def indexsents(self, sentences, writer):
t1 = FieldType()
t1.setStored(True)
t1.setTokenized(False)
t1.setIndexOptions(IndexOptions.DOCS_AND_FREQS)
t2 = FieldType()
t2.setStored(False)
t2.setTokenized(True)
t2.setIndexOptions(IndexOptions.DOCS_AND_FREQS_AND_POSITIONS)
for i, sent in enumerate(sentences):
# print 'adding',i, sent
try:
root = os.getcwd()
# contents = unicode(sent, 'iso-8859-1')
doc = Document()
doc.add(Field('name', str(i), t1))
doc.add(Field('path', root, t1))
if len(sent) > 0:
doc.add(Field('contents', sent.lower(), t2))
else:
print('warning: no content in %s' % str(i))
writer.addDocument(doc)
except Exception as e:
print('Failed in indexsents:', e)
writer.commit()
writer.close()
def indexDocs(self, root, writer):
t1 = FieldType() # for short items, e.g. file name.
t1.setIndexed(True)
t1.setStored(True)
t1.setTokenized(False)
t1.setIndexOptions(FieldInfo.IndexOptions.DOCS_AND_FREQS) # DOCS_AND_FREQS_AND_POSITIONS_OFFSETS
t2 = FieldType() # for content
t2.setIndexed(True)
t2.setStored(False) # don't store the original text
t2.setTokenized(True)
t2.setIndexOptions(FieldInfo.IndexOptions.DOCS_AND_FREQS_AND_POSITIONS)
for root, dirnames, filenames in os.walk(root):
for filename in filenames:
print "adding", filename
try:
path = os.path.join(root, filename)
file = open(path)
contents = unicode(file.read(), 'iso-8859-1')
file.close()
doc = Document()
doc.add(Field("name", filename, t1))
doc.add(Field("path", root, t1))
if len(contents) > 0:
doc.add(Field("contents", contents, t2))
else:
print "warning: no content in %s" % filename
writer.addDocument(doc)
except Exception, e:
print "Failed in indexDocs:", e
def indexDocs(self, root, writer):
t1 = FieldType()
t1.setIndexed(True)
t1.setStored(True)
t1.setTokenized(True)
t1.setIndexOptions(FieldInfo.IndexOptions.DOCS_AND_FREQS)
t2 = FieldType()
t2.setIndexed(True)
t2.setStored(False)
t2.setTokenized(True)
t2.setIndexOptions(FieldInfo.IndexOptions.DOCS_AND_FREQS_AND_POSITIONS)
for root, dirnames, filenames in os.walk(root):
# traverse through the doc directory
for filename in filenames:
# if not filename.endswith('.cdc'):
# continue
try:
# only add the filename and path for indexing
path = os.path.join(root, filename)
print "adding file : ", path
file = open(path)
contents = unicode(file.read(), 'utf-8')
file.close()
doc = Document()
doc.add(Field("name", filename, t1))
doc.add(Field("path", root, t1))
if len(contents) > 0:
doc.add(Field("contents", contents, t2))
else:
print "warning: no content in ", filename
writer.addDocument(doc)
except Exception, e:
print "failed in indexDocs:", e
def index_image(self, root, writer):
t1 = FieldType()
t1.setStored(True)
t1.setTokenized(False)
t1.setIndexOptions(IndexOptions.DOCS_AND_FREQS)
t2 = FieldType()
t2.setStored(False)
t2.setTokenized(True)
t2.setIndexOptions(IndexOptions.DOCS_AND_FREQS_AND_POSITIONS)
with open(os.path.join(root, "index.txt"), mode="r",
encoding="utf8") as index:
count = 1
for line in index:
print("\r", count, end="", sep="")
try:
image_url, content = line.strip().split()[:2]
except ValueError as e:
print(e)
continue
doc = Document()
doc.add(Field("raw_content", content, t1))
content = " ".join(
word for word in jieba.cut_for_search(content)
if word.strip() and word not in self.stop_words)
doc.add(Field("url", image_url, t1))
doc.add(Field("content", content, t2))
writer.addDocument(doc)
count += 1
print("\n{count} image(s) added.".format(count=count))
def Indexing(self, writer):
print("Indexing Segmented File [", SEGMENTATION_FILE, "]")
with open(SEGMENTATION_FILE, 'r') as f:
line_count = 0
for line in f:
# 建立 context 的 fieldtype，需要搭建索引、存储、向量化
fieldtype_context = FieldType()
fieldtype_context.setIndexOptions(IndexOptions.DOCS_AND_FREQS)
fieldtype_context.setStored(True)
fieldtype_context.setTokenized(True)
# 建立 phrase 的 fieldtype，只需要保存
fieldtype_phrase = FieldType()
fieldtype_phrase.setStored(True)
# 对分词好的内容进行处理，把词语和词性分开来存储
processed_context, processed_phrase = self.process_line(line)
doc = Document()
# context field是用于记录文章的内容
doc.add(Field('context', processed_context, fieldtype_context))
# phrase field适用于记录文章每个词所对应的词性
doc.add(Field('phrase', processed_phrase, fieldtype_phrase))
# 把document写入索引库
writer.addDocument(doc)
# 跟踪程序运行情况用
print("\r", str(line_count), " lines", end="", flush=True)
line_count = line_count + 1
if line_count > self.index_limit and not self.training:
break
writer.close()
print()
def indexDocs(self, root, writer):
t1 = FieldType()
t1.setIndexed(True)
t1.setStored(True)
t1.setTokenized(False)
t1.setIndexOptions(FieldInfo.IndexOptions.DOCS_AND_FREQS)
t2 = FieldType()
t2.setIndexed(True)
t2.setStored(False)
t2.setTokenized(True)
t2.setIndexOptions(FieldInfo.IndexOptions.DOCS_AND_FREQS_AND_POSITIONS)
for root, dirnames, filenames in os.walk(root):
for filename in filenames:
if not filename.endswith('.txt'):
continue
print "adding", filename
try:
path = os.path.join(root, filename)
file = open(path)
contents = unicode(file.read(), 'gbk')
file.close()
doc = Document()
doc.add(Field("name", filename, t1))
doc.add(Field("path", root, t1))
if len(contents) > 0:
doc.add(Field("contents", contents, t2))
else:
print "warning: no content in %s" % filename
writer.addDocument(doc)
except Exception, e:
print "Failed in indexDocs:", e
def index(self, root):
t = FieldType()
t.setIndexed(True)
t.setStored(True)
t.setTokenized(True)
t.setStoreTermVectors(True)
for path, dirs, files in os.walk(root):
for file in files:
filePath = os.path.join(path, file)
fd = open(filePath)
content = unicode(fd.read(), 'iso-8859-1')
fd.close()
doc = Document()
doc.add(Field('name', file, StringField.TYPE_STORED))
parent = os.path.split(path)[1]
doc.add(Field('parent', parent, StringField.TYPE_STORED))
if len(content) > 0:
doc.add(Field('content', content, t))
print 'Indexing %s' % file
self.mWriter.addDocument(doc)
self.mWriter.commit()
self.mWriter.close()
def indexDocs(self, root, writer):
t1 = FieldType()
t1.setStored(True)
t1.setTokenized(False)
t1.setIndexOptions(IndexOptions.DOCS_AND_FREQS)
t2 = FieldType()
t2.setStored(False)
t2.setTokenized(True)
t2.setStoreTermVectors(True)
t2.setStoreTermVectorOffsets(True)
t2.setStoreTermVectorPositions(True)
t2.setIndexOptions(
IndexOptions.DOCS_AND_FREQS_AND_POSITIONS_AND_OFFSETS)
file_path = root + 'r52-train-all-terms.txt'
fd = open(file_path)
contents = fd.readlines()
fd.close()
contents_list = [x.strip() for x in contents]
for i in xrange(len(contents_list)):
try:
[topic, content] = contents_list[i].split('\t')
doc = Document()
doc.add(Field("id", str(i), t1))
doc.add(Field("topic", topic, t1))
doc.add(Field("contents", content, t2))
writer.addDocument(doc)
except Exception, e:
print "Failed in indexDocs:", e
def indexDocs(self, root, writer):
t1 = FieldType()
t1.setIndexed(True)
t1.setStored(True)
t1.setTokenized(False)
t1.setIndexOptions(FieldInfo.IndexOptions.DOCS_AND_FREQS)
t2 = FieldType()
t2.setIndexed(True)
t2.setStored(False)
t2.setTokenized(True)
t2.setIndexOptions(FieldInfo.IndexOptions.DOCS_AND_FREQS_AND_POSITIONS)
index_file = open("index.txt", 'r')
for line in index_file.readlines():
try:
src = line.strip().split('\t')[0]
filename = line.strip().split('\t')[1]
tag = line.strip().split('\t')[2]
path = os.path.join(root, filename)
doc = Document()
doc.add(Field("name", filename, t1))
doc.add(Field("path", root, t1))
doc.add(Field("src", src, t1))
if len(tag) > 0:
doc.add(Field("tag", tag, t2))
else:
print "warning: no tag in %s" % filename
writer.addDocument(doc)
except Exception, e:
print "Failed in indexDocs:", e
def build_index(document_path, dir_path):
lucene.initVM()
index_dir = SimpleFSDirectory(Paths.get(dir_path))
analyzer = StandardAnalyzer()
config = IndexWriterConfig(analyzer)
config.setOpenMode(IndexWriterConfig.OpenMode.CREATE)
index_writer = IndexWriter(index_dir, config)
t1 = FieldType()
t1.setStored(True)
t1.setTokenized(True)
t1.setIndexOptions(IndexOptions.DOCS_AND_FREQS)
t2 = FieldType()
t2.setStored(True)
t2.setTokenized(False)
with open(document_path) as input_file:
for line in input_file:
segs = line.strip().split(" ")
music_path, music_tags = segs[0], segs[1].split(",")
document = Document()
document.add(Field("content", " ".join(music_tags), t1))
document.add(Field("url", music_path, t2))
index_writer.addDocument(document)
index_writer.close()
def indexDocs(self, root, writer):
t1 = FieldType()
t1.setIndexed(True)
t1.setStored(True)
t1.setTokenized(False)
t1.setIndexOptions(FieldInfo.IndexOptions.DOCS_AND_FREQS)
t2 = FieldType()
t2.setIndexed(True)
t2.setStored(False)
t2.setTokenized(True)
t2.setIndexOptions(FieldInfo.IndexOptions.DOCS_AND_FREQS_AND_POSITIONS)
for root, dirnames, filenames in os.walk(root):
for filename in filenames:
if not filename.endswith('.html'):
continue
print "adding", filename
try:
path = os.path.join(root, filename)
file = open(path)
contents = unicode(file.read(), 'iso-8859-1')
file.close()
doc = Document()
doc.add(Field("name", filename, t1))
doc.add(Field("path", root, t1))
if len(contents) > 0:
doc.add(Field("contents", contents, t2))
else:
print "warning: no content in %s" % filename
writer.addDocument(doc)
except Exception, e:
print "Failed in indexDocs:", e
def indexDocs(self, root, writer):
t1 = FieldType()
t1.setIndexed(False)
t1.setStored(True)
t1.setTokenized(False)
t3 = FieldType()
t3.setIndexed(True)
t3.setStored(False)
t3.setTokenized(True)#利用预先设置的analyzer进行分词，这里是根据空格
t3.setIndexOptions(FieldInfo.IndexOptions.DOCS_AND_FREQS_AND_POSITIONS)
total=0
file = open(root,"r")
for line in file.readlines():
try:
imgurl, itemurl, content = line.split('\t')
total+=1
print total
print "adding", content
contents = ' '.join(jieba.cut(content))
doc = Document()
doc.add(Field("imgurl", imgurl, t1))
doc.add(Field("itemurl", itemurl, t1))
doc.add(Field("title", content, t1))
doc.add(Field("contents",contents,t3))
writer.addDocument(doc)
except Exception, e:
print "Failed in indexDocs:", e
def indexsents(self, sentences, writer):
t1 = FieldType()
t1.setStored(True)
t1.setTokenized(False)
t1.setIndexOptions(IndexOptions.DOCS_AND_FREQS)
t2 = FieldType()
t2.setStored(False)
t2.setTokenized(True)
t2.setIndexOptions(IndexOptions.DOCS_AND_FREQS_AND_POSITIONS)
for i, sent in enumerate(sentences):
#print "adding",i, sent
try:
root = os.getcwd()
#contents = unicode(sent, 'iso-8859-1')
doc = Document()
doc.add(Field("name", str(i), t1))
doc.add(Field("path", root, t1))
if len(sent) > 0:
doc.add(Field("contents", sent.lower(), t2))
else:
print "warning: no content in %s" % str(i)
writer.addDocument(doc)
except Exception, e:
print "Failed in indexsents:", e
def indexer(self, root, writer):
t1 = FieldType()
t1.setStored(True)
t1.setTokenized(True)
t1.setIndexOptions(IndexOptions.DOCS_AND_FREQS)
def repalcer(text):
chars = '\\`*_{}[]()>#+-.!$‘'
for c in chars:
if c in text:
text = text.replace(c, ' ')
return text
for root, dirnames, filenames in os.walk(root):
i = 0
for filename in filenames:
i += 1
with open(os.path.join(root, filename)) as f:
for line in f.readlines():
line = line.split(' ', 2)
docname = line[0] + ' ' + line[1]
name = repalcer(line[0])
contents = line[2]
doc = Document()
doc.add(Field('docname', docname, t1))
doc.add(Field('name', name, t1))
doc.add(Field('contents', contents, t1))
writer.addDocument(doc)
print('File %d done indexing' % i)
def indexDocs(self, root, iw):
t1 = FieldType()
t1.setStored(True)
t1.setTokenized(True)
t1.setIndexOptions(IndexOptions.DOCS_AND_FREQS_AND_POSITIONS)
t2 = FieldType()
t2.setStored(True)
t2.setTokenized(True)
t2.setStoreTermVectors(True)
t2.setIndexOptions(IndexOptions.DOCS_AND_FREQS_AND_POSITIONS)
for filename in os.listdir(root):
if not filename.endswith(".txt"):
print("file is not a txt file. we skip it.")
