Polish dictionary

San Zamoyski · 05-15-2016, 01:43 PM

Hi!

I'm building polish dictionary for my kindle 3 (and will share it somewhere...). But i've found one quite annoying problem:
there are for example two words: laskę and łaskę.
first one is recognised well, but second one shows definition for laskę, but the meaning should be different.

can i make this dictionary work with this kind of words?

regards!

my opf file:

Code:

<?xml version="1.0"?><!DOCTYPE package SYSTEM "oeb1.ent">

<!-- the command line instruction 'prcgen dictionary.opf' will produce the dictionary.prc file in the same folder-->
<!-- the command line instruction 'mobigen dictionary.opf' will produce the dictionary.mobi file in the same folder-->

<package unique-identifier="uid" xmlns:dc="Dublin Core">

    <metadata>
        <dc-metadata>
                <dc:Identifier id="uid">Słownik Języka Polskiego, sjp.pl</dc:Identifier>
                <!-- Title of the document -->
                <dc:Title><h2>[ZM] sjp.pl (2016-05-14b)</h2></dc:Title>
                <dc:Language>pl-PL</dc:Language>
        </dc-metadata>
        <x-metadata>
                <DictionaryInLanguage>pl-PL</DictionaryInLanguage>
                <DictionaryOutLanguage>pl-PL</DictionaryOutLanguage>
        </x-metadata>
    </metadata>
    <!-- list of all the files needed to produce the .prc file -->
    <manifest>
        <item id="dictionary0" href="slo0.html" media-type="text/x-oeb1-document"/>
(...)
        <item id="dictionary25" href="slo25.html" media-type="text/x-oeb1-document"/>
    </manifest>
        <!-- list of the html files in the correct order  -->
    <spine>
            <itemref idref="dictionary0"/>
            <itemref idref="dictionary1"/>
(...)
            <itemref idref="dictionary25"/>
    </spine>
    <guide> 
        <reference type="search" title="Dictionary Search" onclick= "index_search()"/> 
    </guide>
</package>

and łaska sample.

Code:

<?xml version="1.0" encoding="utf-8"?>
<html xmlns:idx="www.mobipocket.com" xmlns:mbp="www.mobipocket.com" xmlns:xlink="http://www.w3.org/1999/xlink">
    <body>
        <mbp:pagebreak/>
        <mbp:frameset>
            <mbp:slave-frame display="bottom" device="all" breadth="auto" leftmargin="0" rightmargin="0" bottommargin="0" topmargin="0">
                <div align="center" bgcolor="yellow"/>
                    <a onclick="index_search()">Dictionary Search</a>
                </div>
            </mbp:slave-frame>
            <mbp:pagebreak/>
(...)
<idx:entry name="word" scriptable="yes">
                        <h2>
                            <idx:orth>
                                łaska
                                <idx:infl inflgrp="odmiany">
                                    <idx:iform name="odmiana" value="łasce" />
                                    <idx:iform name="odmiana" value="łask" />
                                    <idx:iform name="odmiana" value="łaskach" />
                                    <idx:iform name="odmiana" value="łaskami" />
                                    <idx:iform name="odmiana" value="łaską" />
                                    <idx:iform name="odmiana" value="łaskę" />
                                    <idx:iform name="odmiana" value="łaski" />
                                    <idx:iform name="odmiana" value="łasko" />
                                    <idx:iform name="odmiana" value="łaskom" />
</idx:infl>
                            </idx:orth>
                            <idx:key key="łaska">
                        </h2>
                        <br/>
                        łaskawość, miłość, przychylność<br/>
ssak, mały drapieżnik z rodziny łasicowatych; łasica<br/>
przymiotnik od: Łask, rzadziej od: Łazy<br/>
miasto w Polsce<br/>
nazwisko
                    </idx:entry>
                    <hr />

