I’m trying out DTPO, and placed a lot of PDF books into my database. What I am finding is that the relevant suggestions aren’t granular enough; I want help identifying sections of different pdf’s that are related. Or at the very least, suggestions of relevant annotations within other pdf documents, or suggestions of relevant pages of other documents.
Ultimately, I am looking for a way to intake sources of information, identify and extract key pieces of information from those sources, then compare those notes between multiple sources, be able to easily go back to the original sources and search for and extract new information, and be able to refer to all of these notes in one view while writing up new documents.
Is there some key feature of DTPO I’m overlooking that would help me out here?
No. The See Also is not going to identify sections of a document. You would need separate documents for this to work as expected.
So how can I split apart documents for DT to analyze?
On the page you want to start a split from, you can choose Edit > Split Document (or right-click the page in the PDF sidebar and choose Split Document).
You can also drag a page from the PDF out to another location in a database. However, if you do this with a single page, I strongly suggest you edit the title as it will be generated from the contents of the first paragraph.
It’s not uncommon for documents to contain information about multiple topics, one or more of which may be of significant interest to the user.
One approach to helping the AI assistants “see” a topic could be to split such documents into segments that contain the topics. That would require grunt work. As I don’t have a research assistant to whom I could assign that grunt work, I don’t choose this approach. In any case, I don’t want to vandalize my references by chopping them up.
Instead, I prefer creating a rich text note that summarizes a topic covered in a reference and includes a link or an exact search string that indicates the topic’s location in the referenced document. In the case of paginated PDFs, paste the appropriate Page Link into the rich text note. For other non-paginated documents, copy a text string of perhaps 5 or 6 words that’s in or near the topic location, and paste the string, enclosed within quotation marks, into the rich text note. That text string, including the quote marks, can be used in a phrase search. It’s easy to find a string that will be pretty unique, even in my database that contains more than 30,000 documents.