continue
print("adding", filename)
path = os.path.join(root, filename)
self.parseBook(path, t1, t2, iw)
# Prints a set of statistics displaying missing data
# Authorerror = number of authors not found
# Titleerror = number of titles not found
# Documenterror = number of documents where text could not be extracted so entire document was indexed
print("AuthorError: {}".format(self.authorcount))
print("TitleError: {}".format(self.titlecount))
print("DocumentError: {}".format(self.errorcount))
iw.close()
def indexDocs(self, root, writer):
t1 = FieldType()
t1.setStored(True)
t1.setTokenized(True)
t1.setIndexOptions(IndexOptions.DOCS_AND_FREQS)
wikiFile = ZipFile(root, 'r')
files = wikiFile.namelist()
i = 0
for file in files[1:]:
i += 1
wiki = wikiFile.open(file, 'r')
for line in wiki:
for line in codecs.iterdecode(wiki, 'utf8'):
normailized = unicodedata.normalize('NFD',
line).split(' ', 2)
if not normailized[1].isdigit(): continue
docname = normailized[0] + ' ' + normailized[1]
name = re.sub(r'[^a-zA-Z0-9]', ' ', normailized[0])
contents = normailized[2]
doc = Document()
doc.add(Field('docname', docname, t1))
doc.add(Field('name', name, t1))
doc.add(Field('contents', contents, t1))
writer.addDocument(doc)
print('File %d done indexing' % i, file)
def indexDocs(self, img_url, toi, tid):
try:
t1 = FieldType()
t1.setIndexed(True)
t1.setStored(True)
t1.setTokenized(True)
t1.setIndexOptions(
FieldInfo.IndexOptions.DOCS_AND_FREQS_AND_POSITIONS)
print("Adding", img_url)
name = "Pictures/1.jpg"
conn = urllib.urlopen(img_url)
f = open(name, 'wb')
f.write(conn.read())
f.close()
img = cv2.imread(name)
sdf = img_search_color(img)
storeDir = 'Picture_new/' + sdf.strs
if not os.path.exists(storeDir):
os.mkdir(storeDir)
cv2.imwrite(storeDir + '/' + str(toi) + '___' + str(tid) + '.jpg',
img)
'''storeDir2 = 'Picture_user/'+str(tid)
if not os.path.exists(storeDir2):
n = 0
os.mkdir(storeDir2)
else :
n = len(os.listdir(storeDir2))
cv2.imwrite(storeDir2+'/'+str(toi)+ '_'+ str(n) + '.jpg',img)'''
except Exception, e:
print("Failed in indexDocs:", e)
def index_docs(root, writer):
# metadata: name and path
metadata = FieldType()
metadata.setStored(True) # as is value
metadata.setTokenized(False)
metadata.setIndexOptions(IndexOptions.DOCS_AND_FREQS)
# content: abstract and body
content_type = FieldType()
content_type.setStored(True) # to highlight on search results
content_type.setTokenized(True) # tokenize words
content_type.setIndexOptions(IndexOptions.DOCS_AND_FREQS_AND_POSITIONS)
for directory, _, file_names in walk(root):
for file_name in file_names:
name, extension = splitext(file_name)
if extension not in DOC_FORMATS:
continue # skip unsupported formats
file_path = join(directory, file_name)
print ' ', file_path
# Build indexed document
doc = Document()
doc.add(Field('name', file_name, metadata))
doc.add(Field('path', directory, metadata))
# Read file contents
content = process(file_path, 'utf-8', method='pdfminer')
abstract = extract_abstract(content)
doc.add(Field('content', content, content_type))
doc.add(Field('abstract', abstract, content_type))
writer.addDocument(doc)
def index_docs(self, tweets, writer):
t1 = FieldType()
t1.setIndexed(True)
t1.setStored(True)
t1.setTokenized(True)
t1.setIndexOptions(FieldInfo.IndexOptions.DOCS_AND_FREQS_AND_POSITIONS)
t1.setStoreTermVectors(True)
t1.setStoreTermVectorOffsets(True)
# add each tweet to the index
for tweet in tweets:
try:
# strip out URLs because they provide false index matches
contents = []
for word in tweet[1].text.split():
if word.startswith("http://") or word.startswith("https://"):
continue
contents.append(word)
contents = " ".join(contents)
if len(contents) == 0: continue
doc = Document()
doc.add(Field("contents", contents, t1))
writer.addDocument(doc)
except Exception, e:
print "Failed in index_docs:", e
def indexDocs(self, root, writer):
t1 = FieldType()
t1.setStored(True)
t1.setTokenized(False)
t1.setIndexOptions(IndexOptions.DOCS_AND_FREQS)
t2 = FieldType()
t2.setStored(True)
t2.setTokenized(True)
t2.setIndexOptions(IndexOptions.DOCS_AND_FREQS_AND_POSITIONS)
inFile = open(str(args["inputFile"]))
indexName = inFile.readline()
while (indexName != ''):
print "adding", indexName
doc = Document()
doc.add(Field("name", indexName, t1))
#doc.add(Field("path", root, t1))
text = inFile.readline()
if (len(text) > 0):
print("contents: %s\n" % text)
doc.add(Field("contents", text, t2))
else:
print "warning: no content in %s" % indexName
indexName = inFile.readline()
writer.addDocument(doc)
inFile.close()
def indexDocs(self, sourceDir, writer):
t1 = FieldType()
t1.setStored(True)
t1.setTokenized(False)
t1.setIndexOptions(IndexOptions.DOCS_AND_FREQS)
t2 = FieldType()
t2.setStored(False)
t2.setTokenized(True)
t2.setIndexOptions(IndexOptions.DOCS_AND_FREQS_AND_POSITIONS)
for sourceDir, dirnames, filenames in os.walk(sourceDir):
for filename in filenames:
if not filename.endswith('.txt'):
continue
print(filename)
try:
path = os.path.join(sourceDir, filename)
file = open(path, 'r', encoding="utf-8")
contents = file.read()
#contents = str(filecontent, 'utf-8')
#contents = filecontent.encode('utf-8')
#print('path', path, len(contents))
doc = Document()
doc.add(Field("name", filename, t1)) # filename (title)
#doc.add(Field("path", root, t1))
if len(contents) > 0:
doc.add(Field(queryField, contents, t2)) # content
else:
print("warning: no content in %s" % filename)
writer.addDocument(doc)
file.close()
except NameError:
print("Failed in indexDocs:")
def indexDocs(self, url, writer):
type1 = FieldType()
type1.setIndexed(True)
type1.setStored(True)
type1.setTokenized(False)
type1.setIndexOptions(FieldInfo.IndexOptions.DOCS_AND_FREQS)
type2 = FieldType()
type2.setIndexed(True)
type2.setStored(True)
type2.setTokenized(True)
type2.setIndexOptions(FieldInfo.IndexOptions.DOCS_AND_FREQS_AND_POSITIONS_AND_OFFSETS)
# Read Feeds
feeds = feedparser.parse(url)
for item in feeds["entries"]:
print "adding", item["title"]
try:
link = item["link"]
contents = item["description"].encode("utf-8")
contents = re.sub('<[^<]+?>', '', ''.join(contents))
title = item["title"]
doc = Document()
doc.add(Field("url", link, type1))
doc.add(Field("title", title, type1))
if len(contents) > 0:
doc.add(Field("contents", contents, type2))
else:
print "warning: no content in %s" % item["title"]
writer.addDocument(doc)
except Exception, e:
print "Failed in indexDocs:", e
def build_index(self, dict_data):
print("loading data...")
