Current section
Files
Jump to
Current section
Files
lib/cabo_cha.ex
defmodule CaboCha do
@moduledoc """
Elixir bindings for CaboCha, a Japanese dependency structure analyzer.
Parse function resturns a list of map.
The map's keys meaning is a follows.
- `chunk`: 文節(Chunk) -- This is map which includes follows.
+ `id`: 文節id(Chunk id)
+ `link`: 係り先の文節id(Linked chunk id)
+ `rel`: 不明(Unknown)
+ `head`: 主辞の形態素id(Head morpheme id)
+ `func`: 機能語の形態素id(Function word morpheme id)
+ `score`: 係り関係のスコア(Relational score)
- `morphs`: 分節の中の単語の形態素のリスト(List of morpheme in chunk) -- The list have few maps which includes follows.
+ `id`: 形態素id(Morpheme id)
+ `surface`: 表層系(Surface)
+ `pos`: 品詞(Part of speech)
+ `pos1`: 品詞細分類1(Part of speech subcategory1)
+ `pos2`: 品詞細分類2(Part of speech subcategory2)
+ `pos3`: 品詞細分類3(Part of speech subcategory3)
+ `conjugation_form`: 活用形(Conjugation form)
+ `conjugation`: 活用形(conjucation)
+ `base`: 基本形・原型(Lexical form)
+ `yomi`: 読み(Reading)
+ `pronunciation`: 発音(Pronunciation)
"""
@doc """
Parse given string and returns CaboCha's list.
If read_from_file is true, The file is parsed using str as the filename
## Examples
```elixir
iex> CaboCha.parse("太郎は花子が読んでいる本を次郎に渡した")
[
[
%{
"chunk" => %{
"func" => 1,
"head" => 0,
"id" => 0,
"link" => 5,
"rel" => "D",
"score" => -0.742128
},
"morphs" => [
%{
"base" => "太郎",
"conjugation" => "",
"conjugation_form" => "",
"id" => 0,
"pos" => "名詞",
"pos1" => "固有名詞",
"pos2" => "人名",
"pos3" => "名",
"pronunciation" => "タロー",
"surface" => "太郎",
"yomi" => "タロウ"
},
%{
"base" => "は",
"conjugation" => "",
"conjugation_form" => "",
"id" => 1,
"pos" => "助詞",
"pos1" => "係助詞",
"pos2" => "",
"pos3" => "",
"pronunciation" => "ワ",
"surface" => "は",
"yomi" => "ハ"
}
]
},
%{
"chunk" => %{
"func" => 3,
"head" => 2,
"id" => 1,
"link" => 2,
"rel" => "D",
"score" => 1.700175
},
"morphs" => [
%{
"base" => "花子",
"conjugation" => "",
"conjugation_form" => "",
"id" => 2,
"pos" => "名詞",
"pos1" => "固有名詞",
"pos2" => "人名",
"pos3" => "名",
"pronunciation" => "ハナコ",
"surface" => "花子",
"yomi" => "ハナコ"
},
%{
"base" => "が",
"conjugation" => "",
"conjugation_form" => "",
"id" => 3,
"pos" => "助詞",
"pos1" => "格助詞",
"pos2" => "一般",
"pos3" => "",
"pronunciation" => "ガ",
"surface" => "が",
"yomi" => "ガ"
}
]
},
%{
"chunk" => %{
"func" => 6,
"head" => 4,
"id" => 2,
"link" => 3,
"rel" => "D",
"score" => 1.825021
},
"morphs" => [
%{
"base" => "読む",
"conjugation" => "連用タ接続",
"conjugation_form" => "五段・マ行",
"id" => 4,
"pos" => "動詞",
"pos1" => "自立",
"pos2" => "",
"pos3" => "",
"pronunciation" => "ヨン",
"surface" => "読ん",
"yomi" => "ヨン"
},
%{
"base" => "で",
"conjugation" => "",
"conjugation_form" => "",
"id" => 5,
"pos" => "助詞",
"pos1" => "接続助詞",
"pos2" => "",
"pos3" => "",
"pronunciation" => "デ",
"surface" => "で",
"yomi" => "デ"
},
%{
"base" => "いる",
"conjugation" => "基本形",
"conjugation_form" => "一段",
"id" => 6,
"pos" => "動詞",
"pos1" => "非自立",
"pos2" => "",
"pos3" => "",
"pronunciation" => "イル",
"surface" => "いる",
"yomi" => "イル"
}
]
},
%{
"chunk" => %{
"func" => 8,