(...)
<idx:entry name="word" scriptable="yes">
                        <h2>
                            <idx:orth>
                                laska
                                <idx:infl inflgrp="odmiany">
                                    <idx:iform name="odmiana" value="lasce" />
                                    <idx:iform name="odmiana" value="lasek" />
                                    <idx:iform name="odmiana" value="laskach" />
                                    <idx:iform name="odmiana" value="laskami" />
                                    <idx:iform name="odmiana" value="laską" />
                                    <idx:iform name="odmiana" value="laskę" />
                                    <idx:iform name="odmiana" value="laski" />
                                    <idx:iform name="odmiana" value="lasko" />
                                    <idx:iform name="odmiana" value="laskom" />

                                </idx:infl>
                            </idx:orth>
                            <idx:key key="laska">
                        </h2>
                        <br/>   
                        rodzaj kijka używanego przez osoby starsze oraz niepełnosprawne do podpierania się w czasie chodzenia$
książkowo: symbol niektórych godności i dostojeństw;<br/>
długi i cienki pręt czego;<br/> 
potocznie: członek męski;<br/>  
środowiskowo: ocena niedostateczna;<br/>
młodzieżowo o bardzo zgrabnej, atrakcyjnej dziewczynie; dziunia, lacha, lachon<br/>
gwary laskie - mieszane gwary polsko-czeskie (w rejonie Raciborza)<br/>
nazwisko<br/>
nazwisko
                    </idx:entry>
                    <hr />

        </mbp:frameset>
    </body>
</html>

edit.
adding <output encoding="utf-8" flatten-dynamic-dir="yes"/> won't help.

San Zamoyski · 05-18-2016, 03:56 AM

bump?

knc1 · 05-18-2016, 09:22 AM

Maybe nobody knows the answer.

San Zamoyski · 05-18-2016, 11:35 AM

but it might be that someone knows that it is not possible at all...

Galunid · 05-18-2016, 02:00 PM

I'd suggest checking this: http://1manfactory.com/create-your-o...uage-for-free/
Although I'm not sure, whether it'll help you. Also check the file encoding.
<output encoding="utf-8" flatten-dynamic-dir="yes"/> Should stay.

baf · 05-19-2016, 11:27 AM

It will not work in search box, as Kindle uses fuzzy search. However it should work in a pop-up box, looking up highlighted word. At least it works on my Kindle Touch.

You could also experiment with "exact" keyword, but I don't see any difference when using it. See Kindle Publishing Guidelance:

Quote:

By default, the Kindle device uses a fuzzy algorithm for matching diacritics during word lookup. Languages that use contrastive diacritics to distinguish between distinct word forms should use the exact="yes" attribute in the <idx:iform /> tag to force exact match of diacritics during lookup.

Also note that

Code:

<idx:key>..</idx:key>

tags are depreciated according to the guidelines.

Doitsu · 05-20-2016, 01:28 PM

You'll need to add the exact="yes" attribute:

For example:

Code:

<idx:iform value="łasce" exact="yes" />

There were also some syntax errors in your examples. I attached a working dictionary, which differentiates between łaska and laska.

Spoiler:

Code:

<html xmlns:idx="http://www.mobipocket.com/idx" xmlns:mbp="www.mobipocket.com">
<head>
  <title></title>
  <meta http-equiv="Content-Type" content="text/html; charset=utf-8" />
</head>

<body>
<mbp:frameset>

    <idx:entry name="polish" scriptable="yes">
        <idx:orth>
            łaska
            <idx:infl>
                <idx:iform value="łasce" exact="yes" />
                <idx:iform value="łask" exact="yes" />
                <idx:iform value="łaskach" exact="yes" />
                <idx:iform value="łaskami" exact="yes" />
                <idx:iform value="łaską" exact="yes" />
                <idx:iform value="łaskę" exact="yes" />
                <idx:iform value="łaski" exact="yes" />
                <idx:iform value="łasko" exact="yes"  />
                <idx:iform value="łaskom" exact="yes" />
            </idx:infl>
        </idx:orth>
            <br/>
            łaskawość, miłość, przychylność<br/>
            ssak, mały drapieżnik z rodziny łasicowatych; łasica<br/>
            przymiotnik od: Łask, rzadziej od: Łazy<br/>
            miasto w Polsce<br/>
            nazwisko
    </idx:entry>

