Providing derived formats is not necessary. BTW, the vert format should be derived from the annotated version.
missing bibl element in component files
I am not sure which metadata are available, but at least the newspaper title and date are available.
Documentation: https://clarin-eric.github.io/PressMint/#exa-sourceDescComp
Facsimiles
Are you planning to add links to facsimiles? I have checked the possible source (https://adt.arcanum.com/en/view/PestiHirlap_1884_01/), but only thumbs are available for free (https://static-cdn.arcanum.com/adt/pdf/PestiHirlap/PestiHirlap_1884_01/thumb/page0000.jpg)
paragraphs
It seems that you have created one paragraph per line, which is not correct, because a sentence can overflow over paragraphs, so the sentence segmentation will be wrong in the annotated version:
<p>Hát nem is volt az valami nehéz! Be van</p>
<p>ehhez már dresszirozva a municipiumi jegyzők</p>
<p>tolla !</p>for preserving line beginnings, you can use lb elements:
<lb/>Hát nem is volt az valami nehéz! Be van
<lb/>ehhez már dresszirozva a municipiumi jegyzők
<lb>tolla !And you can also add cb for columns, pb for pages. I believe you are starting from images, so it is a pity to lose this information
And it would be nice if you could recognize paragraphs. In CZ data, I am doing this based on indentation and line length analysis, through the floating window average length/indent counting.
weird characters
There are a lot of various characters, and by running
You can create tables with character stats, placed in the Sample/PressMint-HU folder
There are Greek/Cyrillic/Arabic and even Chinese characters.
Providing derived formats is not necessary. BTW, the
vertformat should be derived from the annotated version.missing bibl element in component files
biblin componentI am not sure which metadata are available, but at least the newspaper title and date are available.
Documentation: https://clarin-eric.github.io/PressMint/#exa-sourceDescComp
Facsimiles
Are you planning to add links to facsimiles? I have checked the possible source (https://adt.arcanum.com/en/view/PestiHirlap_1884_01/), but only thumbs are available for free (https://static-cdn.arcanum.com/adt/pdf/PestiHirlap/PestiHirlap_1884_01/thumb/page0000.jpg)
paragraphs
pelement should stand for paragraphs not linesIt seems that you have created one paragraph per line, which is not correct, because a sentence can overflow over paragraphs, so the sentence segmentation will be wrong in the annotated version:
for preserving line beginnings, you can use
lbelements:And you can also add
cbfor columns,pbfor pages. I believe you are starting from images, so it is a pity to lose this informationAnd it would be nice if you could recognize paragraphs. In CZ data, I am doing this based on indentation and line length analysis, through the floating window average length/indent counting.
weird characters
There are a lot of various characters, and by running
You can create tables with character stats, placed in the Sample/PressMint-HU folder
There are Greek/Cyrillic/Arabic and even Chinese characters.