def _create_query_for_field(self, field, value, analyzer=None):
"""generate a field query
this functions creates a field->value query
@param field: the fieldname to be used
@type field: str
@param value: the wanted value of the field
@type value: str
@param analyzer: the analyzer to be used
possible analyzers are:
- L{CommonDatabase.ANALYZER_TOKENIZE}
the field value is splitted to be matched word-wise
- L{CommonDatabase.ANALYZER_PARTIAL}
the field value must start with the query string
- L{CommonDatabase.ANALYZER_EXACT}
keep special characters and the like
@type analyzer: bool
@return: resulting query object
@rtype: PyLucene.Query
"""
if analyzer is None:
analyzer = self.analyzer
if analyzer == self.ANALYZER_EXACT:
analyzer_obj = self.ExactAnalyzer()
else:
value = _escape_term_value(value)
analyzer_obj = PyLucene.StandardAnalyzer()
if (analyzer & self.ANALYZER_PARTIAL) > 0:
# PyLucene uses explicit wildcards for partial matching
value += "*"
return PyLucene.QueryParser.parse(value, field, analyzer_obj)
def create_index(self, arg):
""" Post download setup callback for creating a lucene index """
moreinfo("Creating lucene index")
storeDir = "index"
if not os.path.exists(storeDir):
os.mkdir(storeDir)
store = PyLucene.FSDirectory.getDirectory(storeDir, True)
self.lucene_writer = PyLucene.IndexWriter(store,
PyLucene.StandardAnalyzer(),
True)
# Uncomment this line to enable a PorterStemmer analyzer
# self.lucene_writer = PyLucene.IndexWriter(store, PorterStemmerAnalyzer(), True)
self.lucene_writer.setMaxFieldLength(1048576)
count = 0
urllist = []
for urlobj in self._urldict.values():
filename = urlobj.get_full_filename()
url = urlobj.get_full_url()
try:
urllist.index(url)
continue
except ValueError:
urllist.append(url)
if not filename in self._downloaddict['_savedfiles']: continue
data = ''
moreinfo('Adding index for URL', url)
if os.path.isfile(filename):
try:
data = unicode(open(filename).read(), 'iso-8859-1')
except UnicodeDecodeError, e:
data = ''
doc = PyLucene.Document()
doc.add(
PyLucene.Field("name", filename, PyLucene.Field.Store.YES,
PyLucene.Field.Index.UN_TOKENIZED))
doc.add(
PyLucene.Field("path", url, PyLucene.Field.Store.YES,
PyLucene.Field.Index.UN_TOKENIZED))
if data and len(data) > 0:
doc.add(
PyLucene.Field("contents", data, PyLucene.Field.Store.YES,
PyLucene.Field.Index.TOKENIZED))
else:
extrainfo("warning: no content in %s" % filename)
self.lucene_writer.addDocument(doc)
count += 1
def _create_query_for_field(self, field, value, analyzer=None):
"""generate a field query
this functions creates a field->value query
:param field: The fieldname to be used
:type field: str
:param value: The wanted value of the field
:type value: str
:param analyzer: The analyzer to be used
Possible analyzers are:
- :attr:`CommonDatabase.ANALYZER_TOKENIZE`
the field value is splitted to be matched word-wise
- :attr:`CommonDatabase.ANALYZER_PARTIAL`
the field value must start with the query string
- :attr:`CommonDatabase.ANALYZER_EXACT`
keep special characters and the like
:type analyzer: bool
:return: resulting query object
:rtype: PyLucene.Query
"""
if analyzer is None:
analyzer = self.analyzer
if analyzer == self.ANALYZER_EXACT:
analyzer_obj = PyLucene.KeywordAnalyzer()
else:
value = self._escape_term_value(value)
analyzer_obj = PyLucene.StandardAnalyzer()
qp = PyLucene.QueryParser(field, analyzer_obj)
if (analyzer & self.ANALYZER_PARTIAL > 0):
# PyLucene uses explicit wildcards for partial matching
value += "*"
return qp.parse(value)
def index_node(self, iba_node):
self.delete_node(iba_node.nid)
create = len(os.listdir('index')) == 0
analyzer = PyLucene.StandardAnalyzer()
writer = PyLucene.IndexWriter("index", analyzer, create)
writer.addDocument(self._document_node(iba_node))
writer.close()
self.count = self.count + 1
def search_node_by_name2(self, name):
if self.searcher is None:
self.searcher = PyLucene.IndexSearcher("index")
query = PyLucene.QueryParser(COLUMN_NAME,
PyLucene.StandardAnalyzer()).parse(name)
hits = self.searcher.search(query)
result = self.hits_to_list(hits)
return result
def __init__(self, store_dir=STORE_DIR, destroy=False, analyzer=None):
self.store_dir = store_dir
self.analyzer = analyzer or lucene.StandardAnalyzer()
self.feed_modifier = IndexModifier(store_dir=os.path.join(
store_dir, 'feeds'),
destroy=destroy,
analyzer=analyzer)
self.entry_modifier = IndexModifier(store_dir=os.path.join(
store_dir, 'entries'),
destroy=destroy,
analyzer=analyzer)
def __init__(self, basedir, analyzer=None, create_allowed=True):
"""Initialize or open an indexing database.
