[RELEASE] PSPedict V0.2 (Japanese Dictionary)

[PixelDragon]

Hi All

I'm back with an updated version of PSPedict, now V0.2.

Three extra edict files have now been included.
Computer Terms (10,547 entries)
Legal Terms ( 3,232 entries)
Life Sciences (31,382 entries)


PSPedict is now up to version 0.3
[RELEASE] PSPedict V0.3

Download:
Download PSPedict V0.2

Source Code:
Download PSPedict V0.2 Source Code
(I would suggest people new to programming C do not use my source as a learning tool,
basically because it will be a mess and my knowledge is not that great )

Original release (V0.1) thread
[RELEASE] PSPedict V0.1 (Japanese Dictionary)

Changelog for V0.2:
- Included support for other EDICT files, Computer Terms, Legal Terms
and Life Sciences. Unfortunately, Proper Names is just too large to
include. The extra EDICT files can be loaded via the Menu (the icon
is next to the exit icon and looks like a book with an up arrow).
Please be aware that study lists and found sets now are exclusive to
their own EDICT files. This means that when you have the Computer Terms
edict open, you can only view found sets and study lists relevant to that
EDICT file. This is because found sets and study lists use an index rather
than saving the actual contents of an entry. This is also to conserve memory.

- Added Ini file support to store search/sort options. If there are
any problems with the file, just delete "PSPedict.ini" in the data
folder and the file will be recreated when you next launch the app.

- Match Whole Word has been enabled for Kana/Kanji. What this basically
does is expect to match the whole string only. For example, if you search
for hiragana 'hi' with match hole word enabled, if a string has 'hima' the
match will fail. Strings consisting of 'hi' only will be matched.

- Fixed the screen flow slightly so it is more natural. For example, if you
exit out of the On Screen Keyboard (OSK), you will now return to the screen
where you opened the OSK.

- Deletion of previously saved found set files is now implimented. This can be
done from the "Load a Found Set" menu item.

Installation
============
See the readme in the download file for full information.
Simply copy the contents of the folder "PSPedict" to the
game folder of your psp, e.g. "X:\PSP\GAME" where X: is
the drive letter for your PSP. If you are asked to overwrite
any files, select yes/yes to all.

Please note, installing a new version will not overwrite your
saved found sets or study lists.

Todo list
=========
- Looking at the possibility of including support for Tanaka Corpus example scentences, can't
promise anything though.
- Tried looking into using the full OSK as suggested by tinmanx, but couldn't find a way.

Thanks/Credits
==============
Jim Breen for EDICT and KANJIDIC. (WWWJDIC: Word Search)
Brunni for OSLib. (home - OldSchool Library)
Sakya for OSLib Mod. (Sakya's Homepage)
BenHur for Intrafont. (intraFont and other PSP Releases by BenHur)
DAX for consistently pumping out quality CFW. Legend.
Anyone who has downloaded, supported, given feedback, etc...

Some Pictures







Comments and suggestions are welcome.

Enjoy!

[RyanF]

thanks for the update

[ymg545]

Thanks for the update. It's an awesome application.

Btw, you should remove the file "gpl.txt". If you want to release this application under GPL license. You should provide a way to get the source code.

[PixelDragon]

Originally Posted by ymg545

If you want to release this application under GPL license. You should provide a way to get the source code.

Thanks. I fully intended to release the source code, but I'm not a great programmer, so anyone looking at the code will probably say, "WTF?". I'll post a link to the source code shortly.

Cheers.

Edit: A link to the source code as been added to the first post. Thanks.

[thecobra]

Originally Posted by PixelDragon

Thanks. I fully intended to release the source code, but I'm not a great programmer, so anyone looking at the code will probably say, "WTF?". I'll post a link to the source code shortly.

Cheers.

Edit: A link to the source code as been added to the first post. Thanks.

It really look nice. By the way I thought Zero in hiragana was writing like this "zero " did not knew it could also be writing as "maru". hmm something new i learn today

[batanen]

Thank you for the changes and quick update! Now it works as I wished. I am a little busy at the moment but I found some time to make a few tests.

1) Match whole string -option works with hiragana now. Thank you very much! There is a small mistake: the text "English only" shouldn't be there because now it works with hiragana too.

2) I was quite satisfied to find out that there is a test which checks, whether the selected sjs-file is found. Though, there could be an additional option for canceling so the user is not forced to quit the application if the file is not found. If the user cancels, he can try to choose another file.

