So I believe I’ve found a hack to working with Devonthink and Google Drive, that several folks have been interested in.
You can’t actually sync Devonthink to Google Drive itself because of many reasons that have been covered here previously.
HOWEVER. What I’ve been doing recently is I’ve been syncing Google Drive to my Mac, using the Google Drive Desktop application.
Then I use Devonthink to Index that folder on my Mac. Anything I throw into Devonthink, because it just organizes it into that folder system, it gets pushed to the Google Drive cloud, and vice versa, anything I push into Google Drive, shows up in Devonthink.
I’ve been extremely happy with the results, because I get the raw files showing up in Google Drive, but able to very quickly search/index/OCR through Devonthink on my Mac, which is far superior to Google Drive’s search system.
Anyone else try this or have suggestions to improve security or compatibility?
Not only Google Drive. Currently I have exactly that same configuration with Synology Drive. What is more: not all files are locally present (this requires to touch some non-recommended DT hidden options, and you are your own disaster-prone hero).
As with all Google dealings you have to be wary of when they change things without warning or announcements.
Sometime in the last 3 years they changed GDrive from being a folder on your main file system to being a separately mounted partition. That change at least was somewhat signalled before they did it.
It also works relatively cleanly if you’re a one person shop. Things get complicated with shared folders and corporate accounts and how much those people have ownership of files you’ve synced to your system.
Then there’s an issue if you need to reinstall an OS, does GDrive remember what you had stored locally or do you have to download everything again and then hope DT remembers where everything is.
As @rfog points out, you become your own disaster prone hero of maintaining your files and Google is an mischievous imp threatening to change things at their whim.
Those are the caveats but when it works its great!
If I can add one caveat I found when Indexing a OneDrive folder. Devonthink needs all the files locally (i.e. Not spirited off to the cloud). When OneDrive removed a file to the cloud, Devonthink would cause it to be re-downloaded so it was present locally. This slowed Devonthink as it was in a constant ‘in the cloud/on disk’ battle withOneDrive. This means your can’t save space by spiriting things off to the cloud. They get pulled back. The OneDrive option for the folder being indexed has to be ‘always present on disk’ (or whatever they call it).
Indexing requires files to be accessible through file system calls. Which n turn means that they have to be either local or on a system accessible via a network file system like SMB/CIFS or NFS. Google drive, Onedrive etc do not offer a network file system. So, the files have to be locally available to be indexable.
Conversely, any file that gets “spirited away” from your local disk must be downloaded again to be indexed.
A future release will improve this by e.g. not automatically downloading indexed cloud documents for previewing but on demand instead. This will be optional, of course.
It won’t be a feature of DEVONthink’s synchronization as items in cloud folders are synchronized by the cloud service/app on the Mac, not by DEVONthink. It’s only a setting for previewing but the handling is similar to DEVONthink To Go in this case.
