Importing from Evernote

DEVONthink Tips
1

If I import a number of files form evernote that are image based (say 1,000 images of documents, receipts, etc), will the OCR decoding be imported as well? If not is there a way to batch interpret them?

For example, after the import if I do a search in Evernote for text in the image, it is found (even though not in title, etc), but I cannot seem to do same search in Devonthink Pro Office.

If can’t batch, going forward when I import images will they automatically be OCR processed, or do I need to do something?

thank you

2

Evernote’s OCR isn’t supported. The only workaround is to store the images in the database and to perform OCR on your own (see Data > Convert > to searchable PDF) but this requires DEVONthink Pro Office.

3

Just to add quickly: Evernote isn’t doing OCR in the way that we normally think of it. There’s a brief explainer about the implications here:
discussion.evernote.com/topic/5 … ent=301397

Evernote is optically recognizing text, but does something different with this than conventional OCR. Conventional OCR usually comes up with a single text character corresponding to the image of the character, and maps that text to the image. Sometimes the text is wrong (a poor quality image, different language in a single document, bad OCR engine, etc), but hopefully most of the time it’s right. This is one reason why handwriting is typically not recognized by OCR engines: while typed characters can be relatively unambiguous, a handwritten character is very ambiguous at the best of times. Is it a cursive L or a cursive capital I?

In contrast, what Evernote does is take the image of text (or handwriting) and develop a tree of candidate characters and words. There is no “canonical” text layer behind the image of text, there’s a large number of possible text characters and words. This is why in Evernote, you cannot select text in an OCRd image. What would you be selecting? If you selected an image of a character, the text layer below would actually be a bunch of characters with probabilities attached. How would Evernote know what you wanted to select?

This does, however, work for search. If you hand write the word “Flag”, there’s a strong chance searching for “flag” and “flog” could return the same document (depending on the quality of the scan and the writing). This is somewhat of a necessity when you want to provide searchable hand writing or images of text taken in less-than-perfect conditions (e.g., a wine bottle label or a projection of text at a conference).

The downside to this is that there’s really no effective way to embed this information into a file in a non-proprietary way. As such, all of the text information is stuck in Evernote, since PDF files do not support embedding multiple text characters with attached probabilities behind the image layer.

So all of this is to say: Evernote uses a very proprietary (and rather excellent) means of OCRing images of text that is proprietary to Evernote and is not transportable. DTPO cannot use this OCR information locked into Evernote. DTPO can OCR PDFs (and will actually embed the text information), but it cannot OCR images or handwriting. You will need to convert images (e.g., jpegs) of typed text to PDFs (Preview.app can do this on the Mac) and then have DTPO OCR the resulting PDF.

For handwriting… I’m still looking for a solution but outside of proprietary options such as those used in Evernote, there’s really not much!

1 Like
4

This is very helpful. Thank you
Seems might make sense for me to use both – and just figure out a good way to keep them in sync

5

Personally I wouldn’t even attempt to “keep them in sync”. That, in my opinion (and experience) will lead to frustration and wasted time.

If there is some type of content you prefer/need Evernote’s powers to deal with, then that should live in Evernote and only Evernote. Stuff that doesn’t need Evernote’s capabilities could be in DTPO.

I would strongly recommend NOT trying to duplicate content between the two applications. This will very like be error prone and really yield little benefit (and could actually cause issues like data loss if you mess up an import).

While I was moving from Evernote to DTPO several years ago, I moved all of my personal and work documents out of Evernote to DTPO databases. However there was, for a time, some amount of information that I left in Evernote because it was reliant on some Evernote features, namely information in shared notebooks that I used with my partner to deal with our collective household concerns. Evernote stuff lived in Evernote and only Evernote. DTPO stuff lived in DTPO and only DTPO (well, not only DTPO since much of it was indexed and actually lived in Finder).

This means that there was never any question about where new information should go, or where an existing document could be found. If it concerned the household, it goes in/could be found in Evernote. If it concerned my personal affairs, it was in/should be deposited in DTPO.

Eventually I moved my household affairs into shared Dropbox folders, eliminating Evernote completely and allowing me to index those folders in DTPO.

But at no time, ever, did any information live in both Evernote and DTPO.

If your primary concern is around searching the contents of photographs of (handwritten) text, then that content may be best left in Evernote.

