summaryrefslogtreecommitdiff
path: root/contrib/eaglesconv/README
diff options
context:
space:
mode:
authorNick Frolov <nf@mkmks.org>2011-12-31 02:36:24 +0000
committerNick Frolov <nf@mkmks.org>2011-12-31 02:36:24 +0000
commit8e2c1823ed74b32d980b27ec2bd8972c80488309 (patch)
treed653937afe22f139612fa4efa079c94548bea80d /contrib/eaglesconv/README
parentba451d203aef642e98790e054f31c5aae6d8d861 (diff)
A Russian dictionary
A Russian dictionary generated from a wordlist created by the FreeLing project. The accompanying converter can be used to convert other wordlists in EAGLES format to GF grammars.
Diffstat (limited to 'contrib/eaglesconv/README')
-rw-r--r--contrib/eaglesconv/README24
1 files changed, 24 insertions, 0 deletions
diff --git a/contrib/eaglesconv/README b/contrib/eaglesconv/README
new file mode 100644
index 000000000..e3c84c61d
--- /dev/null
+++ b/contrib/eaglesconv/README
@@ -0,0 +1,24 @@
+How to use:
+
+1) Sort the wordlist so it can be split into sublists. It is necessary because
+the converter is quite memory-hungry, and you might not have enough RAM to
+process the whole wordlist at once.
+
+./CollectLemmas dicc.src | sort > lemmas.src
+
+2) Split the sorted wordlist.
+
+split -l 500000 lemmas.src
+
+3) Splitting has probably left forms of some lemmas spread across two
+sublists. Manually edit sublists so all forms for a lemma are present in just
+one sublist.
+
+4) Run the converter.
+
+./run_conv.sh xa*
+
+5) The converter has produced abstract and concrete syntaxes for the
+dictionary. You can try them out with GF:
+
+gf DictRus.gf \ No newline at end of file