3) It may not be the best choice that the names of every sjs-file are determined beforehand. The program could search through the data folder for sjs-files when "Load an Edict file" is selected. Then the user can make his own dictionary-files regardless the name. At the moment the user may modify an existing sjs-file to make an own kind of dictionary.

Edit: A quick bug notice.
- Choose "Load an Edict file"
- Choose "legal terms", for example
- Select "Show Kanjidic"
- Search <hiragana>hi</hiragana> (when the option "Match whole string" is selected)
- Nothing is found. Conclusion, I think it always tries to search from the legal terms file even though the Kanjidic was selected. This might be the case with every other dictionary file, not counting the normal edict.

[PixelDragon]

@thecobra Thanks for you kind comment. I'm glad you learned something new

Originally Posted by batanen

1) Match whole string -option works with hiragana now. Thank you very much! There is a small mistake: the text "English only" shouldn't be there because now it works with hiragana too.

The "English only" was left in as it is now meant to refer to the use of the '*' as a wildcard character, which can only be used with English strings.

Originally Posted by batanen

2) ...there could be an additional option for canceling so the user is not forced to quit the application if the file is not found. If the user cancels, he can try to choose another file.

I agree it might be a good thing to do, but as it is only limited to the four EDICT files for now, it's probably not necessary.

Originally Posted by batanen

3) It may not be the best choice that the names of every sjs-file are determined beforehand. The program could search through the data folder for sjs-files when "Load an Edict file" is selected. Then the user can make his own dictionary-files regardless the name. At the moment the user may modify an existing sjs-file to make an own kind of dictionary.

A good point. I hadn't thought that anyone would want to create their own EDICT files. Actually, I'm not sure why someone would want to. I may think about this for a future version.

Originally Posted by batanen

Edit: A quick bug notice.
- Choose "Load an Edict file"
- Choose "legal terms", for example
- Select "Show Kanjidic"
- Search <hiragana>hi</hiragana> (when the option "Match whole string" is selected)
- Nothing is found. Conclusion, I think it always tries to search from the legal terms file even though the Kanjidic was selected. This might be the case with every other dictionary file, not counting the normal edict.

This is actually not a bug. The single string <hiragana>hi</hiragana> does not exist in the "legal terms" EDICT file. The search function only searches the EDICT file, not the Kanjidic file. Usually, if the kanji has a reading (as in "hi" for Day/Sun), searching in the general EDICT will probably find it. I didn't include Kanjidic in the search as this would generally be double handling.

Thanks again batanen, I appreciate your comments and feedback. Keep it coming.

[jojojoris]

Can you add an extra line in the window where you see the 2 Japanese and the English word underneath each other which contains the hepburn version (Hepburn romanization - Wikipedia, the free encyclopedia) of the Japanese text so me and the rest of the non-Japanese people know how to pronounce it?

I would like that. Me and my friends think japanese is way more cooler than dutch (my language),

[PixelDragon]

Hi jorjojoris. I could, but to be honest, a lot of Japanese learning resources do not promote Romaji (Japanese Romanization), and personally I agree with that. You can Learn Hiragana/Katakana basically in a weekend. Kanji is the hard part. I don't know anything about Dutch, but I think Japanese is pretty cool too.

[batanen]

Hi!

I tried to convert the Tanaka Corpus for PSPedict, but in the end I encountered too many problems. My main plan was
- to remove 1st three characters from every line.
- delete every second line.
- add the beginning slash '/' when char is 'A..Z', 'a..z', '(', ')' or '"' when found first time on the line.
- add the ending slash '/' when '#' is found. After that nothing is taken.

Problems were:
1) The script didn't quite work in every situation.
2) PSPedict doesn't like the UTF-8 file format. It uses some special file format, so actually the user cannot create own dictionaries, at least not very easily. This means it doesn't matter much if the file names are determined beforehand. So, you can forget what I said in the part 3.
3) The complete file was too big, PSPedict couldn't allocate enough memory.

In any case, while using PSPedict, I felt it would be nice if PSPedict could remember a previously done search input. In a case where you accidentally input a wrong character or want to search something similar after previous search, it would be frustrating to type everything all over again. If you find this idea useful, maybe a separate string for the Japanese and English text would work.

Here is the latest bug found. If the user press [X] in the Study List menu, the PSP will crash if the list is empty.

Originally Posted by PixelDragon

Thanks again batanen, I appreciate your comments and feedback. Keep it coming.