t1 = FieldType()
t1.setStored(True)
t1.setTokenized(False)
t1.setIndexOptions(IndexOptions.DOCS_AND_FREQS)
t2 = FieldType()
t2.setStored(True)
t2.setTokenized(True)
t2.setIndexOptions(IndexOptions.DOCS_AND_FREQS_AND_POSITIONS)
for k, v in dict_data.items():
doc = Document()
doc.add(Field("id", k, t1))
doc.add(Field("content", v, t2))
self.writer.addDocument(doc)
ticker = Ticker()
print("commit index")
threading.Thread(target=ticker.run).start()
self.writer.commit()
self.writer.close()
ticker.tick = False
print("done")
def index_docs(self, train_set, writer):
t1 = FieldType()
t1.setIndexed(True)
t1.setStored(True)
t1.setTokenized(False)
t1.setIndexOptions(FieldInfo.IndexOptions.DOCS_AND_FREQS)
t2 = FieldType()
t2.setIndexed(True)
t2.setStored(True)
t2.setTokenized(True)
t2.setStoreTermVectorOffsets(True)
t2.setStoreTermVectorPayloads(True)
t2.setStoreTermVectorPositions(True)
t2.setStoreTermVectors(True)
t2.setIndexOptions(FieldInfo.IndexOptions.DOCS_AND_FREQS_AND_POSITIONS)
for ii in train_set:
doc = Document()
doc.add(Field("answer", ii['Answer'], t1))
doc.add(Field("qid", ii['Question ID'], t1))
doc.add(Field("category", ii['category'], t1))
doc.add(Field("position", ii['Sentence Position'], t1))
doc.add(Field("question", ii['Question Text'], t2))
doc.add(Field("wiki_plain",
self.wiki_reader.get_text(ii['Answer']), t2))
writer.addDocument(doc)
def _createNoTermsFrequencyFieldType():
f = FieldType()
f.setIndexed(True)
f.setTokenized(True)
f.setOmitNorms(True)
f.setIndexOptions(FieldInfo.IndexOptions.DOCS_ONLY)
f.freeze()
return f
def _createNoTermsFrequencyFieldType():
f = FieldType()
f.setIndexed(True)
f.setTokenized(True)
f.setOmitNorms(True)
f.setIndexOptions(FieldInfo.IndexOptions.DOCS_ONLY)
f.freeze()
return f
def indexDocs(self, root, writer):
# t1 is used for filenames and t2 is used for contents
t1 = FieldType()
t1.setIndexed(True)
t1.setStored(True)
t1.setTokenized(False)
t1.setIndexOptions(FieldInfo.IndexOptions.DOCS_AND_FREQS)
t2 = FieldType()
t2.setIndexed(True)
t2.setStored(False)
t2.setTokenized(True)
t2.setIndexOptions(FieldInfo.IndexOptions.DOCS_AND_FREQS_AND_POSITIONS)
for root, dirnames, filenames in os.walk(root):
for filename in filenames:
# We can index only a certain types of files
if not (filename.endswith('.txt') or filename.endswith('.pdf') or filename.endswith('.xml') or filename.endswith('.doc') or filename.endswith('.odt')):
continue
try:
file_path = os.path.join(root, filename)
outfile_path = file_path
# First convert PDF and DOC files to text
if filename.endswith('.pdf'):
outfile = filename.replace('.pdf', '.txt')
outfile_path = os.path.join(root, outfile)
cmd = 'pdftotext ' + '-layout ' + "'"+ file_path + "'" + ' ' + "'" + outfile_path + "'"
subprocess.check_output(cmd, shell=True)
file_path = outfile_path
elif filename.endswith('.doc'):
outfile = filename.replace('.doc', '.txt')
outfile_path = os.path.join(root, outfile)
cmd = 'antiword ' + file_path + ' >> ' + outfile_path
subprocess.check_output(cmd, shell=True)
file_path = outfile_path
elif filename.endswith('.odt'):
outfile = filename.replace('.odt', '.txt')
outfile_path = os.path.join(root, outfile)
cmd = 'odttotext ' + '-layout ' + "'"+ file_path + "'" + ' ' + "'" + outfile_path + "'"
subprocess.check_output(cmd, shell=True)
file_path = outfile_path
file = open(file_path)
contents = unicode(file.read(), 'iso-8859-1')
file.close()
doc = Document()
doc.add(Field("name", filename, t1))
doc.add(Field("path", root, t1))
if len(contents) > 0:
doc.add(Field("contents", contents, t2))
else:
logging.debug('warning: no content in %s', filename)
writer.addDocument(doc)
except Exception, e:
logging.debug('Failed in indexDocs: %s', e)
def indexDocs(self, root, indextxt, writer):
t1 = FieldType()
t1.setIndexed(True) #t1为需索引, 需保存, 需分词
t1.setStored(True)
t1.setTokenized(True)
t1.setIndexOptions(FieldInfo.IndexOptions.DOCS_AND_FREQS)
t2 = FieldType() #t2为不需索引, 需保存, 需分词
t2.setIndexed(False)
t2.setStored(True)
t2.setTokenized(True)
t2.setIndexOptions(FieldInfo.IndexOptions.DOCS_AND_FREQS_AND_POSITIONS)
for root, dirnames, filenames in os.walk(root):
for filename in filenames:
print "adding", filename
try:
path = os.path.join(root, filename)
file = open(path)
content = unicode(file.read(), 'utf-8')
content = content.encode('utf-8')
list1 = content.split('\n')
file.close()
doc = Document()
# 读取存于文件的内容并保存于doc
url = list1[0]
print('url : ' + url)
doc.add(Field("url", url, t1))
name = list1[1]
print('name : ' + name)
doc.add(Field("name", name, t1))
collectnum = list1[2]
print('collect_num : ' + collectnum)
doc.add(Field("collect_num", collectnum, t2))
img_url = list1[3]
print('img_url : ' + img_url)
doc.add(Field("img_url", img_url, t2))
zhuliao = list1[4]
print(zhuliao)
doc.add(Field("zhuliao", zhuliao, t1))
zuofa = list1[5]
zuofa = '\n'.join(zuofa.split('\t'))
print('zuofa : ' + zuofa)
doc.add(Field("zuofa", zuofa, t2))
writer.addDocument(doc)
except Exception, e:
print "Failed in indexDocs:", e
def indexDocs_playlist(self, writer):
t1 = FieldType()
t1.setIndexed(False)
t1.setStored(True)
t1.setTokenized(False)
t2 = FieldType()
t2.setIndexed(True)
t2.setStored(True)
t2.setTokenized(True)
t2.setIndexOptions(FieldInfo.IndexOptions.DOCS_AND_FREQS)
# playlist
success = 0
fail = 0
noinfo = 0
playlists = open('data/playlist_details2.db', "r")
for line in playlists.readlines():
if len(line) < 20:
noinfo += 1
continue
information = line.split(',')
try:
playID = information[0]
playname = ' '.join(information[1:-8])
author, playImage, tags, songIDs, sharecount, playcount, subscribedcount, commentcount = information[
-8:]
except Exception, e:
fail += 1
print "fail"
continue
playname = ' '.join(jieba.cut(playname))
author = ' '.join(jieba.cut(author))
tags = ' '.join(jieba.cut(tags))
# playname = ' '.join(pynlpir.segment(playname, pos_tagging=False))
# author = ' '.join(pynlpir.segment(author, pos_tagging=False))
tags = tags.replace("|", " ")
doc = Document()
doc.add(Field("ID", playID, t2))
doc.add(Field("name", playname, t2))
doc.add(Field("author", author, t2))
doc.add(Field("image", playImage, t1))
doc.add(Field("tags", tags, t2))
doc.add(Field("songIDs", songIDs, t1))
doc.add(Field("sharecount", sharecount, t1))
doc.add(Field("playcount", playcount, t1))
doc.add(Field("subscribedcount", subscribedcount, t1))
doc.add(Field("commentcount", commentcount, t1))
writer.addDocument(doc)
print "歌单", playname, "成功添加"
success += 1
def indexDocs(self, root, writer):
t1 = FieldType()
t1.setStored(True)
t1.setTokenized(False)
t1.setIndexOptions(IndexOptions.NONE) # Not Indexed
t2 = FieldType()
t2.setStored(False)
t2.setTokenized(True)
t2.setIndexOptions(IndexOptions.DOCS_AND_FREQS_AND_POSITIONS
) # Indexes documents, frequencies and positions.
count = 0
for root, dirnames, filenames in os.walk(root):
for filename in filenames:
if not filename.endswith('.shtml'):
continue
try:
path = os.path.join(root, filename)
with open(path, 'r') as file:
contents = file.read()
soup = BeautifulSoup(contents, features="html.parser")
doc = Document()
doc.add(Field("name", filename, t1))
doc.add(Field("path", path, t1))
url = self.relation[filename]
doc.add(Field("url", url, t1))
if len(contents) > 0:
title = soup.find('title').text #开始处理开头
content = "".join(soup.findAll(text=True))
content = jieba.lcut(content)
content = ' '.join(content)
doc.add(Field("title", title, t1))
doc.add(Field("contents", content, t2))
else:
doc.add(Field("title", "", t1))
doc.add(Field("contents", "", t2))
print("warning: no content in %s" % filename)
writer.addDocument(doc)
except Exception as e:
print("Failed in indexDocs:", e)
count += 1
if (count % 100 == 0):
writer.commit()
print(count)
class LuceneDocumentField(object):
"""Internal handler class for possible field types"""
def __init__(self):
"""Init possible field types"""
# FIELD_ID: stored, indexed, non-tokenized
self.field_id = FieldType()
self.field_id.setIndexed(True)
self.field_id.setStored(True)
self.field_id.setTokenized(False)
# FIELD_ID_TV: stored, indexed, not tokenized, with term vectors (without positions)
# for storing IDs with term vector info
self.field_id_tv = FieldType()
self.field_id_tv.setIndexed(True)
self.field_id_tv.setStored(True)
self.field_id_tv.setTokenized(False)
self.field_id_tv.setStoreTermVectors(True)
# FIELD_TEXT: stored, indexed, tokenized, with positions
self.field_text = FieldType()
self.field_text.setIndexed(True)
self.field_text.setStored(True)
self.field_text.setTokenized(True)
# FIELD_TEXT_TV: stored, indexed, tokenized, with term vectors (without positions)
self.field_text_tv = FieldType()
self.field_text_tv.setIndexed(True)
self.field_text_tv.setStored(True)
self.field_text_tv.setTokenized(True)
self.field_text_tv.setStoreTermVectors(True)
# FIELD_TEXT_TVP: stored, indexed, tokenized, with term vectors and positions
# (but no character offsets)
self.field_text_tvp = FieldType()
self.field_text_tvp.setIndexed(True)
self.field_text_tvp.setStored(True)
self.field_text_tvp.setTokenized(True)
self.field_text_tvp.setStoreTermVectors(True)
self.field_text_tvp.setStoreTermVectorPositions(True)
def get_field(self, type):
"""Get Lucene FieldType object for the corresponding internal FIELDTYPE_ value"""
if type == Lucene.FIELDTYPE_ID:
return self.field_id
elif type == Lucene.FIELDTYPE_ID_TV:
return self.field_id_tv
elif type == Lucene.FIELDTYPE_TEXT:
return self.field_text
elif type == Lucene.FIELDTYPE_TEXT_TV:
return self.field_text_tv
elif type == Lucene.FIELDTYPE_TEXT_TVP:
return self.field_text_tvp
else:
raise Exception("Unknown field type")
class LuceneDocumentField(object):
"""Internal handler class for possible field types"""
def __init__(self):
"""Init possible field types"""
# FIELD_ID: stored, indexed, non-tokenized
self.field_id = FieldType()
self.field_id.setIndexed(True)
self.field_id.setStored(True)
self.field_id.setTokenized(False)
# FIELD_ID_TV: stored, indexed, not tokenized, with term vectors (without positions)
# for storing IDs with term vector info
self.field_id_tv = FieldType()
self.field_id_tv.setIndexed(True)
self.field_id_tv.setStored(True)
self.field_id_tv.setTokenized(False)
self.field_id_tv.setStoreTermVectors(True)
# FIELD_TEXT: stored, indexed, tokenized, with positions
self.field_text = FieldType()
self.field_text.setIndexed(True)
self.field_text.setStored(True)
self.field_text.setTokenized(True)
# FIELD_TEXT_TV: stored, indexed, tokenized, with term vectors (without positions)
self.field_text_tv = FieldType()
self.field_text_tv.setIndexed(True)
self.field_text_tv.setStored(True)
self.field_text_tv.setTokenized(True)
self.field_text_tv.setStoreTermVectors(True)
# FIELD_TEXT_TVP: stored, indexed, tokenized, with term vectors and positions
# (but no character offsets)
self.field_text_tvp = FieldType()
self.field_text_tvp.setIndexed(True)
self.field_text_tvp.setStored(True)
self.field_text_tvp.setTokenized(True)
self.field_text_tvp.setStoreTermVectors(True)
self.field_text_tvp.setStoreTermVectorPositions(True)
def get_field(self, type):
"""Get Lucene FieldType object for the corresponding internal FIELDTYPE_ value"""
if type == Lucene.FIELDTYPE_ID:
return self.field_id
elif type == Lucene.FIELDTYPE_ID_TV:
return self.field_id_tv
elif type == Lucene.FIELDTYPE_TEXT:
return self.field_text
elif type == Lucene.FIELDTYPE_TEXT_TV:
return self.field_text_tv
elif type == Lucene.FIELDTYPE_TEXT_TVP:
return self.field_text_tvp
else:
raise Exception("Unknown field type")
def create_minidoc(termstring, field='text'):
# To store term vectors (used for query expansion) we have to use a custom fieldtype
customfield = FieldType()
customfield.setIndexOptions(IndexOptions.DOCS_AND_FREQS)
customfield.setStored(True)
customfield.setTokenized(True)
customfield.setStoreTermVectors(True)
doc = Document()
doc.add(Field(field, termstring, customfield))
return doc
class Indexer(object):
def __init__(self, **kwargs):
""" Initialize a new instance of the Indexer
:param output: The output directory of the underlying index
:param anaylzer: The overloaded analyzer to work with
"""
self.output = kwargs.get("root", "index")
if not os.path.exists(self.output):
os.mkdir(self.output)
self.analyzer = kwargs.get("analyzer", StandardAnalyzer(Version.LUCENE_CURRENT))
self.analyzer = LimitTokenCountAnalyzer(self.analyzer, 1048576)
self.config = IndexWriterConfig(Version.LUCENE_CURRENT, self.analyzer)
self.config.setOpenMode(IndexWriterConfig.OpenMode.CREATE)
self.store = SimpleFSDirectory(File(self.output))
self.writer = IndexWriter(self.store, self.config)
self.create_field_types()
def index(self, document):
""" Given a new document, add it to the index.
