Changeset 14277 for lang/python

Timestamp:

06/20/08 07:06:24 (5 months ago)

Author:

ayu

Message:

• added non-blocking search function
  

Location:

lang/python/incsearch

Files:

• include/DocDb.h (modified) (1 diff)
• incsearch/indexer.py (modified) (3 diffs)
• src/DocDb.cpp (modified) (1 diff)

lang/python/incsearch/include/DocDb.h

@@ -24 +24 @@
   DocDbReader(const std::string& dbname);
   ~DocDbReader();
+  //検索関数
   void find(const std::string& searchstr, std::vector<std::string>& answer);
+  //検索関数（排他しない版）
+  void find_dangerous(const std::string& searchstr, std::vector<std::string>& answer);
 };
 

lang/python/incsearch/incsearch/indexer.py

@@ -3 +3 @@
 # Copyright (C)  2007 Ayukawa Hiroshi
 import os.path
+import re
+import glob
+from  incsearch.db import DocDbWriter
 
-def createindex(globber, extractor, outdirbase):
+FUNC = re.compile(r'<tt id[^>]*>([^<]+)</tt>')
+def extractor(cont):
+    scanner = FUNC.scanner(cont)
+    m = scanner.search()
+    while m:
+        yield dict(pos=m.start(), title=m.group(1), word=m.group(1))
+        m = scanner.search()
+
+def globber(pat, base=""):
+    for f in  glob.glob(pat):
+        if f.startswith(base):
+            yield f[len(base):]
+        else:
+            yield f
+
+def createindex(dbfile, globber, outdirbase, extractor=extractor):
+    u"""
+    extractor  ファイル内の情報列を返す。dict(pos=出現文字位置, title=タイトル)
+    """
+    db = DocDbWriter(dbfile)
     for fname in globber:
         #内容を読み込む
+        print "fname", fname
         fp = file(fname)
         whole = fp.read()
@@ -20 +43 @@
         tagged = []
         pre = 0
-        for i, ipoint in enumerate(extractor(whole)):  #TODO:  ここで後でindex構築のためのデータ保存
+        for i, ipoint in enumerate(extractor(whole)):
+            ipoint.update({"fname": fname})
+            db.put(ipoint["word"], "%s\t%d\t%s" % (fname, ipoint["pos"], ipoint["title"]))
             nametag = "<a name='incsearch_%d'/>" % i
-            tagged.append(whole[pre:i])
+            tagged.append(whole[pre:ipoint["pos"]])
             tagged.append(nametag)
-            pre = i
+            pre = ipoint["pos"]
         tagged.append(whole[pre:])
         gp = file(os.path.join(outdir, os.path.basename(fname)), "w")
@@ -32 +57 @@
 
 def test():
-    pass
+    createindex("testdb", globber("data/lib/*"), "data/index")
 
 def main():

lang/python/incsearch/src/DocDb.cpp

@@ -57 +57 @@
     }
   }
+  void DocDbReader::find_dangerous(const std::string& searchstr, std::vector<std::string>& answer){
+    //一文字だと多いので何もしません。
+    if(searchstr.size() < 2) return;
+
+    try{ 
+      //排他なし！
+
+      // カーソルを候補の先頭に置く
+      villa.curjump(searchstr.c_str(), searchstr.size()+1);
+      // カーソルを走査する
+      for(;;){
+        char* key = villa.curkey();
+        if(strstr(key, searchstr.c_str()) != key){
+          free(key);
+          break;
+        }
+        char* val = villa.curval();
+        std::string a = key;
+        a.append("\t");
+        a.append(val);
+        answer.push_back(a);
+        free(key);
+        free(val);
+        villa.curnext();
+      }
+    }catch(qdbm::Villa_error& e){
+      if(e.code() != qdbm::Villa::ENOITEM) throw e;
+    }
+  }
 
 }