def TranslateCaptions(self, translationContext, terminology_names=[]):
marker = "<span>"
sourceLanguageCode = translationContext["sourceLang"]
targetLanguageCodes = translationContext["targetLangList"]
translate_role = translationContext["roleArn"]
bucket = translationContext["bucket"]
inputPath = translationContext["inputLocation"]
outputPath = translationContext["outputlocation"]
jobPrefix = translationContext["jobPrefix"]
try:
translate_client = AwsHelper().getClient('translate')
targetLanguageCode = targetLanguageCodes[0]
self.logger.debug(
"Starting translation to {}".format(targetLanguageCode))
singletonTargetList = []
singletonTargetList.append(targetLanguageCode)
millis = int(round(time.time() * 1000))
job_name = jobPrefix + str(millis)
self.logger.debug("JobName: {}".format(job_name))
terminology_name = []
if len(terminology_names) > 0:
for item in terminology_names:
if targetLanguageCode in item['TargetLanguageCodes']:
terminology_name.append(item['Name'])
break
if len(terminology_name) == 0:
self.logger.debug(int("No custom terminology specified."))
else:
self.logger.debug(
"Using custom terminology {}".format(terminology_name))
# Save the delimited transcript text to S3
response = translate_client.start_text_translation_job(
JobName=job_name,
InputDataConfig={
'S3Uri': "s3://{}/{}".format(bucket, inputPath),
'ContentType': "text/html"
},
OutputDataConfig={
'S3Uri': "s3://{}/{}".format(bucket, outputPath)
},
DataAccessRoleArn=translate_role,
SourceLanguageCode=sourceLanguageCode,
TargetLanguageCodes=singletonTargetList,
TerminologyNames=terminology_name)
jobinfo = {
"JobId": response["JobId"],
"TargetLanguageCode": targetLanguageCode
}
return jobinfo
except Exception as e:
self.logger.error(e)
raise e
def startTranslationJob(bucketName, sourceCode, destCode, access_role):
translate = AwsHelper().getClient('translate')
try:
millis = int(round(time.time() * 1000))
response = translate.start_text_translation_job(
JobName="TranslateJob-json-{}".format(millis),
InputDataConfig={
'S3Uri': "s3://{}/xmlin/".format(bucketName),
'ContentType': 'text/html'
},
OutputDataConfig={'S3Uri': "s3://{}/xmlout/".format(bucketName)},
DataAccessRoleArn=access_role,
SourceLanguageCode=sourceCode,
TargetLanguageCodes=[destCode])
print(response["JobId"])
except ClientError as e:
logger.error("An error occured starting the Translate Batch Job: %s" %
e)