在位于Google云存储的PDF文件上使用textract

Question

我想将textract应用到Google云存储上托管的pdf文件。

我已经用download_as_string()使用并解析了一个txt文件（一切正常）

storage_client = storage.Client()
bucket = storage_client.get_bucket('bucket-for-pdf')
blob = bucket.get_blob('keywords.txt')
keywords_file = blob.download_as_string().decode('utf8')

storage_client = storage.Client()
bucket = storage_client.get_bucket('bucket-for-pdf')
blob = bucket.get_blob(file_name)
file_name = blob.download_to_file(file_name)
return textract.process(file_name, language='eng',
                        encoding='utf-8').decode('utf-8')

上述代码导致错误：

self._stream.write(chunk)
AttributeError: 'str' object has no attribute 'write'

Answer 1

download_to_file方法接受文件对象，而不是文件名。试着做一些类似的事情：

file_name = "/tmp/my-file"
storage_client = storage.Client()
bucket = storage_client.get_bucket('bucket-for-pdf')

with open(file_name, "wb") as file_obj:
    blob.download_to_file(file_obj)

return textract.process(file_name, language='eng', encoding='utf-8').decode('utf-8')