Extracting a PDF in Coda
Using an OCR pack to read text and images
PDF stands for Portable Document Format. It’s a file format that preserves the layout and formatting of a document no matter what device it’s viewed on. This makes PDFs a great way to share documents in Coda that need to look exactly the same way on every device. PDFs are the dominant file type for contracts, invoices and manuals.
To read PDFs we can use OCR. OCR stands for Optical Character Recognition. It’s a technology that can take a scanned image of a document (like a PDF) and convert it into text that can be edited and searched. This is useful because it allows you to turn a physical document or a scanned image into a digital document that you can easily work with.
In the Coda gallery there is an OCR pack (paid, 5 dollar per maker) that extracts the data from any pdf applying this OCR tech. In this blog we show how we first apply this pack and second how we use Coda AI to isolate relevant data.
The installation of the pack
The installation goes fast and is simple. You go right top to insert, you select Packs and you type (a part of) the name. My rule of thumb is one pack per doc, maybe two to keep the doc comprehensible and thus to avoid unnecessarily complexity, more about this concern here.
First go to the pack section.
Next you type the name of the pack and the pack search engine behind looks for the letters as possible match.
The result is that we have the pack active in your document with 14 days free of charge. Afterwards, 5 Dollar per maker — which is a weird pricing model, it should be on workspace level I believe, but that as a side note.
Practical
You create a table and one column is a file type, here you put per row one pdf file, not two or more: one. You create a second column and here you apply one of the functions the pack offers:
OCR::ScanPDFs([Christiaan Huizer],thisRow.files)This function has two elements, the account and the file to read. The account is related to a key for which you have to contact the pack maker. This is a necessary step to avoid abuse. The OCR tech behind the pack is not free.
Limitation.
The OCR pack allows for max 15 pages. May you need to process more pages at once per PDF, you can reach out to the maker. I did and he was very helpful in understanding my needs and proposed some solutions.
The next step
Once you have the content of your file in the column, you can apply Coda AI to ask for specific details. To be clear, you first have to print the content before you can interrogate the file. It is not possible to ask for feedback without having the file unpacked. That is a pity because it can make the doc heavy. After all, Coda is and remains a web based application and the less data in your table, the faster you work in general. I asked the pack maker and I understand why he cannot offer an all in one solution (prompting the extracted content) directly, it is related with a rather expensive technology.
That said, this OCR pack is a wonderful and powerful contribution permitting Coda users to look into files without opening them. Since the content of the PDF is printed in a column, we can use the powerful searchbar (top left) to find anything. This pack works better than PDF extraction packs that do not use OCR and the main reason is that the OCR tech seems to be able to understand texts which are a bit vague or unclear (light grey letters etc which happen often when you scan a contract).
This pack is also a good tool when you work with all sorts of legal documents and you want to check specific details as input for further action.
You should notice that Coda AI is not yet powerful enough to deliver reliable result in all cases, it mostly does however. You may consider other AI packs to get a second opinion. I did not test that for the simple reason that my Coda concept centers around simplicity.
Long story short, when you are working with PDFs and you want to gather insights, this is a good way forward. I am not affiliated to this pack or the pack maker, it is just that I liked the solution that I decided writing about it.
My name is Christiaan and blog about Coda. Since the summer of 2023 often (but not only) about how to Coda with AI to support organisations dealing with texts and templates. The latest Coda AI update was on Dec 7, 2023.
Why I focus on Coda AI you can read here: ⤵️
I hope you enjoyed this article. If you have questions feel free to reach out. Though this article is for free, my work (including advice) won’t be, but there is always room for a chat to see what can be done. You find my (for free) contributions to the Coda Community and on Twitter.
Coda comes with a set of building blocks ー like pages for infinite depth, tables that talk to each other, and buttons that take action inside or outside your doc ーso anyone can make a doc as powerful as an app (source).
Not to forget: the Coda Community provides great insights for free once you add a sample doc.