:param document: The document to add to the indexer
"""
try:
self.writer.addDocument(document)
except Exception:
logger.exception("Failed to index the supplied document")
def shutdown(self):
""" Shutdown the currently processing indexer.
"""
try:
# self.writer.optimize()
self.writer.close()
except Exception:
logger.exception("Failed to shutdown the indexer correctly")
def create_field_types(self):
""" Create the field types that will be used to specify
what actions lucene should take on the various fields
supplied to index.
"""
self.field_clean = FieldType()
self.field_clean.setIndexed(True)
self.field_clean.setStored(True)
self.field_clean.setTokenized(False)
self.field_clean.setIndexOptions(FieldInfo.IndexOptions.DOCS_AND_FREQS)
self.field_dirty = FieldType()
self.field_dirty.setIndexed(True)
self.field_dirty.setStored(False)
self.field_dirty.setTokenized(True)
self.field_dirty.setIndexOptions(FieldInfo.IndexOptions.DOCS_AND_FREQS_AND_POSITIONS)
def indexDocs(self, root, writer):
t1 = FieldType()
t1.setIndexed(True)
t1.setStored(True)
t1.setTokenized(False)
t1.setIndexOptions(FieldInfo.IndexOptions.DOCS_AND_FREQS)
t2 = FieldType()
t2.setIndexed(True)
t2.setStored(False)
t2.setTokenized(True)
t2.setIndexOptions(FieldInfo.IndexOptions.DOCS_AND_FREQS_AND_POSITIONS)
files = os.listdir(root)
for filename in files:
print "adding", filename
try:
path = os.path.join(root, filename)
file = open(path)
total = file.read()
totallist=total.split("\n")
geshou=totallist[0]
geming=totallist[1]
zhuanji=totallist[2]
imgurl=totallist[3]
liupai=totallist[4]
shijian=totallist[5]
jianjie=totallist[6]
geci=totallist[7]
contents = geming+geshou+zhuanji+liupai+geci
seg_result = jieba.cut(contents)
contents = ' '.join(seg_result)
doc = Document()
doc.add(Field("contents",contents,t2))
doc.add(Field("geming", geming, t1))
doc.add(Field("geshou", geshou, t1))
doc.add(Field("zhuanji", zhuanji, t1))
doc.add(Field("liupai",liupai,t1))
doc.add(Field("geci",geci,t1))
doc.add(Field("imgurl", imgurl, t1))
doc.add(Field("shijian",shijian,t1))
doc.add(Field("jianjie",jianjie,t1))
writer.addDocument(doc)
file.close()
except Exception, e:
print "Failed in indexDocs:", e
def indexDocs(self, root, writer):
t1 = FieldType()
t1.setIndexed(True)
t1.setStored(True)
t1.setTokenized(False)
t1.setIndexOptions(FieldInfo.IndexOptions.DOCS_AND_FREQS)
t2 = FieldType()
t2.setIndexed(True)
t2.setStored(False)
t2.setTokenized(True)
t2.setIndexOptions(FieldInfo.IndexOptions.DOCS_AND_FREQS_AND_POSITIONS)
for root, dirnames, filenames in os.walk(root):
for filename in filenames:
# if not filename.endswith('.txt'):
# continue
print "adding", filename
try:
path = os.path.join(root, filename)
pwd = os.path.join(os.getcwd() + '/' + root, filename)
file = open(path)
url = file.readline()
title = file.readline()
tag = file.readline()
imgurl = file.readline()
price = file.readline()
wellrate = file.readline()
comment = file.readline()
contents = unicode(file.read(), 'utf8')
file.close()
doc = Document()
doc.add(Field('url', url, t1))
doc.add(Field('title', title, t1))
doc.add(Field('imgurl', imgurl, t1))
doc.add(Field('price', price, t1))
doc.add(Field('wellrate', wellrate, t1))
doc.add(Field('comment', comment, t1))
if len(tag) > 2:
doc.add(Field('tag', tag, t2))
else:
doc.add(Field('tag', ' ', t2))
#doc.add(Field('comment', comment, t1))
writer.addDocument(doc)
except Exception, e:
print "Failed in indexDocs:", e
def testAdd(self, goodname, salenum, price, shopname, url, picturename, comment, historyprice):
analyzer = WhitespaceAnalyzer(Version.LUCENE_CURRENT)
analyzer = LimitTokenCountAnalyzer(analyzer, 1048576)
config = IndexWriterConfig(Version.LUCENE_CURRENT, analyzer)
config.setOpenMode(IndexWriterConfig.OpenMode.CREATE_OR_APPEND)
writer = IndexWriter(self.dir, config)
# True，建立新索引，False，建立增量索引
noIndexedString = FieldType()
noIndexedString.setTokenized(False)
noIndexedString.setIndexed(False)
noIndexedString.setStored(True)
try:
print "adding", goodname
goodname_s = unicode(goodname, 'utf8')
seg_list_good = jieba.cut(goodname_s, cut_all=False)
goodname_s = " ".join(seg_list_good) # 默认模式
shopname_s = unicode(shopname, 'utf8')
seg_list_shop = jieba.cut(shopname_s, cut_all=False)
shopname_s = " ".join(seg_list_shop) # 默认模式
shopnameField = Field("shopName", shopname, noIndexedString)
shopnameField_s = TextField("shopName_s", shopname_s, Field.Store.NO)
goodnameField = Field("goodName", goodname, noIndexedString)
goodnameField_s = TextField("goodName_s", goodname_s, Field.Store.NO)
salenumField = IntField("saleNum", salenum, Field.Store.YES)
priceField = DoubleField("price", price, Field.Store.YES)
urlField = Field("url", url, noIndexedString)
pictureField = StringField("pictureName", picturename, Field.Store.YES)
commentField = Field("comments", comment, noIndexedString)
historyPriceField = Field("historyPrice", historyprice, noIndexedString)
doc = Document()
doc.add(shopnameField)
doc.add(shopnameField_s)
doc.add(goodnameField)
doc.add(goodnameField_s)
doc.add(salenumField)
doc.add(priceField)
doc.add(urlField)
doc.add(pictureField)
doc.add(commentField)
doc.add(historyPriceField)
writer.addDocument(doc)
except Exception, e:
print "Failed in indexDocs:", e
def indexDocs(self, root, writer):
t1 = FieldType() #t1 is used in path and URL fields
t1.setIndexed(False)
t1.setStored(True)
t1.setTokenized(False)
t2 = FieldType() #t2 is used to index contents
t2.setIndexed(True)
t2.setStored(False)
t2.setTokenized(True)
t2.setIndexOptions(FieldInfo.IndexOptions.DOCS_AND_FREQS_AND_POSITIONS)
t3 = FieldType() #t3 is used to index titles
t3.setIndexed(True)
t3.setStored(True)
t3.setTokenized(True)
t3.setIndexOptions(FieldInfo.IndexOptions.DOCS_AND_FREQS_AND_POSITIONS)
t4 = FieldType() #t3 is used to index sites
t4.setIndexed(True)
t4.setStored(True)
t4.setTokenized(False)
t4.setIndexOptions(FieldInfo.IndexOptions.DOCS_AND_FREQS_AND_POSITIONS)
indextxt = open(self.filedir, 'r')
while True:
t = indextxt.readline()
if (len(t) == 0):
indextxt.close()
return
t = t.split()
filename = t[1]
URL = t[0]
# for root, dirnames, filenames in os.walk(root):
# for filename in filenames:
print "updating", filename
try:
writer.deleteDocuments(Term("name", filename))
path = os.path.join(root, filename)
file = open(path)
title = file.readline()
print title
contents = unicode(file.read())
file.close()
doc = Document()
doc.add(Field("name", filename, t1))
doc.add(Field("path", path, t1))
doc.add(Field("url", URL, t1))
doc.add(Field("title", title, t3))
doc.add(Field("site", get_site(URL), t4))
if len(contents) > 0:
doc.add(Field("contents", contents, t2))
else:
print "warning: no content in %s" % filename
writer.addDocument(doc)
except Exception, e:
print "Failed in indexDocs:", e
def indexDocs(self, dialog, root, writer):
t1 = FieldType()
t1.setIndexed(True)
t1.setStored(True)
t1.setTokenized(False)
t1.setIndexOptions(FieldInfo.IndexOptions.DOCS_AND_FREQS)
t2 = FieldType()
t2.setIndexed(True)
t2.setStored(False)
t2.setTokenized(True)
t2.setIndexOptions(FieldInfo.IndexOptions.DOCS_AND_FREQS_AND_POSITIONS)
f = open(dialog)
lines = f.readlines()
for line in lines:
info = line.split('\t')
url = info[0]
filename = info[1]
title = info[2]
print "adding", filename
try:
path = os.path.join(root, filename)
file = open(path)
contents = unicode(file.read(), 'utf-8')
file.close()
doc = Document()
doc.add(Field('name', filename, t1))
doc.add(Field('path', path, t1))
doc.add(
Field('url', url, Field.Store.YES,
Field.Index.NOT_ANALYZED))
if len(contents) > 0:
soup = BeautifulSoup(contents)
try:
title = soup.title.text
doc.add(
Field('title', title, Field.Store.YES,
Field.Index.ANALYZED))
except Exception, e:
doc.add(
Field('title', 'none', Field.Store.YES,
Field.Index.ANALYZED))
contents = ''.join(soup.findAll(text=True))
doc.add(Field("contents", analysis(contents), t2))
else:
print "warning: no content in %s" % filename
writer.addDocument(doc)
def main():
INDEX_DIR = "full_index1"
DOCUMENTS_DIR = "/media/joseph/Windows8_OS/Users/Joseph/AppData/Local/lxss/home/jwymbs23/data_science_projects/french_pamphlets/frc-data-master/OCR_text/"
# Initialize lucene and JVM
lucene.initVM(vmargs=['-Djava.awt.headless=true'])
print("lucene version is:", lucene.VERSION, '\n')
store = getStore(INDEX_DIR)
analyzer = getAnalyzer()
writer = getWriter(store=store, analyzer=analyzer, create=True)
#get list of documents
doc_list = getDoclist(DOCUMENTS_DIR)
ftype = FieldType()
ftype.setIndexOptions(IndexOptions.DOCS_AND_FREQS_AND_POSITIONS)
ftype.setTokenized(True)
ftype.setStoreTermVectors(True)
ftype.freeze()
for cd, doc_name in enumerate(doc_list):
if not cd % 1000:
print(cd, '--', len(doc_list))
with open(doc_name, 'r') as d:
doc_lines = d.readlines()
full_text = ''.join([i.strip() for i in doc_lines]).lower()
try:
# create a document that would we added to the index
doc = Document()
# Add fields to this document
#could process fname here instead of in the dataframe later
doc.add(
Field("identifier",
doc_name.split('/')[-1], TextField.TYPE_STORED)
) #Store.YES))#, Field.Index.ANALYZED))
doc.add(
Field("vectext", full_text, ftype)
) #TextField.TYPE_STORED, TermVector.YES, ))#Store.YES))#, Field.Index.ANALYZED))
doc.add(Field("text", full_text, TextField.TYPE_STORED))
# Add the document to the index
writer.addDocument(doc)
except:
print("Failed in indexDocs: ", doc_name)
#writer.optimize()
writer.commit()
def index_article(writer, art_id, art_body):
art_id_field = FieldType()
art_id_field.setIndexed(True)
art_id_field.setStored(True)
art_id_field.setTokenized(False)
art_id_field.setIndexOptions(FieldInfo.IndexOptions.DOCS_AND_FREQS)
art_body_field = FieldType()
art_body_field.setIndexed(True)
art_body_field.setStored(True)
art_body_field.setTokenized(True)
art_body_field.setIndexOptions(FieldInfo.IndexOptions.DOCS_AND_FREQS_AND_POSITIONS)
doc = Document()
doc.add(Field("art_id", str(art_id), art_id_field))
doc.add(Field("art_body", art_body, art_body_field))
writer.addDocument(doc)
def setUp(self):
super(Test_Bug1842, self).setUp()
self.analyzer = StandardAnalyzer()
w1 = self.getWriter(analyzer=self.analyzer)
doc1 = Document()
ftype = FieldType()
ftype.setIndexOptions(IndexOptions.DOCS_AND_FREQS_AND_POSITIONS)
ftype.setTokenized(True)
ftype.setStoreTermVectors(True)
ftype.freeze()
doc1.add(Field("all", "blah blah blah Gesundheit", ftype))
doc1.add(Field('id', '1', StringField.TYPE_NOT_STORED))
w1.addDocument(doc1)
w1.close()
def lazyImport():
global imported
if imported:
return
from meresco.pylucene import getJVM
getJVM()
from java.nio.file import Paths
from org.apache.lucene.document import Document, StringField, Field, FieldType
from org.apache.lucene.search import IndexSearcher, TermQuery
from org.apache.lucene.index import DirectoryReader, Term, IndexWriter, IndexWriterConfig, IndexOptions
from org.apache.lucene.store import FSDirectory
from org.apache.lucene.util import Version
from org.apache.lucene.analysis.core import WhitespaceAnalyzer
UNINDEXED_TYPE = FieldType()
UNINDEXED_TYPE.setIndexOptions(IndexOptions.NONE)
UNINDEXED_TYPE.setStored(True)
UNINDEXED_TYPE.setTokenized(False)
imported = True
globals().update(locals())
def indexDocs(self, root, writer):
#Create a new FieldType with default properties.
