fails to convert from older PAGE versions, fails to show cause

[bertsky]

I am trying to make use of the PRImA Layout Dataset, which is still in namespace version 2010-01-12, and thus needs to be converted to 2019.

However, the converter fails to run on some files, e.g. XML/00000122.xml, just saying:

Error writing target PAGE XML file

This, obviously, is not helpful at all.

I managed to get the program running in debug mode with Eclipse (by importing the PrimaText.jar from the release archive of the closed prima-text source).

Eclipse had trouble identifying the writer source file, but I was able to point it manually to XmlPageWriter_2019_07_15.java and set a breakpoint there.

It turns out that the validator was dissatisfied due to validation failures caused by @primaryScript not following the enum restriction:

cvc-enumeration-valid: Value 'Latin' is not facet-valid with respect to enumeration '[Adlm - Adlam, Afak - Afaka, Aghb - Caucasian Albanian, Ahom - Ahom, Tai Ahom, Arab - Arabic, Aran - Arabic (Nastaliq variant), Armi - Imperial Aramaic, Armn - Armenian, Avst - Avestan, Bali - Balinese, Bamu - Bamum, Bass - Bassa Vah, Batk - Batak, Beng - Bengali, Bhks - Bhaiksuki, Blis - Blissymbols, Bopo - Bopomofo, Brah - Brahmi, Brai - Braille, Bugi - Buginese, Buhd - Buhid, Cakm - Chakma, Cans - Unified Canadian Aboriginal Syllabics, Cari - Carian, Cham - Cham, Cher - Cherokee, Cirt - Cirth, Copt - Coptic, Cprt - Cypriot, Cyrl - Cyrillic, Cyrs - Cyrillic (Old Church Slavonic variant), Deva - Devanagari (Nagari), Dsrt - Deseret (Mormon), Dupl - Duployan shorthand, Duployan stenography, Egyd - Egyptian demotic, Egyh - Egyptian hieratic, Egyp - Egyptian hieroglyphs, Elba - Elbasan, Ethi - Ethiopic, Geok - Khutsuri (Asomtavruli and Nuskhuri), Geor - Georgian (Mkhedruli), Glag - Glagolitic, Goth - Gothic, Gran - Grantha, Grek - Greek, Gujr - Gujarati, Guru - Gurmukhi, Hanb - Han with Bopomofo, Hang - Hangul, Hani - Han (Hanzi, Kanji, Hanja), Hano - Hanunoo (Hanunóo), Hans - Han (Simplified variant), Hant - Han (Traditional variant), Hatr - Hatran, Hebr - Hebrew, Hira - Hiragana, Hluw - Anatolian Hieroglyphs, Hmng - Pahawh Hmong, Hrkt - Japanese syllabaries, Hung - Old Hungarian (Hungarian Runic), Inds - Indus (Harappan), Ital - Old Italic (Etruscan, Oscan etc.), Jamo - Jamo, Java - Javanese, Jpan - Japanese, Jurc - Jurchen, Kali - Kayah Li, Kana - Katakana, Khar - Kharoshthi, Khmr - Khmer, Khoj - Khojki, Kitl - Khitan large script, Kits - Khitan small script, Knda - Kannada, Kore - Korean (alias for Hangul + Han), Kpel - Kpelle, Kthi - Kaithi, Lana - Tai Tham (Lanna), Laoo - Lao, Latf - Latin (Fraktur variant), Latg - Latin (Gaelic variant), Latn - Latin, Leke - Leke, Lepc - Lepcha (Róng), Limb - Limbu, Lina - Linear A, Linb - Linear B, Lisu - Lisu (Fraser), Loma - Loma, Lyci - Lycian, Lydi - Lydian, Mahj - Mahajani, Mand - Mandaic, Mandaean, Mani - Manichaean, Marc - Marchen, Maya - Mayan hieroglyphs, Mend - Mende Kikakui, Merc - Meroitic Cursive, Mero - Meroitic Hieroglyphs, Mlym - Malayalam, Modi - Modi, Moḍī, Mong - Mongolian, Moon - Moon (Moon code, Moon script, Moon type), Mroo - Mro, Mru, Mtei - Meitei Mayek (Meithei, Meetei), Mult - Multani, Mymr - Myanmar (Burmese), Narb - Old North Arabian (Ancient North Arabian), Nbat - Nabataean, Newa - Newa, Newar, Newari, Nkgb - Nakhi Geba, Nkoo - N’Ko, Nshu - Nüshu, Ogam - Ogham, Olck - Ol Chiki (Ol Cemet’, Ol, Santali), Orkh - Old Turkic, Orkhon Runic, Orya - Oriya, Osge - Osage, Osma - Osmanya, Palm - Palmyrene, Pauc - Pau Cin Hau, Perm - Old Permic, Phag - Phags-pa, Phli - Inscriptional Pahlavi, Phlp - Psalter Pahlavi, Phlv - Book Pahlavi, Phnx - Phoenician, Piqd - Klingon (KLI pIqaD), Plrd - Miao (Pollard), Prti - Inscriptional Parthian, Rjng - Rejang (Redjang, Kaganga), Roro - Rongorongo, Runr - Runic, Samr - Samaritan, Sara - Sarati, Sarb - Old South Arabian, Saur - Saurashtra, Sgnw - SignWriting, Shaw - Shavian (Shaw), Shrd - Sharada, Śāradā, Sidd - Siddham, Sind - Khudawadi, Sindhi, Sinh - Sinhala, Sora - Sora Sompeng, Sund - Sundanese, Sylo - Syloti Nagri, Syrc - Syriac, Syre - Syriac (Estrangelo variant), Syrj - Syriac (Western variant), Syrn - Syriac (Eastern variant), Tagb - Tagbanwa, Takr - Takri, Tale - Tai Le, Talu - New Tai Lue, Taml - Tamil, Tang - Tangut, Tavt - Tai Viet, Telu - Telugu, Teng - Tengwar, Tfng - Tifinagh (Berber), Tglg - Tagalog (Baybayin, Alibata), Thaa - Thaana, Thai - Thai, Tibt - Tibetan, Tirh - Tirhuta, Ugar - Ugaritic, Vaii - Vai, Visp - Visible Speech, Wara - Warang Citi (Varang Kshiti), Wole - Woleai, Xpeo - Old Persian, Xsux - Cuneiform, Sumero-Akkadian, Yiii - Yi, Zinh - Code for inherited script, Zmth - Mathematical notation, Zsye - Symbols (Emoji variant), Zsym - Symbols, Zxxx - Code for unwritten documents, Zyyy - Code for undetermined script, Zzzz - Code for uncoded script, other]'. It must be a value from the enumeration.

This should be trivial.

So I wonder:

• why is this error not properly shown / propagated?
• why does the converter not prevent this in the first place (by mapping from the older ScriptSimpleType enum to the new one)?
• why not at least allow the user to keep the result file, even if it does not validate?

[bertsky]

Also, is there any workaround for this? It seems to be a catch-22 situation: Latin is not allowed on the output side, but Latn - Latin is not allowed on the input side (where it does show a decent error message BTW). I guess I'll just have to delete all @script, @primaryScript and @secondaryScript values, right?

[bertsky]

For the sake of completeness: the PRImA Layout Dataset contains more problems: Some regions have no Coords/Point whatsoever.