Database Features

Each time you start an upload, pneumatic creates a table called uploads in a SQLite database in a folder called pneumatic_db in the current directory. The uploads table captures the following local file data plus data from DocumentCloud’s API response:

  • id: The unique document id stored by DocumentCloud. Typically, it’s in the form of a number followed by several words. e.g. 12345-some-document-title.
  • title: The document’s title.
  • file_name: Name of the file submitted for upload.
  • full_path: The full directory path to the file, including the file name.
  • upload_time: Timestamp indicating the time of upload.
  • pages: The number of pages in the document.
  • file_hash: A unique identifier calculated based on the document. Useful for finding identical documents.
  • result: HTTP response code from DocumentCloud. 200 indicates success.
  • canonical_url: The URL to the document displayed in the viewer on DocumentCloud.
  • pdf_url: The URL to the PDF.
  • text_url: The URL to the extracted text of the document.
  • exclude_flag: Flag indicating whether pneumatic excluded the file from being uploaded. For information on why files might be excluded, see the documentation on File Exclusions.
  • exclude_reason: The reason a file was excluded from upload.
  • error_msg: Server error message if upload attempt failed.

Printing the database name and location

After you initiate a DocumentCloudUploader object and begin the upload, you can instruct pneumatic to print the path to and name of the database file:


Viewing database contents

There are number of tools available for viewing and querying the SQLite database’s contents. One example that’s simple to use is the DB Browser for SQLite GUI app, available for Windows and macOS. If you prefer to work from the command line, see the SQLite command line app documentation.

pneumatic also provides a method to dump the database to a CSV file, which you can open with Excel or other applications. To dump the database to CSV in the same script where you initiated an upload client, use the following:


If you’re coming back to an old database, or you forgot to dump the CSV while the object was alive, you can still dump to a CSV by importing a Database() object:

from pneumatic import Database

db = Database()

Updating database contents

The API response to a DocumentCloud upload does not include values for pages or file_hash. These are calculated as part of DocumentCloud’s processing of the document. Once your documents are finished processing, you can use pneumatic to retrieve those values by using the update_processed_files method.

You can use the method in a current session, which will update the database active in the session, or you can pass in the name of any other database file pneumatic created.

To process an older database:

from pneumatic import DocumentCloudUploader

uploader = DocumentCloudUploader('', 'your-password')

To use within the current session:


Note that using this method within a current session may not retrieve all file information if it’s run before files are done processing.