def test_big_db(self):
from tts.db import DB
from tts.synthesizer import SpeechSynthesizer
from tts.srv import SynthesizerRequest
import uuid
db = DB()
speech_synthesizer = SpeechSynthesizer(engine='DUMMY',
max_cache_bytes=401)
speech_synthesizer.engine.set_file_sizes(100)
for i in range(20):
request = SynthesizerRequest(text=uuid.uuid4().hex, metadata={})
response = speech_synthesizer._node_request_handler(request)
self.assertEqual(db.get_num_files(), 4)
speech_synthesizer = SpeechSynthesizer(engine='DUMMY',
max_cache_bytes=40001)
speech_synthesizer.engine.set_file_sizes(1000)
for i in range(80):
request = SynthesizerRequest(text=uuid.uuid4().hex, metadata={})
response = speech_synthesizer._node_request_handler(request)
self.assertEqual(db.get_num_files(), 40)
def test_lost_file(self):
from tts.db import DB
from tts.synthesizer import SpeechSynthesizer
from tts.srv import SynthesizerRequest
import uuid
import os
import json
db = DB()
init_num_files = db.get_num_files()
req_text = uuid.uuid4().hex
speech_synthesizer = SpeechSynthesizer(engine='DUMMY')
request = SynthesizerRequest(text=req_text, metadata={})
response = speech_synthesizer._node_request_handler(request)
res_dict = json.loads(response.result)
audio_file1 = res_dict['Audio File']
self.assertEqual(db.get_num_files(), init_num_files + 1)
os.remove(audio_file1)
request = SynthesizerRequest(text=req_text, metadata={})
response = speech_synthesizer._node_request_handler(request)
res_dict = json.loads(response.result)
audio_file2 = res_dict['Audio File']
self.assertEqual(db.get_num_files(), init_num_files + 1)
self.assertEqual(audio_file1, audio_file2)
self.assertTrue(os.path.exists(audio_file2))
def test_multiple_novel(self):
from tts.db import DB
from tts.synthesizer import SpeechSynthesizer
from tts.srv import SynthesizerRequest
import uuid
db = DB()
init_num_files = db.get_num_files()
for i in range(4):
speech_synthesizer = SpeechSynthesizer(engine='DUMMY')
request = SynthesizerRequest(text=uuid.uuid4().hex, metadata={})
response = speech_synthesizer._node_request_handler(request)
self.assertEqual(db.get_num_files(), init_num_files + i + 1)
def test_no_connection_novel(self):
from tts.db import DB
from tts.synthesizer import SpeechSynthesizer
from tts.srv import SynthesizerRequest
import uuid
db = DB()
init_num_files = db.get_num_files()
speech_synthesizer = SpeechSynthesizer(engine='DUMMY')
speech_synthesizer.engine.set_connection(False)
request = SynthesizerRequest(text=uuid.uuid4().hex, metadata={})
response = speech_synthesizer._node_request_handler(request)
self.assertEqual(db.get_num_files(), init_num_files)
def test_file_cleanup_priority(self):
from tts.db import DB
from tts.synthesizer import SpeechSynthesizer
from tts.srv import SynthesizerRequest
import uuid
import json
import os
db = DB()
speech_synthesizer = SpeechSynthesizer(engine='DUMMY',
max_cache_bytes=401)
speech_synthesizer.engine.set_file_sizes(100)
special_text = uuid.uuid4().hex
request = SynthesizerRequest(text=special_text, metadata={})
response = speech_synthesizer._node_request_handler(request)
res_dict = json.loads(response.result)
audio_file1 = res_dict['Audio File']
self.assertTrue(os.path.exists(audio_file1))
special_text2 = uuid.uuid4().hex
request = SynthesizerRequest(text=special_text2, metadata={})
response = speech_synthesizer._node_request_handler(request)
res_dict = json.loads(response.result)
audio_file2 = res_dict['Audio File']
self.assertTrue(os.path.exists(audio_file2))
for z in range(2):
for i in range(2):
request = SynthesizerRequest(text=uuid.uuid4().hex,
metadata={})
response = speech_synthesizer._node_request_handler(request)
request = SynthesizerRequest(text=special_text, metadata={})
response = speech_synthesizer._node_request_handler(request)
res_dict = json.loads(response.result)
audio_file1 = res_dict['Audio File']
self.assertFalse(os.path.exists(audio_file2))
self.assertTrue(os.path.exists(audio_file1))
def _call_engine(self, **kw):
"""Call engine to do the job.
If no output path is found from input, the audio
file will be put into /tmp and the file name will have
a prefix of the md5 hash of the text. If a filename is
not given, the utterance is added to the cache. If a
filename is specified, then we will assume that the
file is being managed by the user and it will not
be added to the cache.
:param kw: what AmazonPolly needs to synthesize
:return: response from AmazonPolly
"""
if 'output_path' not in kw:
tmp_filename = hashlib.md5(json.dumps(kw,
sort_keys=True)).hexdigest()
tmp_filepath = os.path.join(os.sep, 'tmp',
'voice_{}'.format(tmp_filename))
kw['output_path'] = os.path.abspath(tmp_filepath)
rospy.loginfo('managing file with name: {}'.format(tmp_filename))
# because the hash will include information about any file ending choices, we only
# need to look at the hash itself.
db = DB()
db_search_result = db.ex(
'SELECT file, audio_type FROM cache WHERE hash=?',
tmp_filename).fetchone()
current_time = time.time()
file_found = False
if db_search_result: # then there is data
# check if the file exists, if not, remove from db
# TODO: add a test that deletes a file without telling the db and tries to synthesize it
if os.path.exists(db_search_result['file']):
file_found = True
db.ex('update cache set last_accessed=? where hash=?',
current_time, tmp_filename)
synth_result = PollyResponse(
json.dumps({
'Audio File': db_search_result['file'],
'Audio Type': db_search_result['audio_type'],
'Amazon Polly Response Metadata': ''
}))
rospy.loginfo('audio file was already cached at: %s',
db_search_result['file'])
else:
rospy.logwarn(
'A file in the database did not exist on the disk, removing from db'
)
db.remove_file(db_search_result['file'])
if not file_found: # havent cached this yet
rospy.loginfo('Caching file')
synth_result = self.engine(**kw)
res_dict = json.loads(synth_result.result)
if 'Exception' not in res_dict:
file_name = res_dict['Audio File']
if file_name:
file_size = os.path.getsize(file_name)
db.ex(
'''insert into cache(
hash, file, audio_type, last_accessed,size)
values (?,?,?,?,?)''', tmp_filename, file_name,
res_dict['Audio Type'], current_time, file_size)
rospy.loginfo(
'generated new file, saved to %s and cached',
file_name)
# make sure the cache hasn't grown too big
while db.get_size(
) > self.max_cache_bytes and db.get_num_files() > 1:
remove_res = db.ex(
'select file, min(last_accessed), size from cache'
).fetchone()
db.remove_file(remove_res['file'])
rospy.loginfo(
'removing %s to maintain cache size, new size: %i',
remove_res['file'], db.get_size())
else:
synth_result = self.engine(**kw)
return synth_result