    <p>-------------------------------------</p>
    
    <idx:entry name="polish" scriptable="yes">
        <idx:orth>
            laska
            <idx:infl>
                <idx:iform value="lasce" exact="yes" />
                <idx:iform value="lasek" exact="yes" />
                <idx:iform value="laskach" exact="yes" />
                <idx:iform value="laskami" exact="yes" />
                <idx:iform value="laską" exact="yes" />
                <idx:iform value="laskę" exact="yes" />
                <idx:iform value="laski" exact="yes" />
                <idx:iform value="lasko" exact="yes" />
                <idx:iform value="laskom" exact="yes" />
            </idx:infl>
        </idx:orth>
        <br/>   
        rodzaj kijka używanego przez osoby starsze oraz niepełnosprawne do podpierania się w czasie chodzenia<br/>
        książkowo: symbol niektórych godności i dostojeństw;<br/>
        długi i cienki pręt czego;<br/> 
        potocznie: członek męski;<br/>  
        środowiskowo: ocena niedostateczna;<br/>
        młodzieżowo o bardzo zgrabnej, atrakcyjnej dziewczynie; dziunia, lacha, lachon<br/>
        gwary laskie - mieszane gwary polsko-czeskie (w rejonie Raciborza)<br/>
        nazwisko<br/>
        nazwisko
    </idx:entry>
    
</mbp:frameset>
</body>
</html>

baf · 05-20-2016, 05:38 PM

Quote:

Originally Posted by Doitsu

You'll need to add the exact="yes" attribute:

Are you sure this tag makes any difference? IIRC I tested two samples, with 'exact' tag and without it, and the compiled mobi files didn't differ (apart from different uids etc).

Doitsu · 05-20-2016, 05:49 PM

Quote:

Originally Posted by baf

Are you sure this tag makes any difference?

This information comes directly from the Kindle Publishing Guidelines.

Quote:

Originally Posted by baf

IIRC I tested two samples, with 'exact' tag and without it, and the compiled mobi files didn't differ (apart from different uids etc).

If you used KindleUnpack, you can't be sure since it'll only output known attributes and exact="yes" wasn't part of the original Mobipocket specs.

Did you also test the two variations on an actual Kindle and did you get the same results for both versions? (I only tested my version and it worked.)

baf · 05-21-2016, 05:13 AM

Quote:

Originally Posted by Doitsu

If you used KindleUnpack, you can't be sure since it'll only output known attributes and exact="yes" wasn't part of the original Mobipocket specs.

I compared compiled binary files.

Quote:

Originally Posted by Doitsu

Did you also test the two variations on an actual Kindle and did you get the same results for both versions? (I only tested my version and it worked.)

I did some tests on your opf. One test with exact keyword and one without. It confirmed that the binaries don't differ in indices section, nor in metadata, except for some unique ids.

Both dictionaries work the same on my KT.
When searching from the search box (index items) it doesn't recognise different forms of the words. Whatever I type in the search box it brings first entry: "łaska". It is probably limitation of the reader's fuzzy search.
At the same time both dictionaries work correctly when highlighting word in a document. Highlighting "łaska" brings pop-up definition for "łaska", highlighting "laska" shows "laska".

Maybe some other conditions must be fulfilled in order to make the keyword be used.

San Zamoyski · 05-23-2016, 01:59 PM

Hi!

Sorry for delay with answer. I've generated new version of my dictionary, but still, no luck (both definitions from laska).
exact="yes" (to evary one) seems to make no difference.

I've tested Doitsu dictionary - still no luck. Both words Laskę and Łaskę recognised as łaska.

It seems that in-file order makes difference.

Thanks, all of You, for helping solving this problem!

Best regards!

knc1 · 05-23-2016, 03:28 PM

Quote:

Originally Posted by San Zamoyski

Hi!

Sorry for delay with answer. I've generated new version of my dictionary, but still, no luck (both definitions from laska).
exact="yes" (to evary one) seems to make no difference.

I've tested Doitsu dictionary - still no luck. Both words Laskę and Łaskę recognised as łaska.

