def __init__(self, db_dir):
self.db_dir = db_dir # データベースの配置ディレクトリパス
self.db_word = depot.open(os.path.join(db_dir, "Word.qdbm"), "c") # 単語 -> 単語id
self.db_worddoc_tf = depot.open(os.path.join(db_dir, "WordDoc_TF.qdbm"), "c") # 単語id,文書id -> TF
self.db_word_doctf = depot.open(os.path.join(db_dir, "Word_DocTF.qdbm"), "c") # 単語id -> 文書id:TF,...
self.db_doc_wordtf = depot.open(os.path.join(db_dir, "Doc_WordTF.qdbm"), "c") # 文書id -> 単語id:TF,...
self.db_word_stat = depot.open(os.path.join(db_dir, "WordStat.qdbm"), "c") # 単語id -> 単語,DF, sum(TF), SF # csv形式
self.db_wordpair_stat = depot.open(os.path.join(db_dir, "WordPairStat.qdbm"), "c") # 単語1id,単語2id -> DF, SF
#!/usr/bin/python
# coding:utf-8
"""
2013/09/11
QDBMに文書コーパスの単語の統計情報を記録する
"""
from qdbm import depot
import os
db_word = depot.open("Word.qdbm", "n") # 単語 -> 単語id
# db_tf = depot.open("TermFreq.db", "n") # 単語id,文書id -> TF
db_word_stat = depot.open("WordStat.qdbm", "n") # 単語id -> 単語, DF, sum(TF), SF
db_pair_stat = depot.open("WordPairStat.qdbm", "n") # 単語1id, 単語2id -> DF, SF 単語1id < 単語2id
def index(doc):
"""
doc : [sent,...]
sent : [word,...]
"""
word_set = set(word for sent in doc for word in sent)
numWordRec = len(db_word)
wordid_dict = dict()
for word in word_set:
wordid = db_word.setdefault(word, numWordRec)
if wordid == numWordRec:
numWordRec += 1
wordid_dict[word] = wordid
from qdbm import depot
db = depot.open("test.db", "n") # depot.open(ファイル名, フラグ, バケット配列の要素数の目安)
db["apple"] = "red" # key, valueで登録
db["lemon"] = "black"
db["orange"] = "orange"
db["lemon"] = "yellow" # 上書き登録
print db["lemon"] # keyがlemonのvalueを取得
print db.get("orange", "unknown") # default付きで取得
print db.get("melon", "unknown") # default付きで取得
print db.keys() # keyの一覧を取得
for k in db.iterkeys(): # keyのiteratorでぐるぐるまわす
print k
for k, v in db.iteritems(): # (key, value)のiteratorでぐるぐるまわす
print k, v
for v in db.itervalues(): # valueのiteratorでぐるぐるまわす
print v
db.close() # 行儀よく終了
'''
depot.open(ファイル名, [フラグ, [バケット配列の要素数の目安]])
