class pdf_parser:
def __init__(self, tika_jar_path):
self.tika_client = TikaApp(file_jar=tika_jar_path)
def parse(self, doc_path, file=False):
if file:
encoded = base64.b64encode(doc_path)
content = json.loads(
self.tika_client.extract_all_content(payload=encoded))
else:
content = json.loads(
self.tika_client.extract_all_content(path=doc_path))
content_string = re.sub(r'\n(?![\n])', r'',
content[0]['X-TIKA:content'])
content_string = re.sub(r'(\n)(\n+)', r'\1', content_string)
date_string = content[0].get('Last-Modified') \
or content[0].get('Last-Save-Date') \
or content[0].get('Creation-Date') \
or datetime.now().strftime('%Y-%m-%dT%H:%M:%SZ')
if date_string:
date_string = str(dateutil.parser.parse(date_string).date())
df = pd.DataFrame(columns=['date', 'content'])
df.loc[0] = [date_string, content_string]
return df
def main():
args = get_args()
tika = TikaApp(args.jar or os.environ.get("TIKA_APP_JAR", None))
parameters = {
"path": args.file,
"payload": args.payload,
"objectInput": sys.stdin if args.stdin else None
}
try:
if args.detect:
print(tika.detect_content_type(**parameters))
if args.text:
print(tika.extract_only_content(**parameters))
if args.language:
print(tika.detect_language(**parameters))
if args.all:
parameters["pretty_print"] = True
print(tika.extract_all_content(**parameters))
if args.metadata:
parameters["pretty_print"] = True
print(tika.extract_only_metadata(**parameters))
except IOError:
pass
def tika(conf, attachments):
"""This method updates the attachments results
with the Tika reports.
Args:
attachments (list): all attachments of email
conf (dict): conf of this post processor
Returns:
This method updates the attachments list given
"""
if conf["enabled"]:
from tikapp import TikaApp
tika = TikaApp(file_jar=conf["path_jar"],
memory_allocation=conf["memory_allocation"])
for a in attachments:
if not a.get("is_filtered", False):
if a["Content-Type"] in conf["whitelist_cont_types"]:
payload = a["payload"]
if a["content_transfer_encoding"] != "base64":
payload = payload.encode("base64")
# tika-app only gets payload in base64
a["tika"] = tika.extract_all_content(payload=payload,
convert_to_obj=True)
def getTextFile(nameOfPDF, jarPath):
'''
getTextFile("nameOfPDF", path) will take the PDF and
output it as a textfile. Path will be taken and used to
specify the .jar file needed for tika.
Note: Needs the tika-app-1.22.jar and tika-app-python files
and folders in the same working directory
'''
# get the Tika Object from current directory
tika_client = TikaApp(file_jar=join(jarPath, "tika-app-1.22.jar"))
# read the pdf
with open(nameOfPDF) as fin:
# get rid of the .pdf to change to .txt for later
if nameOfPDF[-4:] == ".pdf":
foutName = nameOfPDF[:-4]
else:
# if its not .pdf, then just keep the filename
foutName = nameOfPDF
content = tika_client.extract_all_content(objectInput=fin)
# write the pdf to a text file & add .txt to it
with open(foutName + ".txt", "w", encoding='utf-8',
errors='replace') as fout:
fout.write(content)
return foutName + ".txt"
def main():
args = get_args()
command_line = dict()
if args.jar:
command_line = {"TIKA_APP_JAR": args.jar}
defaults = {"TIKA_APP_JAR": "/opt/tika/tika-app-1.15.jar"}
options = ChainMap(command_line, os.environ, defaults)
tika = TikaApp(options['TIKA_APP_JAR'])
try:
if args.file:
f = args.file
if args.detect:
print(tika.detect_content_type(path=f))
if args.text:
print(tika.extract_only_content(path=f))
if args.language:
print(tika.detect_language(path=f))
if args.all:
print(tika.extract_all_content(path=f, pretty_print=True))
elif args.payload:
p = args.payload
if args.detect:
print(tika.detect_content_type(payload=p))
if args.text:
print(tika.extract_only_content(payload=p))
if args.language:
print(tika.detect_language(payload=p))
if args.all:
print(tika.extract_all_content(payload=p, pretty_print=True))
except IOError:
pass
class TikaReader(object):
def __init__(self, path):
self.tika_client = TikaApp(file_jar=path)
def detect_type(self, doc):
return self.tika_client.detect_content_type(doc)
def detect_language(self, doc):