Any derived class must override __init__.
:raise ValueError: The given location exists, but the database type
is incompatible (e.g. created by a different indexing engine)
:raise OSError: the database failed to initialize
:param basedir: The parent directory of the database
:type basedir: str
:param analyzer: Bitwise combination of possible analyzer flags
to be used as the default analyzer for this database.
Leave it empty to use the system default analyzer
(self.ANALYZER_DEFAULT). See self.ANALYZER_TOKENIZE,
self.ANALYZER_PARTIAL, ...
:type analyzer: int
:param create_allowed: create the database, if necessary; default: True
:type create_allowed: bool
"""
jvm = PyLucene.getVMEnv()
jvm.attachCurrentThread()
super(PyLuceneDatabase, self).__init__(basedir,
analyzer=analyzer,
create_allowed=create_allowed)
self.pyl_analyzer = PyLucene.StandardAnalyzer()
self.writer = None
self.reader = None
self.index_version = None
try:
# try to open an existing database
tempreader = PyLucene.IndexReader.open(self.location)
tempreader.close()
except PyLucene.JavaError, err_msg:
# Write an error out, in case this is a real problem instead of an absence of an index
# TODO: turn the following two lines into debug output
#errorstr = str(e).strip() + "\n" + self.errorhandler.traceback_str()
#DEBUG_FOO("could not open index, so going to create: " + errorstr)
# Create the index, so we can open cached readers on it
if not create_allowed:
raise OSError("Indexer: skipping database creation")
try:
# create the parent directory if it does not exist
parent_path = os.path.dirname(self.location)
if not os.path.isdir(parent_path):
# recursively create all directories up to parent_path
os.makedirs(parent_path)
except IOError, err_msg:
raise OSError("Indexer: failed to create the parent " \
+ "directory (%s) of the indexing database: %s" \
% (parent_path, err_msg))
def _create_query_for_string(self, text, require_all=True,
analyzer=None):
"""generate a query for a plain term of a string query
basically this function parses the string and returns the resulting
query
:param text: The query string
:type text: str
:param require_all: boolean operator
(True -> AND (default) / False -> OR)
:type require_all: bool
:param analyzer: The analyzer to be used
Possible analyzers are:
- :attr:`CommonDatabase.ANALYZER_TOKENIZE`
the field value is splitted to be matched word-wise
- :attr:`CommonDatabase.ANALYZER_PARTIAL`
the field value must start with the query string
- :attr:`CommonDatabase.ANALYZER_EXACT`
keep special characters and the like
:type analyzer: bool
:return: resulting query object
:rtype: PyLucene.Query
"""
if analyzer is None:
analyzer = self.analyzer
if analyzer == self.ANALYZER_EXACT:
