Question

This is a general question about whether it is possible, and if so how, to automate the download of a scribd.com search result document.

Scenario:

I have a Scribd account and find a document I want. I normally I then have to click the download button to start the download.

Any ideas for automating this? I’m using the scribd api and python to automatically extract document IDs based on automated queries, but once I get the doc_id’s I have to physically go to each doc page and click the download button to get the physical txt/pdf file. I want to automate this step as well.

Any Ideas?

Answer 1

Looking at the python-scribd documentation or the scribd API reference, any object that can give you a document ID or website URL can also give you a download URL. Or, if you already have a document ID, you can just call get to get an object that can give you a download URL.

Most likely, you’ve got a Document object, which has this method:

get_download_url(self, doc_type=’original’)

Returns a link that can be used to download a static version of the document.

So, wherever you’re calling get_scribd_url, just call get_download_url.

And then, to download the result, Python has urllib2 (2.x) or urllib.request (3.x) built into the standard library, or you can use requests or any other third-party library instead.

Putting it all together as an example:

# do all the stuff to set up the api_key, get a `User` object, etc.

def is_document_i_want(document):
    return document.author == "Me"

urls = [document.get_download_url() for document in user.all()
        if is_document_i_want(document)]

for url in urls:
    path = urllib.parse.urlparse(url).path
    name = os.path.basename(path)
    u = urllib.request.urlopen(url)
    with open(name, 'w') as f:
        f.write(u.read())
    print('Wrote {} as {}'.format(url, name))

Presumably you’re going to want to use something like user.find instead of user.all. Or, if you’ve already written the code that gets the document IDs and don’t want to change it, you can use user.get with each one.

And if you want to post-filter the results, you probably want to use attributes beyond the basic ones (or you would have just passed them to the query), which means you need to call load on each document before you can access them (so add document.load() at the top of the is_document_i_want function). But really, there’s nothing complicated here.