"head" => 7,
"id" => 3,
"link" => 5,
"rel" => "D",
"score" => -0.742128
},
"morphs" => [
%{
"base" => "本",
"conjugation" => "",
"conjugation_form" => "",
"id" => 7,
"pos" => "名詞",
"pos1" => "一般",
"pos2" => "",
"pos3" => "",
"pronunciation" => "ホン",
"surface" => "本",
"yomi" => "ホン"
},
%{
"base" => "を",
"conjugation" => "",
"conjugation_form" => "",
"id" => 8,
"pos" => "助詞",
"pos1" => "格助詞",
"pos2" => "一般",
"pos3" => "",
"pronunciation" => "ヲ",
"surface" => "を",
"yomi" => "ヲ"
}
]
},
%{
"chunk" => %{
"func" => 11,
"head" => 10,
"id" => 4,
"link" => 5,
"rel" => "D",
"score" => -0.742128
},
"morphs" => [
%{
"base" => "次",
"conjugation" => "",
"conjugation_form" => "",
"id" => 9,
"pos" => "名詞",
"pos1" => "一般",
"pos2" => "",
"pos3" => "",
"pronunciation" => "ツギ",
"surface" => "次",
"yomi" => "ツギ"
},
%{
"base" => "郎",
"conjugation" => "",
"conjugation_form" => "",
"id" => 10,
"pos" => "名詞",
"pos1" => "一般",
"pos2" => "",
"pos3" => "",
"pronunciation" => "ロー",
"surface" => "郎",
"yomi" => "ロウ"
},
%{
"base" => "に",
"conjugation" => "",
"conjugation_form" => "",
"id" => 11,
"pos" => "助詞",
"pos1" => "格助詞",
"pos2" => "一般",
"pos3" => "",
"pronunciation" => "ニ",
"surface" => "に",
"yomi" => "ニ"
}
]
},
%{
"chunk" => %{
"func" => 13,
"head" => 12,
"id" => 5,
"link" => -1,
"rel" => "D",
"score" => 0.0
},
"morphs" => [
%{
"base" => "渡す",
"conjugation" => "連用形",
"conjugation_form" => "五段・サ行",
"id" => 12,
"pos" => "動詞",
"pos1" => "自立",
"pos2" => "",
"pos3" => "",
"pronunciation" => "ワタシ",
"surface" => "渡し",
"yomi" => "ワタシ"
},
%{
"base" => "た",
"conjugation" => "基本形",
"conjugation_form" => "特殊・タ",
"id" => 13,
"pos" => "助動詞",
"pos1" => "",
"pos2" => "",
"pos3" => "",
"pronunciation" => "タ",
"surface" => "た",
"yomi" => "タ"
}
]
}
]
]
```
"""
@spec parse(String.t, boolean()) :: [[Map.t, ...], ...]
def parse(str, read_from_file \\ false) do
command = case read_from_file do
true ->
"cabocha -f3 #{str}"
_ ->
"""
cat <<EOS.e42197de978c41367f46aa6429ed5c8e | cabocha -f3
#{str}
EOS.e42197de978c41367f46aa6429ed5c8e
"""
end
result = command
|> String.to_charlist
|> :os.cmd
|> List.to_string
|> String.trim
|> Floki.find("sentence")
result
|> Enum.map(fn {"sentence", [], info} ->
info
|> Enum.map(fn {_, chunk_info, toks} ->
chunk = Enum.reduce(chunk_info, %{}, fn({key, value}, acc) ->
case key do
"score" ->
Map.put(acc, key, String.to_float(value))
"rel" ->
Map.put(acc, key, value)
_ ->
Map.put(acc, key, String.to_integer(value))
end
end)
toks = create_toks(toks)
%{"chunk" => chunk, "morphs" => toks}
end)
end)
end
defp create_toks(toks) do
toks
|> Enum.map(fn {_, [{_, id}, {_, feature}], [surface]} ->
%{"id" => String.to_integer(id)}
|> Map.merge(%{"surface" => surface})
|> Map.merge(parse_feature(feature))
end)
end
defp parse_feature(feature) do
Regex.named_captures(~r/
^
(?:
(?<pos>[^,]+),
\*?(?<pos1>[^,]*),
\*?(?<pos2>[^,]*),
\*?(?<pos3>[^,]*),
\*?(?<conjugation_form>[^,]*),
\*?(?<conjugation>[^,]*),
(?<base>[^,]*)
(?:
,(?<yomi>[^,]*)
,(?<pronunciation>[^,]*)
)?
)?
$
/x, feature)
end
end