# exact matching - no substitution ...
# for PyLucene: nothing special is necessary
pass
# don't care about special characters ...
if analyzer == self.ANALYZER_EXACT:
analyzer_obj = self.ExactAnalyzer()
else:
text = _escape_term_value(text)
analyzer_obj = PyLucene.StandardAnalyzer()
qp = PyLucene.QueryParser(analyzer=analyzer_obj)
if require_all:
qp.setDefaultOperator(qp.Operator.AND)
else:
qp.setDefaultOperator(qp.Operator.OR)
if (analyzer & self.ANALYZER_PARTIAL) > 0:
# PyLucene uses explicit wildcards for partial matching
text += "*"
return qp.parse(text)
def __init__(self, store_dir=STORE_DIR, destroy=False, analyzer=None):
self.store_dir = store_dir
self.create = False
if os.path.exists(self.store_dir) and destroy:
shutil.rmtree(self.store_dir)
if not os.path.exists(self.store_dir):
try:
os.makedirs(self.store_dir)
except:
{}
self.create = True
self.store = lucene.FSDirectory.getDirectory(self.store_dir,
self.create)
self.analyzer = analyzer or lucene.StandardAnalyzer()
if self.create:
self.get_writer(self.create).close() #this inits the segment
def search(self,
query_string='',
require_visible=True,
allow_curated=True):
hits = []
query_string = str(query_string)
self.logger.info('Performing search: %s' % query_string)
disassembled_query = disassemble_user_query(query_string)
self.logger.debug('Disassembled query: %s' % str(disassembled_query))
reassembled_query = '+(%s)' % reassemble_user_query(disassembled_query)
self.logger.debug('Reassembled query: %s', reassembled_query)
if not allow_curated:
reassembled_query += \
' -record-status:%s' % canary.loader.QueuedRecord.STATUS_CURATED
if require_visible:
reassembled_query += ' +article-type:[%s TO %s]' % \
(Study.ARTICLE_TYPES['traditional'],
Study.ARTICLE_TYPES['curated'])
reassembled_query += ' +record-status:%s' % \
canary.loader.QueuedRecord.STATUS_CURATED
try:
searcher = PyLucene.IndexSearcher(
PyLucene.FSDirectory.getDirectory(
self.context.config.search_index_dir, False))
analyzer = PyLucene.StandardAnalyzer()
query_parser = PyLucene.QueryParser('all', analyzer)
query_parser.setOperator(PyLucene.QueryParser.DEFAULT_OPERATOR_AND)
query = query_parser.parseQuery(reassembled_query)
self.logger.info('Search query: %s', query)
hits = searcher.search(query)
return hits, searcher
except Exception, e:
self.logger.error('Search failed: %s', e)
#self.logger.error(traceback.format_stack())
if hits \
and searcher:
return hits, searcher
else:
return [], None
def index_files(board, time_delta):
store = PyLucene.FSDirectory.getDirectory(
BOARDSPATH + board + '/' + RECENT_INDEX, True)
writer = PyLucene.IndexWriter(store, PyLucene.StandardAnalyzer(), True)
writer.setMaxFieldLength(1048576) # 1MB
flist = get_all_files(board, time_delta)
for filename, owner, title in flist:
path = BOARDSPATH + board + '/' + filename
if not os.path.exists(path):
continue
f = open(path, 'r')
contents = filter_file(f)
debug(contents)
try:
title = title.decode('gbk')
owner = owner.decode('gbk')
contents = unicode(contents, 'gbk')
except UnicodeDecodeError:
f.close()
debug(filename)
continue
f.close()
if len(contents) > 0:
doc = PyLucene.Document()
doc.add(
PyLucene.Field("name", filename, PyLucene.Field.Store.YES,
PyLucene.Field.Index.UN_TOKENIZED))
doc.add(
PyLucene.Field("owner", owner, PyLucene.Field.Store.YES,
PyLucene.Field.Index.UN_TOKENIZED))
doc.add(
PyLucene.Field("title", title, PyLucene.Field.Store.YES,
PyLucene.Field.Index.UN_TOKENIZED))
doc.add(
PyLucene.Field("contents", contents, PyLucene.Field.Store.YES,
PyLucene.Field.Index.TOKENIZED))
writer.addDocument(doc)
debug('adding ' + filename)
writer.optimize()
writer.close()
def search_node_by_attribute2(self, att_type, att_value):
if self.searcher is None:
self.searcher = PyLucene.IndexSearcher("index")
analyzer = PyLucene.StandardAnalyzer()
if att_type != "" and att_value == "":
parser = PyLucene.QueryParser(COLUMN_ATTRIBUTE_TYPE_NID, analyzer)
query = parser.parse(att_type)
elif att_type == "" and att_value != "":
