up to do this, then have at it.

I presume someone asking for this sort of thing has been a long term JAWS user, who has compiled an extensive dictionary, and does not want to have to start over again from scratch when starting to use NVDA.  So far, based on those who've contacted me, that presumption has been valid.

And you've been around long enough to have seen me make the statement that any screen reader user should have at least basic competence in a screen reader other than their preferred one.  With NVDA and Narrator available, JAWS users have no excuse, and with Narrator available NVDA users have no excuse.  There are times when any screen reader "doesn't play well" with a given thing that another screen reader will.  It's gonna happen, so it makes sense to be prepared for when it does.
 
--

Brian - Windows 10 Pro, 64-Bit, Version 1909, Build 18363  

Science has become just another voice in the room; it has lost its platform.  Now, you simply declare your own truth.

      ~ Dr. Paul A. Offit, in New York Times article, How Anti-Vaccine Sentiment Took Hold in the United States, September 23, 2019