return self.tika_client.detect_language(doc)
def content(self, doc):
return self.tika_client.extract_all_content(doc)
class TikaAnalysis(object):
def __init__(self,
jar=None,
memory_allocation=None,
valid_content_types=set()):
# Init Tika
self._tika_client = TikaApp(file_jar=jar,
memory_allocation=memory_allocation)
self._jar = jar
self._memory_allocation = memory_allocation
self._valid_content_types = valid_content_types
@property
def jar(self):
return self._jar
@jar.setter
def jar(self, value):
self._jar = value
@property
def memory_allocation(self):
return self._memory_allocation
@memory_allocation.setter
def memory_allocation(self, value):
self._memory_allocation = value
@property
def valid_content_types(self):
return self._valid_content_types
@valid_content_types.setter
def valid_content_types(self, value):
if not isinstance(value, set):
raise InvalidContentTypes("Content types must be a set")
self._valid_content_types = value
def add_meta_data(self, attachment):
"""If content_type in valid_content_types this method
extracts meta data and update attachments input results.
"""
if not isinstance(attachment, dict):
raise InvalidAttachment("Attachment result is not a dict")
# The Apache Tika output of archive contains the contents and metadata
# of all archived files.
if attachment['Content-Type'] in self.valid_content_types:
attachment['tika'] = self._tika_client.extract_all_content(
payload=attachment['payload'], convert_to_obj=True)
def handle_backup_file(self,context, savepkt):
DebugMessage(context, 100, "handle_backup_file called with " + str(savepkt) + "\n");
DebugMessage(context, 100, "fname: " + savepkt.fname + " Type: " + str(savepkt.type) + "\n");
if ( savepkt.type == bFileType['FT_REG'] ):
DebugMessage(context, 100, "regulaer file, do something now...\n");
# configure your Elasticsearch server here:
es = Elasticsearch([{'host': '192.168.17.2', 'port': 9200}])
# configure your TikaApp jar file here:
try:
tika_client = TikaApp(file_jar="/usr/local/bin/tika-app-1.20.jar")
except Exception as ex:
JobMessage(context, bJobMessageType['M_ERROR'], 'Error indexing %s. Tika error: %s' % (savepkt.fname, str(ex)))
return bRCs['bRC_OK'];
# tika_client has several parser options
# Next one is for metadata only:
#result_payload=tika_client.extract_only_metadata(savepkt.fname)
# This one includes file contents as text:
try:
result_payload=tika_client.extract_all_content(savepkt.fname)
except Exception as ex:
JobMessage(context, bJobMessageType['M_ERROR'], 'Error extracting contents from %s. Tika error: %s' % (savepkt.fname, str(ex)))
return bRCs['bRC_OK'];
# result_payload is a list of json-strings. Nested structes like
# tar-files or emails with attachments or inline documents are
# returned as distinct json string.
# The first string [0] contains information for the main file
# TODO: care about nested structures, for now we only takte the first/main file
try:
data = json.loads(result_payload)[0]
except Exception as ex:
JobMessage(context, bJobMessageType['M_ERROR'],
'Error reading json fields delivered by Tika examining file %s. Json error: %s' % (savepkt.fname, str(ex)))
return bRCs['bRC_OK'];
# Tika eventually adds "Unkonwn Tags (id)", with id as increasing number, which
# could lead to exceed the keyword limit in elasticsearch indices, we
# remove those tags
for data_keyword in data.keys():
if data_keyword.startswith ("Unknown tag ("):
del data[data_keyword]
# Tika adds some emptylines at the beginning of content, we strip it here
if 'X-TIKA:content' in data:
data['X-TIKA:content'] = data['X-TIKA:content'].strip()
data['bareos_jobId'] = self.jobId
data['bareos_fdname'] = self.fdname
data['bareos_joblevel'] = unichr(self.level)
data['bareos_directory'] = os.path.dirname(savepkt.fname)
try:
esRes = es.index (index="bareos-test", doc_type='_doc', body=data)
except Exception as ex:
JobMessage(context, bJobMessageType['M_ERROR'], 'Error indexing %s. Elastic error: %s' % (savepkt.fname, str(ex)))
return bRCs['bRC_OK'];
def tika(conf, attachments):
"""This method updates the attachments results
with the Tika reports.