parser = PyLucene.QueryParser(COLUMN_ATTRIBUTE_VALUE, analyzer)
query = parser.parse(att_value)
elif att_type != "" and att_value != "":
parser = PyLucene.QueryParser(COLUMN_ATTRIBUTE_VALUE, analyzer)
query = parser.parse(COLUMN_ATTRIBUTE_TYPE_NID + ":" + att_type +
" AND " + att_value)
hits = self.searcher.search(query)
result = self.hits_to_list(hits)
return result
def _create_query_for_string(self, text, require_all=True,
analyzer=None):
"""generate a query for a plain term of a string query
basically this function parses the string and returns the resulting
query
@param text: the query string
@type text: str
@param require_all: boolean operator
(True -> AND (default) / False -> OR)
@type require_all: bool
@param analyzer: the analyzer to be used
possible analyzers are:
- L{CommonDatabase.ANALYZER_TOKENIZE}
the field value is splitted to be matched word-wise
- L{CommonDatabase.ANALYZER_PARTIAL}
the field value must start with the query string
- L{CommonDatabase.ANALYZER_EXACT}
keep special characters and the like
@type analyzer: bool
@return: resulting query object
@rtype: PyLucene.Query
"""
if analyzer is None:
analyzer = self.analyzer
if analyzer == self.ANALYZER_EXACT:
analyzer_obj = PyLucene.KeywordAnalyzer()
else:
text = _escape_term_value(text)
analyzer_obj = PyLucene.StandardAnalyzer()
qp = PyLucene.QueryParser(UNNAMED_FIELD_NAME, analyzer_obj)
if (analyzer & self.ANALYZER_PARTIAL > 0):
# PyLucene uses explicit wildcards for partial matching
text += "*"
if require_all:
qp.setDefaultOperator(qp.Operator.AND)
else:
qp.setDefaultOperator(qp.Operator.OR)
return qp.parse(text)
PyLucene.Field("name", filename,
PyLucene.Field.Store.YES,
PyLucene.Field.Index.UN_TOKENIZED))
doc.add(
PyLucene.Field("path", path, PyLucene.Field.Store.YES,
PyLucene.Field.Index.UN_TOKENIZED))
if len(contents) > 0:
doc.add(
PyLucene.Field("contents", contents,
PyLucene.Field.Store.YES,
PyLucene.Field.Index.TOKENIZED))
else:
print "warning: no content in %s" % filename
writer.addDocument(doc)
except Exception, e:
print "Failed in indexDocs:", e
if __name__ == '__main__':
if len(sys.argv) < 2:
print IndexFiles.__doc__
sys.exit(1)
print 'PyLucene', PyLucene.VERSION, 'Lucene', PyLucene.LUCENE_VERSION
start = datetime.now()
try:
IndexFiles(sys.argv[1], "index", PyLucene.StandardAnalyzer())
end = datetime.now()
print end - start
except Exception, e:
print "Failed: ", e
def create_index(self):
""" Post download setup callback for creating a lucene index """
info("Creating lucene index")
count = 0
urllist = []
urldb = objects.datamgr.get_urldb()
storeDir = "index"
if not os.path.exists(storeDir):
os.mkdir(storeDir)
store = PyLucene.FSDirectory.getDirectory(storeDir, True)
lucene_writer = PyLucene.IndexWriter(store,
PyLucene.StandardAnalyzer(), True)
lucene_writer.setMaxFieldLength(1048576)
for node in urldb.preorder():
urlobj = node.get()
# Only index if web-page or document
if not urlobj.is_webpage() and not urlobj.is_document(): continue
filename = urlobj.get_full_filename()
url = urlobj.get_full_url()
try:
urllist.index(urlobj.index)
continue
except ValueError:
urllist.append(urlobj.index)
if not os.path.isfile(filename): continue
data = ''
extrainfo('Adding index for URL', url)
try:
data = unicode(open(filename).read(), 'iso-8859-1')
except UnicodeDecodeError, e:
data = ''
try:
doc = PyLucene.Document()
doc.add(
PyLucene.Field("name", 'file://' + filename,
PyLucene.Field.Store.YES,
PyLucene.Field.Index.UN_TOKENIZED))
doc.add(
PyLucene.Field("path", url, PyLucene.Field.Store.YES,
PyLucene.Field.Index.UN_TOKENIZED))
if data and len(data) > 0:
doc.add(
PyLucene.Field("contents", data,
PyLucene.Field.Store.YES,
PyLucene.Field.Index.TOKENIZED))
else:
warning("warning: no content in %s" % filename)
lucene_writer.addDocument(doc)
except PyLucene.JavaError, e:
print e
continue
def _analyzer(self):
return PyLucene.StandardAnalyzer([])
def __init__(self, basedir, analyzer=None, create_allowed=True):
"""Initialize or open an indexing database.