It seems that in-file order makes difference.

Thanks, all of You, for helping solving this problem!

Best regards!

In that case, I would re-consider the test methods your using.

San Zamoyski · 05-23-2016, 04:52 PM

Quote:

In that case, I would re-consider the test methods your using.

Of creating or testing dictionary?

quiris · 05-23-2016, 04:57 PM

I also can confirm that exact="yes" does not work as described in Kindle Publishing Guidelines.

Doitsu · 05-24-2016, 12:41 AM

Quote:

Originally Posted by San Zamoyski

I've tested Doitsu dictionary - still no luck. Both words Laskę and Łaskę recognised as łaska.

Maybe it's a firmware issue. I only successfully tested it on my PW2.
Since the Kindle Publishing Guidelines specifically mention the Spanish words una and uña, can you create a short Spanish book with both words, download one of the free Spanish dictionaries from Amazon, e.g. the DRAE, and look up both words? If you get the definition of the unaccented word (una) both times it's definitely a firmware issue.

Quote:

Originally Posted by quiris

I also can confirm that exact="yes" does not work as described in Kindle Publishing Guidelines.

Maybe this feature is planned for the next KindleGen version and a tech writer has prematurely included it or maybe they updated the index generation code a long time ago so that it defaults to exact="yes".
Since you own both a Kindle 4 and a PW1, can you test my dictionary on both machines?

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
GUI [Polish] vs. CLI ebook-polish	phossler	Calibre	6	10-01-2015 08:27 AM
E-reader with Dutch/English or Dutch/Polish dictionary	tttx	Which one should I buy?	17	08-20-2015 05:42 AM
Dictionary question: changing word delimiters for french dictionary	oecherprinte	Amazon Kindle	1	05-09-2011 04:45 AM
Converting PWN Oxford Polish-English dictionary to Mobipocket	owl123	Workshop	1	05-24-2010 07:07 AM
Oxford built-in dictionary disappears after changing default dictionary	YYZscientist	Amazon Kindle	4	01-24-2010 08:42 PM

05-18-2016, 03:56 AM	#2
San Zamoyski Enthusiast Posts: 48 Karma: 41612 Join Date: Jan 2015 Location: Poland, Warsaw Device: Kindle 3G	bump?

05-18-2016, 09:22 AM	#3
knc1 Going Viral Posts: 17,212 Karma: 18210809 Join Date: Feb 2012 Location: Central Texas Device: No K1, PW2, KV, KOA	Maybe nobody knows the answer.

05-18-2016, 11:35 AM	#4
San Zamoyski Enthusiast Posts: 48 Karma: 41612 Join Date: Jan 2015 Location: Poland, Warsaw Device: Kindle 3G	but it might be that someone knows that it is not possible at all...

05-18-2016, 02:00 PM	#5
Galunid Zealot Posts: 122 Karma: 43580 Join Date: Apr 2016 Device: KPW3, Kobo Clara HD, Onyx Boox Nova 2	I'd suggest checking this: http://1manfactory.com/create-your-o...uage-for-free/ Although I'm not sure, whether it'll help you. Also check the file encoding. <output encoding="utf-8" flatten-dynamic-dir="yes"/> Should stay.

05-23-2016, 01:59 PM	#11
San Zamoyski Enthusiast Posts: 48 Karma: 41612 Join Date: Jan 2015 Location: Poland, Warsaw Device: Kindle 3G	Hi! Sorry for delay with answer. I've generated new version of my dictionary, but still, no luck (both definitions from laska). exact="yes" (to evary one) seems to make no difference. I've tested Doitsu dictionary - still no luck. Both words Laskę and Łaskę recognised as łaska. It seems that in-file order makes difference. Thanks, all of You, for helping solving this problem! Best regards!

05-23-2016, 04:57 PM	#14
quiris Groupie Posts: 195 Karma: 42216 Join Date: Oct 2013 Location: Poland Device: Kindles: KOA1, KV	I also can confirm that exact="yes" does not work as described in Kindle Publishing Guidelines.

Advert

Advert