t1 = FieldType()
t1.setIndexed(True)
t1.setStored(True)
t1.setTokenized(False)#True if this field's value should be analyzed by the Analyzer.
t1.setIndexOptions(FieldInfo.IndexOptions.DOCS_AND_FREQS)
#Create a new FieldType with default properties.
t2 = FieldType()
t2.setIndexed(True)
t2.setStored(True)
t2.setTokenized(True)#True if this field's value should be analyzed by the Analyzer.
t2.setIndexOptions(FieldInfo.IndexOptions.DOCS_AND_FREQS_AND_POSITIONS_AND_OFFSETS)
for root, dirnames, filenames in os.walk(root):
for filename in filenames:
if not filename.endswith('.txt'):
continue
print 'adding', filename
try:
path = os.path.join(root, filename)
file = open(path)
contents = file.read()
file.close()
doc = Document()
doc.add(Field('name', filename, t1))
doc.add(Field('path', root, t1))
if len(contents) > 0:
doc.add(Field('contents', contents, t2))
print 'length of content is %d'%(len(contents))
else:
print 'warning: no content in %s' % filename
writer.addDocument(doc)
except Exception, e:
print 'Failed in indexDocs:', e
def indexDocs(self,root,writer):
t1 = FieldType()
t1.setIndexed(True)
t1.setStored(True)
t1.setTokenized(True)
t1.setIndexOptions(FieldInfo.IndexOptions.DOCS_AND_FREQS)
t2 = FieldType()
t2.setIndexed(True)
t2.setStored(False)
t2.setTokenized(True)
t2.setIndexOptions(FieldInfo.IndexOptions.DOCS_AND_FREQS_AND_POSITIONS)
for root, dirnames,filenames in os.walk(root):
# traverse through the doc directory
for filename in filenames:
# only if this file ends with '.c'
if not filename.endswith('.c'):
continue
try:
# only add the filename and path for indexing
path = os.path.join(root,filename)
print "adding file : ",path
file = open(path)
contents = unicode(file.read(),'utf-8')
file.close()
doc = Document()
doc.add(Field("name",filename,t1))
doc.add(Field("path",root,t1))
# if len(contents) > 0:
# doc.add(Field("contents",contents,t2))
# else:
# print "warning: no content in ",filename
writer.addDocument(doc)
except Exception,e:
print "failed in indexDocs:",e
def index_wiki(wiki_xmlfile, index_directory_name):
lucene.initVM()
# Initialize index directory and analyzer.
version = Version.LUCENE_CURRENT
store = FSDirectory.open(File(index_directory_name))
analyzer = StandardAnalyzer(version)
# Creates config file.
config = IndexWriterConfig(version, analyzer)
config.setOpenMode(IndexWriterConfig.OpenMode.CREATE)
writer = IndexWriter(store, config)
# Set document content field type.
content_fieldtype = FieldType()
content_fieldtype.setIndexed(True)
content_fieldtype.setStored(True)
content_fieldtype.setTokenized(True)
content_fieldtype.setIndexOptions(FieldInfo.IndexOptions.DOCS_AND_FREQS_AND_POSITIONS)
# Set document title field type.
title_fieldtype = FieldType()
title_fieldtype.setIndexed(True)
title_fieldtype.setStored(True)
title_fieldtype.setTokenized(True)
title_fieldtype.setIndexOptions(FieldInfo.IndexOptions.DOCS_AND_FREQS_AND_POSITIONS)
# Set document url field type.
url_fieldtype = FieldType()
url_fieldtype.setIndexed(True)
url_fieldtype.setStored(True)
url_fieldtype.setTokenized(False)
url_fieldtype.setIndexOptions(FieldInfo.IndexOptions.DOCS_AND_FREQS_AND_POSITIONS)
for xmldoc in wikicorpusxml((wiki_xmlfile)):
content = xmldoc.partition('>')[2].partition('<')[0].strip()
title = xmldoc.partition(' title="')[2].partition('"')[0].strip()
url = xmldoc.partition(' url="')[2].partition('"')[0].strip()
doc = Document()
doc.add(Field("contents", content, content_fieldtype))
doc.add(Field("title", title, title_fieldtype))
doc.add(Field("url", url, url_fieldtype))
writer.addDocument(doc)
writer.commit()
writer.close()
def __init__(self):
self.mDocumentDirectory = settings.ADMINS_ENGINE.mDocumentDirectory
self.mIndexDirectory = settings.ADMINS_ENGINE.mIndexDirectory
self.mAnalyzers = settings.ADMINS_ENGINE.getIndexingAnalyzers()
############################# Writer Configurattion #####################################
map = HashMap()
map.put('name', self.mAnalyzers['name'])
map.put('parent', self.mAnalyzers['parent'])
map.put('content', self.mAnalyzers['default'])
map.put('id', self.mAnalyzers['id'])
analyzerWrapper = PerFieldAnalyzerWrapper(self.mAnalyzers['default'], map)
self.mWriterConfig = IndexWriterConfig(Version.LUCENE_CURRENT, analyzerWrapper)
self.mWriterConfig.setOpenMode(settings.ADMINS_ENGINE.mOpenMode)
if settings.ADMINS_ENGINE.mSimilarity != None:
self.mWriterConfig.setSimilarity(settings.ADMINS_ENGINE.mSimilarity)
########################################################################################
directory = SimpleFSDirectory(File(self.mIndexDirectory))
self.mIndexWriter = IndexWriter(directory, self.mWriterConfig)
############################# FieldType Prepration #####################
nameField = FieldType()
nameField.setIndexed(True)
nameField.setStored(True)
nameField.setTokenized(True)
nameField.setIndexOptions(FieldInfo.IndexOptions.DOCS_ONLY)
parentField = FieldType()
parentField.setIndexed(True)
parentField.setStored(True)
parentField.setTokenized(True)
parentField.setIndexOptions(FieldInfo.IndexOptions.DOCS_ONLY)
contentField = FieldType()
contentField.setIndexed(True)
contentField.setStored(True)
contentField.setTokenized(True)
contentField.setIndexOptions(FieldInfo.IndexOptions.DOCS_AND_FREQS_AND_POSITIONS_AND_OFFSETS)
idField = FieldType()
idField.setIndexed(True)
idField.setStored(True)
idField.setTokenized(False)
idField.setIndexOptions(FieldInfo.IndexOptions.DOCS_ONLY)
self.mFieldTypes = {
'name' : nameField,
'parent' : parentField,
'content' : contentField,
'id' : idField
}
#######################################################################
self.mLog = ""
def indexCranFull(path, writer):
"""
This method reads in the cran.1400 file and creates an index out of it.