Any derived class must override __init__.
:raise ValueError: The given location exists, but the database type
is incompatible (e.g. created by a different indexing engine)
:raise OSError: the database failed to initialize
:param basedir: The parent directory of the database
:type basedir: str
:param analyzer: Bitwise combination of possible analyzer flags
to be used as the default analyzer for this database.
Leave it empty to use the system default analyzer
(self.ANALYZER_DEFAULT). See self.ANALYZER_TOKENIZE,
self.ANALYZER_PARTIAL, ...
:type analyzer: int
:param create_allowed: create the database, if necessary; default: True
:type create_allowed: bool
"""
jvm = PyLucene.getVMEnv()
jvm.attachCurrentThread()
super(PyLuceneDatabase, self).__init__(basedir,
analyzer=analyzer,
create_allowed=create_allowed)
self.pyl_analyzer = PyLucene.StandardAnalyzer()
self.writer = None
self.reader = None
self.index_version = None
try:
# try to open an existing database
tempreader = PyLucene.IndexReader.open(self.location)
tempreader.close()
except PyLucene.JavaError as err_msg:
# Write an error out, in case this is a real problem instead of an absence of an index
# TODO: turn the following two lines into debug output
#errorstr = str(e).strip() + "\n" + self.errorhandler.traceback_str()
#DEBUG_FOO("could not open index, so going to create: " + errorstr)
# Create the index, so we can open cached readers on it
if not create_allowed:
raise OSError("Indexer: skipping database creation")
try:
# create the parent directory if it does not exist
parent_path = os.path.dirname(self.location)
if not os.path.isdir(parent_path):
# recursively create all directories up to parent_path
os.makedirs(parent_path)
except IOError as err_msg:
raise OSError("Indexer: failed to create the parent "
"directory (%s) of the indexing database: %s" %
(parent_path, err_msg))
try:
tempwriter = PyLucene.IndexWriter(self.location,
self.pyl_analyzer, True)
tempwriter.close()
except PyLucene.JavaError as err_msg:
raise OSError("Indexer: failed to open or create a Lucene"
" database (%s): %s" % (self.location, err_msg))
# the indexer is initialized - now we prepare the searcher
# windows file locking seems inconsistent, so we try 10 times
numtries = 0
#self.dir_lock.acquire(blocking=True)
# read "self.reader", "self.indexVersion" and "self.searcher"
try:
while numtries < 10:
try:
self.reader = PyLucene.IndexReader.open(self.location)
self.indexVersion = self.reader.getCurrentVersion(
self.location)
self.searcher = PyLucene.IndexSearcher(self.reader)
break
except PyLucene.JavaError as e:
# store error message for possible later re-raise (below)
lock_error_msg = e
time.sleep(0.01)
numtries += 1
else:
# locking failed for 10 times
raise OSError("Indexer: failed to lock index database"
" (%s)" % lock_error_msg)
finally:
pass
# self.dir_lock.release()
# initialize the searcher and the reader
self._index_refresh()
def get_search_index_writer (self, clobber=False):
search_index_store = PyLucene.FSDirectory.getDirectory(
self.config.search_index_dir, clobber)
return PyLucene.IndexWriter(search_index_store,
PyLucene.StandardAnalyzer(), clobber)