One thing you should know, if you aren’t already aware of it:

You can always use Item Links and Note Links in DTPO and Evernote, respectively, to make reference to items in one application or the other. For example, of you keep hand-written notes about an academic article in Evernote and have your academic article library stored in DTPO, you can always paste the Evernote link into the Information pane for the PDF, and the DTPO item link in the body of the Evernote note containing your hand written text.

Ultimately you should do what’s best for you, but I’d strongly caution against trying to do any semblance of “syncing” between DTPO and Evernote!

6

Extremely helpful, I am new to this, thank you.
So I am thinking will:

  1. separate DTPO and Evernote as you suggest
  2. for items in Finder folders and backed up with dropbox, rather in import use indexing. When I sync across other Macs (that have identical folders due to dropbox), the index will also sync, and I update indexed items after each use.

Does that make sense? I wasn’t sure about indexing and dropbox, but didn’t seem to make sense to import files to DTPO since they are huge, and already backup on dropbox.

thanks again

7
  1. Probably a good plan :smiley:
  2. Indexing is definitely favourable in circumstances where you want to have access to the documents outside of DTPO. For example, if you want to have them available via dropbox on computers without DTPO, or from Dropbox’s web interface, or from a mobile device you do not have DEVONthink To Go installed.

I have a few apprehensions about your proposed setup though. As I understand it you have (at least) 2 Macs with DTPO. It sounds like you aren’t using the sync feature in DTPO, so my mental model of your setup is this (Note that Folders are called “Groups” in DTPO, but I’ll use Folder here for the sake of being generic):


Mac A
  Database A
      Folder A [indexed, stored in Dropbox]
      Folder B [indexed, stored in Dropbox]
      ...
      Folder X [indexed, stored in Dropbox]

Mac B
  Database B
      Folder A [indexed, stored in Dropbox]
      Folder B [indexed, stored in Dropbox]
      ...
      Folder X [indexed, stored in Dropbox]

Mac A and Mac B are independent and the DTPO databases do not communicate with each other.

In theory, files added/removed/changed in the folders that correspond to Folder A (or B…X) will sync across dropbox and when the Database updates indexed files, the modifications/additions/removals will be reflected in DTPO.

Where I think you might run into difficult is if you add a new folder to Dropbox that is not in an already-indexed folder, lets say it’s Folder Z, it won’t appear in DTPO. You’ll have to instruct DTPO to index that newly-created folder on both Mac A and Mac B.

Similar situation arises if you create a new document in DTPO (rather than creating it directly in the folder in Finder) – unless you say “Move to external folder”, that newly-created file will not show up in the Folder in Finder and won’t sync.

You may also find that some DTPO-specific metadata (e.g., labels, tags, etc) will not survive the sync process (though much of it will, since much of that metadata is industry-standard stuff fully supported by finder).

What I think you might want to try is to use the Sync feature in DTPO and uncheck the “Upload indexed items” setting (which is on by default). This means that rather than two macs with two separate and independent databases that cannot communicate, the database will actually sync between your Macs. Any documents you create inside DTPO will sync across your Macs, metadata for indexed files will sync, and if you create a Folder Z outside of an already-indexed folder, you’ve only got to add it to one database and it will sync across to your other devices.
(see Sync Via Sync Store tutorial)

You’d end up with a database that is synced across devices, which would look like this:


Mac A
  Database A [Synced using one of DTPO's sync options]
      Folder A [indexed, stored in Dropbox]
      Folder B [indexed, stored in Dropbox]
      ...
      Folder X [indexed, stored in Dropbox]

Mac B
  Database A [Synced using one of DTPO's sync options]
      Folder A [indexed, stored in Dropbox]
      Folder B [indexed, stored in Dropbox]
      ...
      Folder X [indexed, stored in Dropbox]

Note that you are working out a single database that is synced between devices (Database A) rather than two independent database on two computers.

If you don’t need access to your dropbox-synced files outside of DTPO, then you can probably just import them to DTPO and use DTPO’s sync to make them accessible on multiple Macs and iOS devices (Using DEVONthink To Go), which is a bit more straightforward than dealing with syncing AND indexed files (I do a mix of both indexing and importing across several databases on 2 macs and 2 iOS devices).

Before you do any of this, make sure you have a good, reliable, backup system in place (other than Dropbox, which is not a backup service), so you are not left with data-loss should something go sideways.

8

Thanks so much, very helpful indeed.

9

No problem. All this kind of information management stuff requires experimentation and trial-and-error to see what works best for you. DTPO is a blessing and curse since it’s so capable, there’s a lot of experimentation to do!