These fields are used when storing documents:
title: title of the document (in a line starting with .T)
author: author of the document (in a line starting with .A)
source: where this article has been published, not yet used (in a line starting with .B)
content: body text of the article (in a line starting with .W)
"""
# Title field type
tft = FieldType()
tft.setIndexed(True)
tft.setStored(True)
tft.setTokenized(TokenizeFields)
tft.setIndexOptions(FieldInfo.IndexOptions.DOCS_AND_FREQS) #only index the document and frequency data
# Author field type
aft = FieldType()
aft.setIndexed(True)
aft.setStored(True)
aft.setTokenized(TokenizeFields)
aft.setIndexOptions(FieldInfo.IndexOptions.DOCS_AND_FREQS_AND_POSITIONS) #index the document, term frequency and position data
# Content field type
cft = FieldType()
cft.setIndexed(True)
cft.setStored(True)
cft.setTokenized(True)
cft.setIndexOptions(FieldInfo.IndexOptions.DOCS_AND_FREQS_AND_POSITIONS)
txt = open(path)
"""
cran documents are listed in this order:
.I # -> ignore
.T -> starting on the next line, multi-line title
.A -> on the next line, multiple authored separated by 'and'
.B -> on the next line, ignore
.W -> starting on the next line, multi-line body (content)
"""
docid = 0
debug = False
while True: # in each iteration, read all the lines corresponding to a document
line = txt.readline()
if line == '':
break
if docid == 1400:
debug = True
if line.startswith('.I'):
docid = int(line.split(' ')[1].strip())
continue
if line.startswith('.T'):
title = ''
while True:
line = txt.readline()
if line.startswith('.A'):
break
title = ' '.join([title, line]).strip()
line = txt.readline()
# authors = ' '.join(line.split('and'))
authors = line.strip()
#.B, its corresponding line and .W
txt.readline()
txt.readline()
txt.readline()
body = ''
while True:
line = txt.readline()
if line.startswith('.I'):
docid = int(line.split(' ')[1].strip())
break
if line == '':
break
body = ' '.join([body, line]).strip()
doc = Document()
doc.add(Field(title_field, title, tft))
doc.add(Field(author_field, authors, aft))
doc.add(Field(content_field, body, cft))
doc.add(IntField(docid_field, docid, Field.Store.YES))
writer.addDocument(doc)
def indexDocs(root, writer):
"""
indexed: name title content
stored: date name tilte sumary
:param root:
:param writer:
:return:
"""
#index and store
t1 = FieldType()
t1.setIndexed(True)
t1.setStored(True)
t1.setTokenized(False)
t1.setIndexOptions(FieldInfo.IndexOptions.DOCS_AND_FREQS)
#only index, but not store
t2 = FieldType()
t2.setIndexed(True)
t2.setStored(False)
t2.setTokenized(True)
t2.setIndexOptions(FieldInfo.IndexOptions.DOCS_AND_FREQS_AND_POSITIONS)
#only store
t3 = FieldType()
t3.setIndexed(False)
t3.setStored(True)
t3.setTokenized(False)
t3.setIndexOptions(FieldInfo.IndexOptions.DOCS_AND_FREQS)
for root, dirnames, filenames in os.walk(root):
print filenames
for filename in filenames:
if not filename.endswith('.md'):
continue
print "adding", filename
try:
path = os.path.join(root, filename)
file = open(path)
contents = unicode(file.read(), 'utf-8')
file.close()
date, name = get_date_name(filename)
title, content = get_post_title_content(contents)
summary = content[:200] if content else ''
print date, name, title
doc = Document()
doc.add(Field('date', date, t3))
doc.add(Field('name', name, t1))
doc.add(Field('title', title, t1))
doc.add(Field('content', content, t2))
doc.add(Field('summary', summary, t3))
# doc.add(Field("name", filename, t1))
# doc.add(Field("path", root, t1))
# if len(contents) > 0:
# doc.add(Field("contents", contents, t2))
# else:
# print "warning: no content in %s" % filename
writer.addDocument(doc)
except Exception, e:
print "Failed in indexDocs:", e
from org.apache.lucene.store import RAMDirectory
from org.apache.lucene.document import Document, Field, FieldType
from org.apache.lucene.util import BytesRef, BytesRefIterator
from org.apache.lucene.index import \
IndexWriterConfig, IndexWriter, DirectoryReader, IndexOptions
if __name__ == '__main__':
lucene.initVM(vmargs=['-Djava.awt.headless=true'])
directory = RAMDirectory()
iconfig = IndexWriterConfig(LimitTokenCountAnalyzer(StandardAnalyzer(), 100))
iwriter = IndexWriter(directory, iconfig)
ft = FieldType()
ft.setStored(True)
ft.setTokenized(True)
ft.setStoreTermVectors(True)
ft.setStoreTermVectorOffsets(True)
ft.setStoreTermVectorPositions(True)
ft.setIndexOptions(IndexOptions.DOCS_AND_FREQS_AND_POSITIONS_AND_OFFSETS)
ts = ["this bernhard is the text to be index text",
"this claudia is the text to be indexed"]
for t in ts:
doc = Document()
doc.add(Field("fieldname", t, ft))
iwriter.addDocument(doc)
iwriter.commit()
iwriter.close()
ireader = DirectoryReader.open(directory)
class Worker(Process):
"""This class represents the lucene worker object, which will run
concurrently with other workers, and contains the actual lucene logic."""
def __init__(self, config, queue, ret):
Process.__init__(self)
self.config = config
self.connection = None
self.channel = None
self.queue = queue
self.ret = ret
def run(self):
print "Booting lucene driver worker...."
lucene.initVM()
self.fieldType1 = FieldType()
self.fieldType1.setIndexed(True)
self.fieldType1.setStored(False)
self.fieldType1.setTokenized(True)
self.fieldType2 = FieldType()
self.fieldType2.setIndexed(True)
self.fieldType2.setStored(True)
self.fieldType2.setTokenized(False)
while(True):
data = self.queue.get()
da = data[1]
response = None
try:
self.fil = File(da['data']['indexdir'])
self.d = NIOFSDirectory(self.fil)
self.analyzer = StandardAnalyzer(Version.LUCENE_CURRENT)
self.conf = IndexWriterConfig(
Version.LUCENE_CURRENT,
self.analyzer)
response = getattr(self, da['action'])(da['data'])
self.d.close()
except Exception as e:
print e
if response is None:
response = {}
self.ret[data[0]] = response
def rebuildIndex(self, data):
writer = IndexWriter(
self.d, self.conf.setOpenMode(IndexWriterConfig.OpenMode.CREATE))
for record in data['records']:
doc = self.buildDocument(data['fields'], record)
writer.addDocument(doc)
writer.commit()
writer.close()
def buildDocument(self, fields, record):
doc = Document()
doc.add(
Field("id",
record["_id"],
self.fieldType2))
for field in fields:
if isinstance(record[field], dict):
self.dictToFields(doc, record[field])
else:
doc.add(
Field(field,
record[field],
self.fieldType1))
return doc
def dictToFields(self, doc, record):
for key in record:
if isinstance(record[key], dict):
self.dictToFields(doc, record[key])
else:
doc.add(
Field(key,
record[key],
self.fieldType1))
def index(self, data):
writer = IndexWriter(
self.d, self.conf)
doc = self.buildDocument(data['fields'], data['record'])
writer.addDocument(doc)
writer.commit()
writer.close()
def updateindex(self, data):
writer = IndexWriter(
self.d, self.conf)
doc = self.buildDocument(data['fields'], data['record'])
writer.updateDocument(lucene.Term("_id", data['record']['_id']), doc)
writer.optimize()
writer.close()
def removeindex(self, data):
writer = IndexWriter(
self.d, self.conf)
writer.deleteDocuments(lucene.Term("_id", data['record']['_id']))
writer.optimize()
writer.close()
def query(self, data):
if self.fil.exists():
searcher = IndexSearcher(DirectoryReader.open(self.d))
query = QueryParser(
Version.LUCENE_30,
"id",
self.analyzer).parse(
data['query'])
hits = searcher.search(query, 100000)
results = {}
results['totalHits'] = hits.totalHits
results['hits'] = {}
for hit in hits.scoreDocs:
record = {}
doc = searcher.doc(hit.doc)
fields = doc.getFields()
record['score'] = hit.score
for field in fields:
if field.name() != "id":
record[field.name()] = field.stringValue()
results['hits'][doc.get('id')] = record
searcher.getIndexReader().close()
return results
def reindex(self):
''' Re-indexes the entire database into Index file'''
start = time.time()
# get all posts
posts = self._tuples_to_dict(self._fetch_all_questions(), self._posts_fields)
if not posts:
raise Exception("FATAL Error: Could not fetch posts from Database")
# open indexer
# lucene.initVM(vmargs=['-Djava.awt.headless=true'])
# print 'lucene', lucene.VERSION
store = SimpleFSDirectory(File(self.index_dir))
analyzer = StandardAnalyzer(Version.LUCENE_CURRENT)
config = IndexWriterConfig(Version.LUCENE_CURRENT, analyzer)
config.setOpenMode(IndexWriterConfig.OpenMode.CREATE_OR_APPEND)
writer = IndexWriter(store, config)
indexedField = FieldType()
indexedField.setIndexed(True)
indexedField.setStored(True)
indexedField.setTokenized(True)
indexedField.setIndexOptions(FieldInfo.IndexOptions.DOCS_AND_FREQS_AND_POSITIONS)
storedField = FieldType()
storedField.setIndexed(False)
storedField.setStored(True)
storedField.setTokenized(False)
storedField.setIndexOptions(FieldInfo.IndexOptions.DOCS_AND_FREQS)
fieldTypes = {
'type' : storedField,
'id' : storedField,
'title' : indexedField,
'question' : indexedField,
'answer' : indexedField,
# 'comment' : indexedField,
'tag' : indexedField,
'extra' : indexedField,
}
# get their comments
num_docs = 0
for post in posts:
if self.status_mode: print "\r {0:.2f} %complete".format(((num_docs/142627.0)*100)),
if self.debug : print "\n","*"*20,"\nIndexing post: ", post['id'], "from ", post['extra']
if self.debug and self.verbose_values: print post
answers = self._tuples_to_dict(self._fetch_all_answers(post['id'], post['extra']), self._answer_fields)
# add comment field
for answer in answers:
num_docs += 1
if self.debug: print "\n","+"*10, "\nMaking new Document"
doc = Document()
if self.debug: print "Adding doc type"
doc.add(Field("type", self.doctype, fieldTypes['type']))
# make fields
if self.debug: print "Adding post fields"
for i in xrange(len(self._posts_fields)):
f = Field(self._posts_fields[i], self._cleanup_tag(post[self._posts_fields[i]]), fieldTypes[self._posts_fields[i]])
f.setBoost(self._fields_boost[self._posts_fields[i]])
doc.add(f)
if self.status_mode: print "\t Indexing answer: ", answer['answer_id']
if self.debug and self.verbose_values: print answer
# answered_doc = copy.deepcopy(doc)
# make comment field
f = Field("answer", self._cleanup_tag(answer['answer']), fieldTypes['answer'])
f.setBoost(self._fields_boost['answer'])
doc.add(f)
# calculate paths
# commented_doc = copy.deepcopy(answered_doc)
# comments = self._comments_to_comment_string(self._tuples_to_dict(self._fetch_all_comments(answer['id']), self._comment_fields))
# if self.debug: print "\t\tAdding comments: ", comments
# commented_doc.add(Field("comment", self._cleanup_tag(comments), fieldTypes['comment']))
# write index
if self.debug: print "\tAdding document {doc_id} to index".format(doc_id=post['id'])
writer.addDocument(doc)
# del answered_doc
# del commented_doc
if self.debug: print "Commiting document to index"
writer.commit()
# close index
if self.status_mode: print "Closing index write"
writer.close()
end = time.time() - start
if self.status_mode: print "\n","-"*20, \
"\nTotal time spent in indexing: ", end, "seconds" \
"\nIndexed {num_docs} documents".format(num_docs=num_docs)
def indexTable(self, writer):
#connection
con = None
#define the index of all the fields
#---------step 2：connect to mysql----------
con = mdb.connect('localhost','root','testgce','douban_movie_v3')
#t_num = FieldType.NumericType it is wrong!!
