Initial version of whoosh based search

realms/config/__init__.py

@@ -86,20 +86,23 @@ DB_URI = 'sqlite:////tmp/wiki.db'
 CACHE_TYPE = 'simple'
 
 # Redis
-#CACHE_TYPE = 'redis'
+# CACHE_TYPE = 'redis'
 CACHE_REDIS_HOST = '127.0.0.1'
 CACHE_REDIS_PORT = 6379
 CACHE_REDIS_DB = '0'
 
 # Memcached
-#CACHE_TYPE = 'memcached'
+# CACHE_TYPE = 'memcached'
 CACHE_MEMCACHED_SERVERS = ['127.0.0.1:11211']
 
-SEARCH_TYPE = 'simple'  # simple is not good for large wikis
+# SEARCH_TYPE = 'simple'  # simple is not good for large wikis
 
 # SEARCH_TYPE = 'elasticsearch'
 ELASTICSEARCH_URL = 'http://127.0.0.1:9200'
 
+SEARCH_TYPE = 'whoosh'
+WHOOSH_INDEX = '/tmp/whoosh'
+
 # Get ReCaptcha Keys for your domain here:
 # https://www.google.com/recaptcha/admin#whyrecaptcha
 RECAPTCHA_ENABLE = False

realms/modules/search/models.py

@@ -6,6 +6,10 @@ def simple(app):
     return SimpleSearch()
 
 
+def whoosh(app):
+    return WhooshSearch(app.config['WHOOSH_INDEX'])
+
+
 def elasticsearch(app):
     from flask.ext.elastic import Elastic
     return ElasticSearch(Elastic(app))
@@ -42,6 +46,65 @@ class SimpleSearch(BaseSearch):
         pass
 
 
+class WhooshSearch(BaseSearch):
+    def __init__(self, index_path):
+        from whoosh import index as whoosh_index
+        from whoosh.fields import Schema, TEXT, ID
+        from whoosh import qparser
+        from whoosh.highlight import UppercaseFormatter
+        from whoosh.analysis import LanguageAnalyzer
+        import os.path
+
+        self.schema = Schema(path=ID(unique=True, stored=True), body=TEXT(analyzer=LanguageAnalyzer("de")))
+        self.formatter = UppercaseFormatter()
+
+        if os.path.exists(index_path):
+            self.search_index = whoosh_index.open_dir(index_path)
+        else:
+            os.mkdir(index_path)
+            self.search_index = whoosh_index.create_in(index_path, self.schema)
+
+        self.query_parser = qparser.QueryParser("body", schema=self.schema)
+        self.query_parser.add_plugin(qparser.FuzzyTermPlugin())
+
+    def index(self, index, doc_type, id_=None, body=None):
+        writer = self.search_index.writer()
+        writer.update_document(path=id_.decode("utf-8"), body=body["content"])
+        writer.commit()
+
+    def index_wiki(self, name, body):
+        self.index('wiki', 'page', id_=name, body=body)
+
+    def delete_index(self, index):
+        writer = self.search_index.writer()
+        writer.delete_by_term('path', index)
+        writer.commit()
+
+    def wiki(self, query):
+        if not query:
+            return []
+
+        q = self.query_parser.parse("%s~2" % (query,))
+
+        with self.search_index.searcher() as s:
+            results = s.search(q)
+
+            results.formatter = self.formatter
+
+            res = []
+            for hit in results:
+                name = hit["path"]
+                page_data = g.current_wiki.get_page(name)["data"].decode("utf-8")
+                content = hit.highlights('body', text=page_data)
+
+                res.append(dict(name=name, content=content))
+
+            return res
+
+    def users(self, query):
+        pass
+
+
 class ElasticSearch(BaseSearch):
     def __init__(self, elastic):
         self.elastic = elastic