[epic] Structured document translation
Open, MediumPublic
Actions

Assigned To

None

Authored By

	santhosh
	Aug 8 2023, 8:09 AM

Description

Along with plaintext translation, support for translating structured documents using MT system is a common and useful feature.

html
- Supported in MinT, uses a two stage beautifulsoup based parsing
- webpage by providing a URL is supported, but has performance issues and bugs with arbitrary webpage translation. Need more testing.
markdown
- supported in MinT by converting markdown to html and translating. The translated html is coverted back to markdown
- Need to explore if markdown parsers like marko etc can help to avoid going through html format. Need more testing
json
- supported in MinT, uses a two stage parsing. Translates values that are strings
svg
- supported in MinT, uses a two stage parsing. Translates text, textpath nodes. Does not support tspan yet.
Wikipedia i18n format strings. See T341544: Wikitext syntax is translated when requesting translations via MinT
Wikitext (T347018)
- The easiest way to support this is going through html format like markdown
OpenDocument formats like odt, odp, ods etc
MS Word
..

Related Objects
Search...

Status	Subtype	Assigned	Task
Open		None	T343782 [epic] Structured document translation
Open		None	T343784 MinT UI enhancements for structured document translation
Open		None	T348229 Different translation when same text is provided as html or plain text
Open		None	T353791 Improve MinT support for rich text
Resolved		santhosh	T363563 Avoid references losing their data (showing as plain-text "[1]") when added to the translation using MinT
Open		None	T347018 MinT support for Wikitext
Resolved		Pginer-WMF	T341544 Wikitext syntax is translated when requesting translations via MinT
Open	BUG REPORT	None	T347005 factorisation of tags is not to be used with format expressions
Open	BUG REPORT	None	T346720 suggestions are desynchronised when braces encountered {{ \| }}
Open	BUG REPORT	None	T346340 suggestions are desynchronised when several [[]] links $1 $2 $3... are present
Open	BUG REPORT	None	T346333 Section title markup "===" translates to "I am here" in some languages
Resolved		Wangombe	T349375 Avoid showing MinT suggestions in Translate for text with significant wikitext markup to break the translation
Open	BUG REPORT	None	T349470 unbalanced number of $xx parameters
Open	BUG REPORT	None	T349445 unclosed [[$1 at the end of message to translate
Open	BUG REPORT	None	T349438 Broken Markdown link syntax due to added whitespace
Open	BUG REPORT	None	T347826 simple $1 $2.. token list is suggested as currencies 1 $ 2 $ ....
Open	BUG REPORT	None	T348215 bad suggestion of single article with plural name form
Open	BUG REPORT	None	T347825 wrong split within well balanced tags
Open	BUG REPORT	None	T350460 Translating strings with variables in apostrophe produces unexpected output
Open	BUG REPORT	None	T350131 sub section name '==== XXX ====' is suggested as 'I am a man'
Open		None	T352692 Explore possible approaches to support wikitext in MinT
Resolved		santhosh	T368521 MinT leaks template metadata in translated content