t_num = FieldType()
t_num.setStored(False)
t1 = FieldType()
t1.setIndexed(True)
t1.setStored(True)
t1.setTokenized(False)
t1.setIndexOptions(FieldInfo.IndexOptions.DOCS_AND_FREQS)
t2 = FieldType()
t2.setIndexed(True)
t2.setStored(False)
t2.setTokenized(True)
t2.setIndexOptions(FieldInfo.IndexOptions.DOCS_AND_FREQS_AND_POSITIONS)
t3 = FieldType()
t3.setIndexed(True)
t3.setStored(True)
t3.setTokenized(True)
t3.setIndexOptions(FieldInfo.IndexOptions.DOCS_AND_FREQS)
maxDict = utils.maxDict
#加权数值范围
base = DOC_BOOST_RANGE[0]
upper = DOC_BOOST_RANGE[1]
with con:
# Careful with codecs
con.set_character_set('utf8')
cur = con.cursor()
# Aagin the codecs
cur.execute('SET NAMES utf8;')
cur.execute('SET CHARACTER SET utf8;')
cur.execute('SET character_set_connection=utf8;')
#------step 3： choose the right table------
cur.execute("SELECT * FROM movie_items")
numrows = int(cur.rowcount)
print 'numrows:',numrows
for i in range(numrows):
print
row = cur.fetchone()
#------step 4：Index your field------
summary = row[SUMMARY]
subject_id = row[SUBJECT_ID]
print 'id'+subject_id
year = utils.formatYear(row[YEAR])
try:
date = DateTools.stringToDate(year.replace('-',' '))
wtfFile = open('wtf.txt','a')
dateStr = DateTools.dateToString(date,DateTools.Resolution.DAY)
except:
wtfFile.write(year+'\n')
doc = Document()
#boosting
boostProb = utils.calcBoostProb(row,maxDict,dateStr)
boost = base + boostProb*(upper-base)
doc.add(FloatField("boost",boost,Field.Store.YES))
doc.add(StringField("year",dateStr,Field.Store.YES))
print 'dateStr:'+dateStr
#A text field is a sequence of terms that has been tokenized while a string field is a single term (although it can also be multivalued.)
do_count = row[DO_COUNT] if row[DO_COUNT] != None else 0
wish_count = row[COLLECT_COUNT] if row[WISH_COUNT] != None else 0
#fields which should not be analyzed
doc.add(FloatField("rating_average",float(row[RATING_AVERAGE]),Field.Store.YES))
doc.add(FloatField("rating_stars", float(row[RATING_STARS]), Field.Store.YES))
doc.add(IntField("reviews_count", int(row[REVIEWS_COUNT]), Field.Store.YES))
#doc.add(FloatField("year", float(row[YEAR]), Field.Store.YES).setBoost(boost))
doc.add(IntField("collect_count", int(row[COLLECT_COUNT]), Field.Store.YES))
doc.add(IntField("do_count", int(do_count), Field.Store.YES))
doc.add(IntField("wish_count", int(wish_count), Field.Store.YES))
doc.add(IntField("subject_id", int(row[SUBJECT_ID]), Field.Store.YES))
doc.add(IntField("comments_count", int(row[COMMENTS_COUNT]), Field.Store.YES))
doc.add(IntField("ratings_count", int(row[RATINGS_COUNT]), Field.Store.YES))
doc.add(StringField("image_small", row[IMAGE_SMALL], Field.Store.YES))
#fields which should be analyzed with WhitespaceAnalyzer
#attention!!! dont use a long sentence like :
#doc.add(Field("genres", row[GENRES].replace(delim,' '), t3).setBoost(boost))
#or you'll get a null pointer error
f = Field("countries", row[COUNTRIES].replace(delim,' '), t3)
f.setBoost(boost)
doc.add(f)
#process casts
raw_casts = row[CASTS].replace(delim,' ')
f = Field("raw_casts", raw_casts , t1)
f.setBoost(boost)
doc.add(f)
#将英文人名中的 ·
raw_casts = raw_casts.replace('·',' ')
if len(raw_casts.split(' '))<CASTS_LEN:
#平局人名长度是4
casts = raw_casts + ' ￥￥￥￥'*(CASTS_LEN-len(raw_casts.split(' ')))
f = Field("casts", casts , t3)
f.setBoost(boost)
doc.add(f)
#process directors
raw_directors = row[DIRECTORS].replace(delim,' ')
f = Field("raw_directors",raw_directors, t1)
f.setBoost(boost)
doc.add(f)
#将英文人名中的 · 替换
raw_directors = raw_directors.replace('·',' ')
if len(raw_directors.split(' '))<DIRECTORS_LEN:
#平局人名长度是4
directors = raw_directors + ' ￥￥￥￥'*(DIRECTORS_LEN-len(raw_directors.split(' ')))
f = Field("directors", directors, t3)
f.setBoost(boost)
doc.add(f)
Field("genres", row[GENRES].replace(delim,' '), t3)
f.setBoost(boost)
doc.add(f)
Field("subtype", row[SUBTYPE].replace(delim,' '), t3)
f.setBoost(boost)
doc.add(f)
#it is wrong cause indexable field has no method setBoost
# fieldList = doc.getFields() # is not a python 'list' , but a 'List' which is unindexable
# for eachField in fieldList:
# eachField.setBoost(boost)
#user_tags 原始字符串要存，reRank要用：
doc.add(StringField("raw_user_tags",row[USER_TAGS],Field.Store.YES))
doc.add(StringField("raw_others_like",row[OTHERS_LIKE],Field.Store.YES))
user_tags_str = ''
others_like_str = ''
tags_len = 0
if row[USER_TAGS]!='':
user_tags_list = row[USER_TAGS].split(delim)
for tag_pair in user_tags_list:
if tag_pair!='':#字符串的最后一个字符是￥，这样split之后最后一个元素是空字符
#print 'tag_pair'+tag_pair+'hhe'
tag_name = tag_pair.split(delim_uo)[0]+' ' # dont forget this space !!
tag_num = tag_pair.split(delim_uo)[1]
tag_num_processed = int(int(tag_num)/TAG_SPAN)+1 #最小为1
#!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
user_tags_str = user_tags_str +' '+ tag_name * tag_num_processed
tags_len = tags_len + tag_num_processed #最后得到总共词的个数
if tags_len<TAGS_AVER_LEN:
#填充tags，目测3是平均长度,所以使用 ￥￥￥
user_tags_str = user_tags_str +' ￥￥￥'*(TAGS_AVER_LEN - tags_len)
#
if row[OTHERS_LIKE]!='':
for like_pair in row[OTHERS_LIKE].split(delim):
if like_pair!='':
others_like_str = others_like_str +' '+like_pair.split(delim_uo)[1]
#start process adjs
if row[ADJS] != None:
raw_adjs = row[ADJS][:-1]
adjs_str = ''
adjs_len = 0
if row[ADJS] != '' and row[ADJS] != '\n':
#'重要=4.0,特殊=4.0'
adjs_str = row[ADJS]
adjs_list = adjs_str.split(',')
for adj_pair in adjs_list:
#print 'adj_pair:'+adj_pair+'hhe'
adj_name = adj_pair.split('=')[0]
adj_num = adj_pair.split('=')[1]
#去换行符,转换int
if adj_num[-1] == '\n':
adj_num = adj_num[0:-1]
adj_num = int(float(adj_num))
add_adj=''
# #同义词
# adj_name_bro = searchDictValue(adjMap,adj_name)
# if adj_name_bro == -1: #表示没有结果，即未找到近义词，不添加
# add_adj = ''
# else:
# add_adj = (adj_name_bro+' ')*adj_num
# raw_adjs = raw_adjs + ',' + adj_name_bro+'='+str(adj_num)
adjs_str = adjs_str + ' ' + (adj_name+' ') * adj_num +add_adj
adjs_len = adjs_len + adj_num #最后得到总共tags的个数
#print raw_adjs
doc.add(StringField("raw_adjs",raw_adjs,Field.Store.YES))
if adjs_len<ADJS_AVER_LEN:
#填充 adjs_str，目测2是平均长度,所以使用 "￥￥"
adjs_str = adjs_str +' ￥￥'*(ADJS_AVER_LEN - adjs_len)
f = Field("adjs", adjs_str, t3)
f.setBoost(boost)
doc.add(f)
f = Field("user_tags", user_tags_str, t3)
f.setBoost(boost)
doc.add(f)
f = Field("others_like", others_like_str, t3)
f.setBoost(boost)
doc.add(f)
#fields which should be analyzed with good analyzer
f = Field("title", row[TITLE], t3)
f.setBoost(boost)
doc.add(f)
f = Field("original_title", row[ORIGINAL_TITLE], t3)
f.setBoost(boost)
doc.add(f)
f = Field("summary_segmentation", row[SUMMARY_SEGMENTATION], t3)
f.setBoost(boost)
doc.add(f)
f = Field("aka", row[AKA], t2)
f.setBoost(boost)
doc.add(f)
if len(summary) > 0:
print subject_id +'--->'+':\n '+ row[TITLE]
try:
summary_unicoded = unicode(summary, 'utf-8') #test the encoding
except Exception,e:
print "Decode Failed: ", e
f = Field('summary', summary, t2)
f.setBoost(boost)
doc.add(f)
else:
print "warning:\n" + subject_id +'---> No content!'
print 'boosting:' + str(boost)
#for debug
if boost>upper:
print boostProb
print maxDict
exit(0)
writer.addDocument(doc)
def indexTable(self, writer):
#connection
con = None
#define the index of all the fields
#---------step 2----------
con = mdb.connect('localhost','root','testgce','moviedata')
#t_num = FieldType.NumericType it is wrong!!
t_num = FieldType()
t_num.setStored(False)
t1 = FieldType()
t1.setIndexed(True)
t1.setStored(True)
t1.setTokenized(False)
t1.setIndexOptions(FieldInfo.IndexOptions.DOCS_AND_FREQS)
t2 = FieldType()
t2.setIndexed(True)
t2.setStored(False)
t2.setTokenized(True)
t2.setIndexOptions(FieldInfo.IndexOptions.DOCS_AND_FREQS_AND_POSITIONS)
t3 = FieldType()
t3.setIndexed(True)
t3.setStored(True)
t3.setTokenized(True)
t3.setIndexOptions(FieldInfo.IndexOptions.DOCS_AND_FREQS)
with con:
# Careful with codecs
con.set_character_set('utf8')
cur = con.cursor()
# Aagin the codecs
cur.execute('SET NAMES utf8;')
cur.execute('SET CHARACTER SET utf8;')
cur.execute('SET character_set_connection=utf8;')
#------step 3------
cur.execute("SELECT * FROM movie_items")
numrows = int(cur.rowcount)
print 'numrows:',numrows
for i in range(numrows):
row = cur.fetchone()
#------step 4------
summary = row[SUMMARY]
subject_id = row[SUBJECT_ID]
print 'id'+subject_id
#print 'summary'+summary+'end'
doc = Document()
#fields which should not be analyzed
doc.add(FloatField("rating_average",float(row[RATING_AVERAGE]),Field.Store.NO))
doc.add(FloatField("rating_stars", float(row[RATING_STARS]), Field.Store.NO))
doc.add(IntField("reviews_count", int(row[REVIEWS_COUNT]), Field.Store.NO))
#doc.add(FloatField("year", float(row[YEAR]), Field.Store.NO))
doc.add(IntField("collect_count", int(row[COLLECT_COUNT]), Field.Store.NO))
doc.add(IntField("subject_id", int(subject_id), Field.Store.YES))
doc.add(IntField("comments_count", int(row[COMMENTS_COUNT]), Field.Store.NO))
doc.add(IntField("ratings_count", int(row[RATINGS_COUNT]), Field.Store.NO))
doc.add(Field("image_small", row[IMAGE_SMALL], t1))
#fields which should be analyzed with WhitespaceAnalyzer
doc.add(Field("countries", row[COUNTRIES].replace(delim,' '), t3))
doc.add(Field("casts", row[CASTS].replace(delim,' '), t3))
doc.add(Field("genres", row[GENRES].replace(delim,' '), t3))
doc.add(Field("subtype", row[SUBTYPE].replace(delim,' '), t2))
doc.add(Field("directors", row[DIRECTORS].replace(delim,' '), t3))
user_tags_str = ''
others_like_str = ''
# print 'user_tags'+row[USER_TAGS]
# print 'others_like'+row[OTHERS_LIKE]
if row[USER_TAGS]!='':
for tag_pair in row[USER_TAGS].split(delim):
if tag_pair!='':#字符串的最后一个字符是:，这样split之后最后一个元素是空字符
user_tags_str = user_tags_str +' '+tag_pair.split(delim_uo)[0]
if row[OTHERS_LIKE]!='':
for like_pair in row[OTHERS_LIKE].split(delim):
if like_pair!='':
others_like_str = others_like_str +' '+like_pair.split(delim_uo)[1]
# print user_tags_str
# print others_like_str
doc.add(Field("user_tags", user_tags_str, t3))
doc.add(Field("others_like", others_like_str, t3))
#fields which should be analyzed with good analyzer
doc.add(Field("title", row[TITLE], t3))
doc.add(Field("original_title", row[ORIGINAL_TITLE], t2))
doc.add(Field("summary_segmentation", row[SUMMARY_SEGMENTATION], t2))
doc.add(Field("aka", row[AKA], t2))
if len(summary) > 0:
print subject_id +'--->'+':\n '+ row[TITLE]
try:
summary_unicoded = unicode(summary, 'utf-8') #test the encoding
except Exception,e:
print "Decode Failed: ", e
doc.add(Field('summary', summary, t2))
else:
print "warning:\n" + subject_id +'---> No content!'
