Debian Codegraph @

data/README

raw
1
2
3 Language Directories
4 --------------------
5
6 ar -- Arabic. To be used with morphology engine; see morphology/ar
7
8 de -- German. Contains dictionaries for a few thousand words, that's
9 all. Incomplete.
10
11 en -- English. Complete set.
12
13 fa -- Farsi. To be used with morphology engine; see morphology/fa
14
15 he -- Hebrew. Experimental prototype. Contains a few dozen words.
16
17 id -- Indonesian. Experimental prototype. Contains a few hundred words.
18
19 kz -- Kazakh. Experimental prototype. Contains a few dozen words.
20
21 lt -- Lithuanian. Experimental prototype. Contains a few hundred words,
22 and implements only a few parts of speech.
23
24 ru -- Russian. Complete set of dictionaries, should parse most text.
25 Support needs to be added for conjunctions (and, or, but...)
26 Could use work on supporting regular expressions for numbers,
27 Russian-ized Latin technical terms, etc. See the English 4.0.regex
28 file for more.
29
30 th -- Thai. Complete set of dictionaries, consist of more than 100,000 words and should parse most text.
31
32 tr -- Turkish. Experimental prototype. Contains a few dozen words.
33
34 vn -- Vietnamese. Experimental prototype. Contains a few hundred words.
35
36 any -- Will parse "any" language, exploring all combinatoric possibilities.
37 This is useful for certain machine-learning tasks, when one might
38 want to iterate over all possible parse trees in every way.
39
40 ady -- Will parse "any" language, exploring all possible morphological
41 decompositions of the words in a sentence, as well as using
42 random linkages between words and morphemes. The assumed
43 morphology splits a word into at most two parts: a stem,
44 having no syntactic structure, and a suffix, which carries
45 the inflection -- that is, the stem carries all the syntactic
46 structure.
47
48 amy -- Will parse "any" language, exploring all possible morphological
49 decompositions of the words in a sentence, as well as using
50 random linkages between words and morphemes. Configured to
51 explore morphology of 3 or more parts: syntactically inactive
52 prefixes and stems, and a single syntactically active suffix.
53
54 demo-sql -- An almost empty demonstration dictionary, showing what an
55 SQLite-backed dictionary might look like.
56
57 demo-atomese -- A demonstration configuration for an AtomSpace-backed
58 dictionary. The lookup of words in the dictionary will be
59 performed by contacting the configured cogserver to obtain
60 disjunct data.
61
62 gen -- An empty template for autogenerated grammars. Part of the
63 language-learning project. This provide just enough boilerplate
64 to allow the `link-generator` binary to run.