|
| 1 | +import logging |
| 2 | + |
| 3 | +from apache_beam import Create |
| 4 | +from apache_beam import DoFn |
| 5 | +from apache_beam import Map |
| 6 | +from apache_beam import ParDo |
| 7 | +from apache_beam import Pipeline |
| 8 | +from apache_beam.options.pipeline_options import GoogleCloudOptions |
| 9 | +from apache_beam.options.pipeline_options import PipelineOptions |
| 10 | +from apache_beam.options.pipeline_options import SetupOptions |
| 11 | +from google.cloud import translate |
| 12 | + |
| 13 | + |
| 14 | +class Translate(DoFn): |
| 15 | + def __init__(self, project, source_language_code, target_language_code): |
| 16 | + self._client = None |
| 17 | + self._project = project |
| 18 | + self._source_code = source_language_code |
| 19 | + self._target_code = target_language_code |
| 20 | + |
| 21 | + def setup(self): |
| 22 | + self._client = translate.TranslationServiceClient() |
| 23 | + |
| 24 | + def process(self, text): |
| 25 | + # See https://cloud.google.com/translate/docs/advanced/quickstart |
| 26 | + request = { |
| 27 | + 'parent': f'projects/{self._project}/locations/global', |
| 28 | + 'contents': [text], |
| 29 | + 'mime_type': 'text/plain', |
| 30 | + 'source_language_code': self._source_code, |
| 31 | + 'target_language_code': self._target_code, |
| 32 | + } |
| 33 | + |
| 34 | + response = self._client.translate_text(request) |
| 35 | + |
| 36 | + translated_text = None |
| 37 | + for translation in response.translations: |
| 38 | + translated_text = translation.translated_text |
| 39 | + break |
| 40 | + |
| 41 | + yield text, translated_text |
| 42 | + |
| 43 | + |
| 44 | +def main(): |
| 45 | + options = PipelineOptions() |
| 46 | + options.view_as(SetupOptions).save_main_session = True |
| 47 | + |
| 48 | + project = options.view_as(GoogleCloudOptions).project |
| 49 | + assert project is not None, '"project" is not specified.' |
| 50 | + |
| 51 | + source_code = 'en-US' |
| 52 | + target_code = 'ja' |
| 53 | + texts = ['Hello', 'Thank you', 'Goodbye'] |
| 54 | + |
| 55 | + p = Pipeline(options=options) |
| 56 | + (p |
| 57 | + | 'Texts' >> Create(texts) |
| 58 | + | 'Translate' >> ParDo(Translate(project, source_code, target_code)) |
| 59 | + | 'Print' >> Map(lambda pair: logging.info('%s -> %s', pair[0], pair[1]))) |
| 60 | + |
| 61 | + p.run() |
| 62 | + |
| 63 | + |
| 64 | +if __name__ == '__main__': |
| 65 | + logging.getLogger().setLevel(logging.INFO) |
| 66 | + main() |
0 commit comments