tables recognition support

[yaleks]

Windows version is able to recognize tables.
In this project this feature disabled or additional information needed to enable?

[Alex Samorukov (samm-os2)]

table recognition is exists inside engine, and it supported at least in HTML and RTF output.
PUMA_Word32_Tables (later goes to gnTables) is responsible for this.
inside doc/pumaAPI.rtf I found description of the similar constant (in Russian, my translation) :
long Get/Set ModeTable – table recognition mode. 0 – tables are not selected, 1- columns and lines are selected, 2 – lines only, 3- no lines ( only for Apuma.dll )

The constant is used in file rshelllines.cpp, which, btw, tells that "Warning: RSL said that the lines do not have to withdraw." on any file. So i think that it should be enabled somewhere around.

[yaleks (yaleks)]

Я знаю русский и проблем с чтением оригинальной документации нет (I know russian and no problems with reading original docs).

OK, I look the code and try to enable this functions.

[Jussi Pakkanen (jpakkane)]

The "lines do not have to withdraw" thing was translated using Google translator. I have no idea what it actually means. Getting a translation for that would be nice. :)

[yaleks (yaleks)]

LDPUMA_Console("Предупреждение: RSL говорит, что снимать линии не надо.\n");

"снимать" is a jargon term in this context. My translation - "measure", may be "take" also appropriate term.

LDPUMA_Console("Warning: RSL said that the lines do not have to measure.\n");

[Alex Samorukov (samm-os2)]

btw, i found source of this error.
It comes from non initialized RVERLINE engine. To fix this we should just put
#define _USE_RVERLINE_
at the beginning of cuneiform_src/Kern/puma/c/initdone.cpp and and cuneiform_src/Kern/rstuff/sources/main/normalise.cpp.

This will enable rverline code (not sure what it is, but at least it not crashes).

[yaleks (yaleks)]

Some relevant files..

Kern/puma/c/debug.cpp:
Этап выделения таблиц. Выделение таблиц по линиям возможно \
только после выполнения этапа выделения линий."

Kern/usage/makeding.cpp:
II и III этапы алгоритма поиска таблиц.

Kern/rstuff/sources/main/Normalise.cpp:
Пропущен этап поиска таблиц

[Alex Samorukov (samm-os2)]

 "It is difficult to see a black cat in a dark room, especially if there is no cat"

I contacted by private mail one f the cuneiform developers, to ask him about tables support (because i was not able to find table selection code in layout() at all). Today i got an answer from him:

dvpsun@
В текущей версии выделение и распознавание таблиц отключено, т.к. открытие
этого набора исходников отложило бы опубликования всего кода под BSD-like
лицензией на длительный срок (не менее 9-12 месяцев). Открытие кода по
таблицам в планах стоит, но о сроках я ничего конкретного сказать не могу.

Translation:
In the current version of the code selection and recognition of the tables turned off, because opening
this set would have postponed the publication of source code under the whole BSD-like
license for a long period (not less than 9-12 months). Opening the code for
tables in the plans is, but terms are unknown.

2 JussiP i think it`s time to change this bug status to WONT`BE FIXED and to put an explanation inside readme for other project hackers.

[Jussi Pakkanen (jpakkane)]

So it shall be done.

[Polevoy Dmitry (openocr-polevoy)]

"снимать" в таком контектсте обычно означает физическое стирание линии с изображения (на пиксельном уровне), т.е. наиболее близкое "стирать"

Translation:
"снимать" in such context usualy means erase or clear line from the image (erase line pixels)
My translation - "clear" or "erase"

[Jussi Pakkanen (jpakkane)]

Thanks. I have updated the translation.

[Uriy Zenkov (zenkovyurij)]

Launch cuneiform
cuneiform
cuneiform: error while loading shared libraries: libpuma.so: cannot open shared object file: No such file or directory.
Then we must run ldconfig.

[Yury V. Zaytsev (zyv)]

What does this comment have to do with this bug?