How to upload files to Google Drive from GAE Python app

Other Questions Google docs
1

I’m working on a Python application running on Google App Engine and I need to upload documents to Google Drive. I’m struggling with the gdata library and how to handle file uploads properly.

Here’s what I have so far:

import gdata.docs.client

docs_client = gdata.docs.client.DocsClient()
docs_client.ClientLogin('username', 'password')

file_source = gdata.MediaSource(# how to handle this part?)
docs_client.Upload(media_source=file_source, title='my_document')

The main issue is that gdata.MediaSource() expects a file path like '/path/to/file.pdf', but in GAE I don’t have access to a traditional file system. My files are stored in the Blobstore or I can access them via direct URLs.

Is there a way to create a MediaSource object from blob data or a URL instead of a file path? I’ve seen some solutions for Java using byte arrays but I can’t find anything similar for Python.

2

Skip the file_path and use file_contents instead when creating MediaSource. For Blobstore files, grab the data with blobstore.BlobReader(blob_key).read() and pass it straight to MediaSource. URL files? Use urlfetch to pull the content and do the same thing. Hit this same problem last year - GAE’s urlfetch works great for this, just set decent timeout values since docs can be huge. Content type usually gets detected automatically, but I’d specify it anyway to avoid upload headaches.

3

try using BytesIO with your blob data. like this: MediaSource(file_contents=BytesIO(your_blob_data), content_type='application/pdf'). should work well on GAE if you get the content type right.

4

Had the same problem migrating from local filesystem to GAE. The trick is that MediaSource takes multiple parameters, not just file_path. You can pass blob content directly: MediaSource(file_contents=blob_data, content_length=len(blob_data), content_type='your/mime-type'). With Blobstore, grab your blob using the key and read it into memory first. For URL files, use urlfetch.fetch() - response.content gives you the raw bytes. Watch out for large files though since GAE has memory limits. Anything over a few MB should use chunked uploads with the resumable upload methods.