writer.addDocument(doc)
def tweetIndexer(self, writer):
t1 = FieldType()
t1.setIndexed(True)
t1.setStored(True)
t1.setTokenized(False)
t1.setIndexOptions(FieldInfo.IndexOptions.DOCS_AND_FREQS)
t2 = FieldType()
t2.setIndexed(True)
t2.setStored(True)
t2.setTokenized(True)
t2.setStoreTermVectorOffsets(True)
t2.setStoreTermVectorPayloads(True)
t2.setStoreTermVectorPositions(True)
t2.setStoreTermVectors(True)
t2.setIndexOptions(FieldInfo.IndexOptions.DOCS_AND_FREQS_AND_POSITIONS)
x = 0
for i in range(0,500):
if not os.path.isfile("json/tweets-" + str(i) + ".json"):
break
print "adding tweets-" + str(i) + ".json"
tweets = open("json/tweets-" + str(i) + ".json", "r")
for line in tweets.readlines():
tweet = json.loads(line)
if 'limit' in tweet:
continue
try:
doc = Document()
doc.add(Field("file", "json/tweets-" + str(i) + ".json", t1))
sname = tweet['user']['screen_name']
tid = str(tweet['id'])
text = tweet['text']
uname = tweet['user']['name']
created = tweet['created_at']
tstamp = tweet['timestamp_ms']
place = ""
if tweet['place']:
place = tweet['place']['full_name'] + ", " + tweet['place']['country']
lat = ""
lng = ""
titles = ""
urls = ""
exist = "false"
if tweet['coordinates']:
lat = str(tweet['coordinates']['coordinates'][1])
lng = str(tweet['coordinates']['coordinates'][0])
else:
lat = str((tweet['place']['bounding_box']['coordinates'][0][0][1] + tweet['place']['bounding_box']['coordinates'][0][2][1])/2)
lng = str((tweet['place']['bounding_box']['coordinates'][0][0][0] + tweet['place']['bounding_box']['coordinates'][0][2][0])/2)
if len(tweet['entities']['urls']) != 0:
exist = "true"
for index in range(len(tweet['entities']['urls'])):
title = tweet['entities']['urls'][index]['url_title']
if title == None:
titles += ",-"
else:
title = title.encode('ascii','ignore')
titles += "," + str(title)
urls += " " + str(tweet['entities']['urls'][index]['expanded_url'])
searchable = text + " " + urls + " " + uname + " " + sname + " " + place
doc.add(Field("lookup", searchable, t2))
doc.add(Field("text", text, t2))
doc.add(Field("user_name", uname, t2))
doc.add(Field("screen_name", sname, t2))
doc.add(Field("tweet_id", tid, t2))
doc.add(Field("created_at", created, t2))
doc.add(Field("geo_lat", lat, t2))
doc.add(Field("geo_lng", lng, t2))
doc.add(Field("url_exist", exist, t2))
doc.add(Field("url_url", urls, t2))
doc.add(Field("url_title", titles, t2))
doc.add(Field("timestamp", tstamp, t2))
writer.addDocument(doc)
x += 1
except Exception, e:
pass
tweets.close()
config = IndexWriterConfig(Version.LUCENE_43, analyzer)
writer = IndexWriter(directory, config)
return writer
def open_searcher(writer):
from org.apache.lucene.search import IndexSearcher
reader = writer.getReader()
searcher = IndexSearcher(reader)
return reader, searcher
from org.apache.lucene.document import Document, Field, FieldType, TextField, StringField
from org.apache.lucene.util import BytesRef, BytesRefIterator
from org.apache.lucene.index import Term
vectorFieldType = FieldType(TextField.TYPE_NOT_STORED)
vectorFieldType.setIndexed(True)
vectorFieldType.setTokenized(True)
vectorFieldType.setStoreTermVectors(True)
vectorFieldType.setStoreTermVectorPositions(False)
writer = open_writer('data/index')
def addToIndex(lxmlNode):
uri = xpathFirst(lxmlNode, '//oa:hasTarget/@rdf:resource')
print uri
seen = set()
doc = Document()
for fieldName in FIELD_NAMES:
seen.clear()
for subpath in [
'', '/*/rdfs:label', '/*/skos:prefLabel', '/*/skos:altLabel',
'/*/dcterms:title', '/*/foaf:name']:
def indexDocs(self, writer):
t1 = FieldType()
t1.setIndexed(True)
t1.setStored(True)
t1.setTokenized(False)
t1.setIndexOptions(FieldInfo.IndexOptions.DOCS_AND_FREQS)
t2 = FieldType()
t2.setIndexed(True)
t2.setStored(False)
t2.setTokenized(True)
t2.setIndexOptions(FieldInfo.IndexOptions.DOCS_AND_FREQS_AND_POSITIONS)
for folder in os.listdir(rawDir):
for fileName in os.listdir(rawDir+'/'+folder):
if not fileName.endswith('.json'):
continue
print fileName
subject_id = fileName.split('_')[0]
# print rawDict
print 'id:'+subject_id
print 'folder:'+str(folder)
rawPath = rawDir+'/' +folder +'/'+subject_id+'_res.json'
adjPath = adjDir+'/' +folder +'/'+subject_id+'_adj.json'
#tfidfPath = tf_idfDir + '/' + subject_id+'_tfidf.json'
print adjPath
rawFile = open(rawPath,'r')
raw = rawFile.read()
if raw =='':
with open(baseDir+'/'+'err_no_raw_content.txt','a') as err:
err.write(subject_id+'\n')
# if subject_id =='6018943':
# rawFile.seek(0)
rawFile.close()
adjFile = open(adjPath,'r')
adj = adjFile.read()
adjFile.close()
raw = getRidOfBOM(raw)
adj = getRidOfBOM(adj)
if raw != '':
rawDict = json.loads(raw)
adjDict = json.loads(adj)
rawAll = rawDict['summary'] +' '+ rawDict['user_tags'] + ' '+rawDict['comments']
summary_adjs = adjDict['summary_adjs']
comments_adjs = adjDict['comments_adjs']
title = adjDict['title']
rating_average = adjDict['rating_average']
comments_count = adjDict['comments_count']
doc = Document()
doc.add(Field("folder",folder,t1))
doc.add(Field("title", title, t1))
doc.add(Field("subject_id", subject_id, t1))
doc.add(IntField("comments_count", comments_count, Field.Store.YES))
doc.add(FloatField("rating_average", rating_average, Field.Store.YES))
if len(summary_adjs)>0:
print summary_adjs
exit()
doc.add(Field("summary_adjs", summary_adjs, t2))
#if len(comments_adjs)>0:
doc.add(Field("comments_adjs", comments_adjs, t2))
writer.addDocument(doc)
def indexDocs(self, root, writer):
for root, dirnames, filenames in os.walk(root):
for filename in filenames:
print "adding", filename
doc_parser = HTMLDocumentParser()
try:
path = os.path.join(root, filename)
file = open(path)
contents = unicode(file.read(), 'iso-8859-1')
doc_parser.feed(contents)
contents = doc_parser.contents
html_doc = HTMLDocument(contents)
flag = False
if flag:
print '=============='
print 'Title: ' + html_doc.title
print 'Description: ' + html_doc.description
print 'Month: ' + html_doc.month
print 'Year: ' + html_doc.year
print 'Authors: ' + str(html_doc.authors)
print 'Keywords: ' + str(html_doc.keywords)
print 'Timestamp: ' + str(html_doc.timestamp)
print ' '
file.close()
doc = Document()
field_filename = FieldType()
field_filename.setIndexed(True)
field_filename.setStored(True)
field_filename.setTokenized(False)
field_filename.setIndexOptions(FieldInfo.IndexOptions.DOCS_AND_FREQS)
doc.add(Field("filename", filename.replace('.html', ''), field_filename))
field_path = FieldType()
field_path.setIndexed(True)
field_path.setStored(True)
field_path.setTokenized(True)
field_path.setIndexOptions(FieldInfo.IndexOptions.DOCS_AND_FREQS_AND_POSITIONS)
doc.add(Field("path", root, field_path))
field_title = FieldType()
field_title.setIndexed(True)
field_title.setStored(True)
field_title.setTokenized(True)
field_title.setIndexOptions(FieldInfo.IndexOptions.DOCS_AND_FREQS_AND_POSITIONS)
doc.add(Field("title", html_doc.title, field_title))
field_description = FieldType()
if html_doc.has_description():
field_description.setIndexed(True)
else:
field_description.setIndexed(True)
field_description.setStored(True)
field_description.setTokenized(True)
field_description.setIndexOptions(FieldInfo.IndexOptions.DOCS_AND_FREQS_AND_POSITIONS)
doc.add(Field("description", html_doc.description, field_description))
field_month = FieldType()
field_month.setIndexed(True)
field_month.setStored(True)
field_month.setTokenized(False)
field_month.setIndexOptions(FieldInfo.IndexOptions.DOCS_AND_FREQS)
doc.add(Field("month", html_doc.month, field_month))
field_year = FieldType()
field_year.setIndexed(True)
field_year.setStored(True)
field_year.setTokenized(False)
field_year.setIndexOptions(FieldInfo.IndexOptions.DOCS_AND_FREQS)
doc.add(Field("year", html_doc.year, field_year))
if html_doc.has_authors():
field_author = FieldType()
field_author.setIndexed(True)
field_author.setStored(True)
field_author.setTokenized(True)
field_author.setIndexOptions(FieldInfo.IndexOptions.DOCS_AND_FREQS)
for author in html_doc.authors:
doc.add(Field("author", author, field_author))
if html_doc.has_keywords():
field_keyword = FieldType()
field_keyword.setIndexed(True)
field_keyword.setStored(True)
field_keyword.setTokenized(True)
field_keyword.setIndexOptions(FieldInfo.IndexOptions.DOCS_AND_FREQS)
for keyword in html_doc.keywords:
doc.add(Field("keyword", keyword, field_keyword))
field_timestamp = FieldType()
field_timestamp.setIndexed(False)
field_timestamp.setStored(True)
field_timestamp.setTokenized(False)
field_timestamp.setIndexOptions(FieldInfo.IndexOptions.DOCS_AND_FREQS)
doc.add(Field("timestamp", html_doc.timestamp, field_timestamp))
if len(contents) > 0:
field_source = FieldType()
field_source.setIndexed(True)
field_source.setStored(True)
field_source.setTokenized(True)
field_source.setIndexOptions(FieldInfo.IndexOptions.DOCS_AND_FREQS_AND_POSITIONS)
doc.add(Field("contents", contents, field_source))
else:
print "warning: no content in %s" % filename
writer.addDocument(doc)
except Exception, e:
print "Failed in indexDocs:", e