Args:
attachments (list): all attachments of email
conf (dict): conf of this post processor
Returns:
This method updates the attachments list given
"""
if conf["enabled"]:
from tikapp import TikaApp
tika = TikaApp(file_jar=conf["path_jar"],
memory_allocation=conf["memory_allocation"])
wtlist = conf.get("whitelist_content_types", [])
if not wtlist:
log.warning(
"Apache Tika analysis setted, without whitelist content types")
return
for a in attachments:
if not a.get("is_filtered", False):
if a["Content-Type"] in wtlist:
payload = a["payload"]
if a["content_transfer_encoding"] != "base64":
try:
payload = payload.encode("base64")
except UnicodeError:
# content_transfer_encoding': u'x-uuencode'
# it's not binary with strange encoding
continue
# tika-app only gets payload in base64
try:
results = tika.extract_all_content(payload=payload,
convert_to_obj=True)
if results:
a["tika"] = results
except JSONDecodeError:
log.warning(
"JSONDecodeError for {!r} in Tika analysis".format(
a["md5"]))
class TikaReader:
# Iniciador de la clase.
def __init__(self, file_process):
# Cliente Tika que utiliza que carga el fichero jar cliente.
self.tika_client = TikaApp(file_jar="tika-app-1.20.jar")
self.file_process = file_process
# Detector del tipo de contenido MIME.
def detect_document_type(self):
return self.tika_client.detect_content_type(self.file_process)
# Detector de lenguaje utilizado en el documento.
def detect_language(self):
return self.tika_client.detect_language(self.file_process)
# Extractor del contenido completo del documento.
def extract_complete_info(self, value=False):
return self.tika_client.extract_all_content(self.file_process,
convert_to_obj=value)
# Extractor de solo el contenido del documento.
def extract_content_info(self):
return self.tika_client.extract_only_content(self.file_process)
class ProcessJSONTika(object):
def __init__(self, path):
self.tika_client = TikaApp(file_jar=path)
def jsonprocessor(self, doc):
return self.tika_client.extract_all_content(doc,
convert_to_obj=True)[0]
def author(self, doc):
return self.jsonprocessor(doc).get('Author', None)
def creationdate(self, doc):
return self.jsonprocessor(doc).get('Creation-Date', None)
def lastmodified(self, doc):
return self.jsonprocessor(doc).get('Last-Modified', None)
def all_content(self, doc):
return self.jsonprocessor(doc)['X-TIKA:content']
def top_10_words(self, doc):
content = self.all_content(doc)
words = word_tokenize(content)
# stopwords
stopWords = set(stopwords.words('english'))
clean_words = [
word for word in words if word.isalpha() and word not in stopWords
]
words_dic = {}
for i in clean_words:
if i in words_dic.keys():
words_dic[i] += 1
else:
words_dic[i] = 1
return sorted(words_dic.items(),
key=operator.itemgetter(1),
reverse=True)[:10]
def tika_extract_all_content(memory=None):
tika_client = TikaApp(file_jar=TIKA_APP_JAR, memory_allocation=memory)
output = tika_client.extract_all_content(path=test_zip)
return output
def handle_backup_file(self, context, savepkt):
DebugMessage(context, 100,
"handle_backup_file called with " + str(savepkt) + "\n")
DebugMessage(
context, 100,
"fname: " + savepkt.fname + " Type: " + str(savepkt.type) + "\n")
if (savepkt.type == bFileType['FT_REG']):
DebugMessage(context, 100, "regulaer file, do something now...\n")
# configure your Elasticsearch server here:
es = Elasticsearch([{'host': '192.168.17.2', 'port': 9200}])
# configure your TikaApp jar file here:
try:
tika_client = TikaApp(
file_jar="/usr/local/bin/tika-app-1.20.jar")
except Exception as ex:
JobMessage(
context, bJobMessageType['M_ERROR'],
'Error indexing %s. Tika error: %s' %
(savepkt.fname, str(ex)))
return bRCs['bRC_OK']
# tika_client has several parser options
# Next one is for metadata only:
#result_payload=tika_client.extract_only_metadata(savepkt.fname)
# This one includes file contents as text:
try:
result_payload = tika_client.extract_all_content(savepkt.fname)
except Exception as ex:
JobMessage(
context, bJobMessageType['M_ERROR'],
'Error extracting contents from %s. Tika error: %s' %
(savepkt.fname, str(ex)))
return bRCs['bRC_OK']
