Skip to content

Formats

Format Files Resources and notes
MDict .mdx .mdd, .1.mdd, … archives; .spx audio auto-converted to WAV
StarDict .ifo + .idx or .idx.gz + .dict or .dict.dz .syn synonyms; res/ folder or res.zip
Aard 2 .slob everything in the file; zlib, bz2 and LZMA2 compression
Lingvo DSL .dsl or .dsl.dz .dsl.files.zip; .ann annotations; UTF-8, UTF-16 and UTF-32 detected
Babylon .bgl everything in the file; source and target character sets detected
ZIM .zim everything in the file (Kiwix, Wikipedia, Wiktionary)
wudict markdown .wudict.md, .wudict.md.gz, .wudict.md.dz, or .md <name>.wudict.files folder or <name>.wudict.files.zip
DictShelf library text.db media.db in the same folder

Indexing

When DictShelf finds a dictionary, it queues a headword index in the background. Exact and prefix searches can use the dictionary's own format while that index is being created. The library index uses less memory and enables the optional contains and full-text searches.

ZIM is not added to the queue. Its own index answers exact and prefix search with little memory, and an indexed copy is several times the size of the file. The app does not create a headword index for ZIM.

The library

wudict markdown

A dictionary as one Markdown (CommonMark) file. The first line is # which is the title; the second line is wudict: 1, which is a marker for a DictShelf dictionary, so other .md files in a dictionary folder are ignored. Each entry is a ## headword heading followed by its text; further ## headings directly below it are alternative headwords.

# My Glossary
wudict: 1

## colour
## color

The property of an object that depends on the light it reflects.

dictshelf dump -format md writes any dictionary readable by DictShelf in this format.

Specification DictShelf dump

Audio

.mp3, .ogg and .wav play in the browser. Browsers cannot play Speex (.spx), so DictShelf converts it to WAV with the decoder compiled into the binary — in the Android app too. SPEEX_BACKEND = "external" uses the speexdec program instead.

No sound