Thank you for your kind words. I am glad if I can be any help with testing and sharing my thoughts, programming is the difficult part.

[PixelDragon]

Originally Posted by batanen

I tried to convert the Tanaka Corpus for PSPedict, but in the end I encountered too many problems...

Problems were:
...
2) PSPedict doesn't like the UTF-8 file format. It uses some special file format, so actually the user cannot create own dictionaries, at least not very easily. This means it doesn't matter much if the file names are determined beforehand. So, you can forget what I said in the part 3.

The files I've used are in Shift JIS (SJIS) format. You can use a program like JWPce (Windows) to convert files to many different formats (e.g. Unicode, UTF-8, JIS, etc...). So, you probably could still make your own dictionary files, but for now, you would have to rename them to one of the existing EDICT files.

Originally Posted by batanen

3) The complete file was too big, PSPedict couldn't allocate enough memory.

I'm actually in the process (stopped to type this) of working on the Tanka Corpus (TC) example sentences implimentation. I'm just toying around with some ideas at the moment. I've been able to successfully parse the Tanaka Corpus file to work with the EDICT file. However, I've been using the subset which has approx 22,000 entries. With my testing, only 20,409 words out of 68,785 for the general EDICT would have example sentences. As this is quite small, I'm looking at using the main TC file which has over 150,000 entries. The file size is pretty big, so I'll have to work around this. I'd like to have maybe 2-3 example sentences per entry if possible. I'm fairly confident I'll be able to work something out. Still might take a while though.

Originally Posted by batanen

In any case, while using PSPedict, I felt it would be nice if PSPedict could remember a previously done search input...

Once again, a good idea. I'll try to impliment this for the next version.

Originally Posted by batanen

Here is the latest bug found. If the user press [X] in the Study List menu, the PSP will crash if the list is empty.

Fixed! I've updated the links at the top of the page to the newly compiled code that should fix that bug. Of interest, it also happened if you pressed [X] on an empty found set.
Should be OK now. You'll notice that I credit you for finding the bug

Originally Posted by batanen

Thank you for your kind words. I am glad if I can be any help with testing and sharing my thoughts, programming is the difficult part.

What can I say! Thanks again!!

[PixelDragon]

Just a quick note on the Tanaka Corpus example sentences. It seems that although the full file contains over 150,000 sentence pairs (approx 660,000 words), there is only at most a maximum of 27,000 unique words. This includes approx 3,000 words that would be proper names, so it looks like there will only be example sentences for approx 20,000 entries in the general EDICT file. I was a little disappointed by this, but I think it will still be worthwhile including example sentences for at least 20,000 dictionary words. Hey, it's better than nothing IMO!

[batanen]

Originally Posted by PixelDragon

The files I've used are in Shift JIS (SJIS) format. You can use a program like JWPce (Windows) to convert files to many different formats (e.g. Unicode, UTF-8, JIS, etc...). So, you probably could still make your own dictionary files, but for now, you would have to rename them to one of the existing EDICT files.

Thank you for the tip! JWPce worked well, though I was a little dissapointed with the limitations of SJIS-format (lack of some characters like ä and ö). Well, it's not that big problem so I might be playing with the idea. If you're wondering for what I might use an own dictionary, I could create a list of the words I learn during the Japanese courses. Then the words would have the exact form (usually the masu-form; though it's easy to make the masu-form from the dictionary form.) and a native language translation.

Originally Posted by PixelDragon

Fixed! I've updated the links at the top of the page to the newly compiled code that should fix that bug. Of interest, it also happened if you pressed [X] on an empty found set.
Should be OK now. You'll notice that I credit you for finding the bug

Oh, thank you... It works now.

Your progress with the Tanaka Corpus sounds very promising. I think 20000 entries are a lot. It's almost a third of the EDICT words.

[PixelDragon]

Hey everyone

Just wanted to quickly announce that I'm almost done implimenting the Tanaka Corpus example sentences. There will be a maximum of three sentences (I think that should be adequate) for words that have example sentences.

As stated in an earlier post, about 20,400 words will have at least one example. That's about a third of the general Edict file.

Here are some pictures of what it looks like...



In the main lists, words that have examples will be shown in blue text...


I've got to do some testing to make sure everything is working OK, so expect a new update soon.

[PixelDragon]

Hi all

PSPedict V0.3 has just been released. Link at the top of this page.

Thanks