# result_payload is a list of json-strings. Nested structes like
# tar-files or emails with attachments or inline documents are
# returned as distinct json string.
# The first string [0] contains information for the main file
# TODO: care about nested structures, for now we only takte the first/main file
try:
data = json.loads(result_payload)[0]
except Exception as ex:
JobMessage(
context, bJobMessageType['M_ERROR'],
'Error reading json fields delivered by Tika examining file %s. Json error: %s'
% (savepkt.fname, str(ex)))
return bRCs['bRC_OK']
# Tika eventually adds "Unkonwn Tags (id)", with id as increasing number, which
# could lead to exceed the keyword limit in elasticsearch indices, we
# remove those tags
for data_keyword in data.keys():
if data_keyword.startswith("Unknown tag ("):
del data[data_keyword]
# Tika adds some emptylines at the beginning of content, we strip it here
if 'X-TIKA:content' in data:
data['X-TIKA:content'] = data['X-TIKA:content'].strip()
data['bareos_jobId'] = self.jobId
data['bareos_fdname'] = self.fdname
data['bareos_joblevel'] = unichr(self.level)
data['bareos_directory'] = os.path.dirname(savepkt.fname)
try:
esRes = es.index(index="bareos-test",
doc_type='_doc',
body=data)
except Exception as ex:
JobMessage(
context, bJobMessageType['M_ERROR'],
'Error indexing %s. Elastic error: %s' %
(savepkt.fname, str(ex)))
return bRCs['bRC_OK']
class TikaProcessing(AbstractProcessing):
""" This class processes the output mail attachments to add
Apache Tika analysis.
Args:
jar (string): path of Apache Tika App jar
valid_content_types (list or set): list of contents types to analyze
memory_allocation (string): memory to give to Apache Tika App
"""
def __init__(self, **kwargs):
super(TikaProcessing, self).__init__(**kwargs)
# Init Tika
self._tika_client = TikaApp(file_jar=self.jar,
memory_allocation=self.memory_allocation)
def __getattr__(self, name):
try:
return self._kwargs[name]
except KeyError:
# Default values
if name in ("memory_allocation"):
return None
else:
msg = "'{0}' object has no attribute '{1}'"
raise AttributeError(msg.format(type(self).__name__, name))
def __setattr__(self, name, value):
super(TikaProcessing, self).__setattr__(name, value)
if name == "valid_content_types":
if not isinstance(value, set) and not isinstance(value, list):
raise InvalidContentTypes("Content types must be set or list")
self._kwargs[name] = value
def _check_arguments(self):
"""This method checks if all mandatory arguments are given. """
if 'jar' not in self._kwargs:
msg = "Argument '{0}' not in object '{1}'"
raise MissingArgument(msg.format('jar', type(self).__name__))
if 'valid_content_types' not in self._kwargs:
msg = "Argument '{0}' not in object '{1}'"
raise MissingArgument(
msg.format('valid_content_types',
type(self).__name__))
def process(self, attachment):
"""This method updates the attachment result
with the Tika output.
Args:
attachment (dict): dict with a raw attachment mail
Returns:
This method updates the attachment dict given
"""
super(TikaProcessing, self).process(attachment)
if attachment['Content-Type'] in self.valid_content_types:
attachment['tika'] = self._tika_client.extract_all_content(
payload=attachment['payload'], convert_to_obj=True)
