def dst_srcs(f):
data = mk41.load_cabocha(f)
phrase_dst_ls = list()
sentence_ls = list()
all_ls = list()
pattern = re.compile(r"。|、|\ ")
# [我輩は, 5]とするリスト作成
for sentence in data:
for line in sentence:
phrase_dst_ls.append(pattern.sub("", line.phrase)) # 句読点等削除
phrase_dst_ls.append(line.dst)
sentence_ls.append(phrase_dst_ls)
phrase_dst_ls = list()
all_ls.append(sentence_ls)
sentence_ls = list()
return all_ls
def dst_srcs(f):
data = mk41.load_cabocha(f)
phrase_dst_ls = list()
sentence_ls = list()
all_ls = list()
pattern = re.compile(r"。|、|\ ")
# [我輩は, 5]とするリスト作成
for sentence in data:
for line in sentence:
phrase_dst_ls.append(pattern.sub("", line.phrase)) # 句読点等削除
phrase_dst_ls.append(line.dst)
sentence_ls.append(phrase_dst_ls)
phrase_dst_ls = list()
all_ls.append(sentence_ls)
sentence_ls = list()
return all_ls
（中略）
このプログラムの出力をファイルに保存し，以下の事項をUNIXコマンドを用いて確認せよ．
・コーパス中で頻出する述語（サ変接続名詞+を+動詞）
・コーパス中で頻出する述語と助詞パターン
"""
"""
chunk = [morphs=[[surface, base, pos, pos1], ...], dst, srcs, phrase]
chunk = [[[我輩, 我輩, 名詞, 代名詞], [は, は, 助詞, 助詞]], 2, [0], 我輩は]
"""
import mk41
import sys, re
data = mk41.load_cabocha(sys.stdin)
pattern = re.compile(r"。|、|「|」|\*| ")
for sentence in data:
for chunk in sentence:
verb = ""
dst_ls = list()
kou_ls = list()
for morph in chunk.morphs:
if morph.pos == "動詞":
verb = morph.base
break
if verb != "" and len(chunk.srcs) > 0:
sahen_wo = ""
for src_index in chunk.srcs:
# coding=utf-8
"""
43. 名詞を含む文節が動詞を含む文節に係るものを抽出
名詞を含む文節が，動詞を含む文節に係るとき，これらをタブ区切り形式で抽出せよ．ただし，句読点などの記号は出力しないようにせよ．
"""
"""
chunk = [morphs=[[surface, base, pos, pos1], ...], dst, srcs, phrase]
chunk = [[[我輩, 我輩, 名詞, 代名詞], [は, は, 助詞, 助詞]], 2, [0], 我輩は]
"""
import sys, re
import mk41
data = mk41.load_cabocha(sys.stdin)
pattern = re.compile(r"。|、| ")
for sentence in data:
for chunk in sentence:
for morpheme in chunk.morphs:
if morpheme.pos == "名詞" and chunk.dst != -1:
noun_src = pattern.sub("", chunk.phrase) # 句読点等削除
for is_verb_morphs in sentence[chunk.dst].morphs:
if is_verb_morphs.pos == "動詞":
verb_dst = pattern.sub(
"", sentence[chunk.dst].phrase) # 句読点等削除
print noun_src + "\t" + verb_dst
break
break
"""
$ python mk43.py < neko.txt.cabocha
どこで 生れたか
def main():
data = mk41.load_cabocha(sys.stdin)
pattern = re.compile(r"。|、|「|」|\*| ")
# noun_flag = 0
path_ls = list()
for sentence in data:
for chunk in sentence:
noun_flag = 0
for morph in chunk.morphs:
if morph.pos == "名詞": # 文節chunkに名詞があるか
noun_flag = 1
break
if noun_flag == 1: # 名詞があったら
all_path = list()
all_path.append(chunk) # 係り元のchunkをall_pathリストに追加
# pass_ls.append(chunk.phrase)
dst_path = chunk.dst
while (dst_path != -1): # 係り先がある限り
dst_chunk = sentence[dst_path]
all_path.append(dst_chunk) # 係り先のchunkをall_pathリストに追加
dst_path = dst_chunk.dst
path_ls.append(all_path)
list_of_nodes = [{"chunk": l[0], "path": l} for l in path_ls]
for node_i, node_j in combinations(list_of_nodes, 2):
if node_j["chunk"] in node_i[
"path"]: # node_iから構文木の根までのパスにnode_jが存在する場合
index_j_in_i = node_i["path"].index(node_j["chunk"])
path_ij = node_i["path"][0:index_j_in_i + 1]
path_ij_surfaces = [chunk.phrase for chunk in path_ij]
path_ij_surfaces[0] = replace_noun_phrase_in_chunk(path_ij[0], "X")
path_ij_surfaces[-1] = "Y"
print " -> ".join(path_ij_surfaces)
else: # node_iと文節node_jから構文木の根までのパス上で共通のnode_kで交わる場合
node_k_chunk = None
h = -1
while (node_i["path"][h] == node_j["path"][h]):
node_k_chunk = node_i["path"][h]
h = h - 1
if node_k_chunk != None:
# node_iからnode_kまでのパスを取り出す
index_k_in_i = node_i["path"].index(node_k_chunk)
path_ik = node_i["path"][0:index_k_in_i]
path_ik_surfaces = [
re.sub(pattern, "", chunk.phrase) for chunk in path_ik
]
path_ik_surfaces[0] = replace_noun_phrase_in_chunk(
path_ik[0], "X")
# node_jからnode_kまでのパスを取り出す
index_k_in_j = node_j["path"].index(node_k_chunk)
path_jk = node_j["path"][0:index_k_in_j]
path_jk_surfaces = [
re.sub(pattern, "", chunk.phrase) for chunk in path_jk
]
path_jk_surfaces[0] = replace_noun_phrase_in_chunk(
path_jk[0], "Y")
print " -> ".join(path_ik_surfaces) \
+ " | " + " -> ".join(path_jk_surfaces) \
+ " | " + re.sub(pattern, "", node_k_chunk.phrase)