วีดีโอ

Hi there. Thank you for your feedback and comment. Converting handwriting to typed text is a critical feature that I think every user expects to have from an e-ink device such as the Supernote. However, the feature described in this video addresses a different need: allowing users to bring external typed content in their notebooks, by mixing selections with their own handwriting and rearranging as they want. In one of the first videos about that text-to-pen strokes feature, I mentioned the origin of that feature: a pastor wanted to copy and paste verses of the Bible in his prepared sermon, because it would save him time copying by hand those verses. I can see a similar need for some people who want to complement notebooks on a given topic with some extract of typed text material. Such users wouldn't just convert entire textbooks into the Supernote format for reading...They want the ability to copy and paste selected content into their own notes, something the Supernote doesn't offer yet natively.

Thank you! Azure vision, please check back this thread in a few days. I may just send a screencast link

¡Gracias! Espero que te ayude.

Glad I could help! Thank you for the feedback.

Thank you for the feedback. Still some tweaks needed to find the optimal resolution for pictures, and I hope to merge soon to the main branch SN import of hybrid pen strokes / complex pictures downsized to sketch functionality.

@@mujmmb side note: is there any reason this shouldn't work with Python installed on the Supernote itself via Termux or something similar?

Thank you. Yes: a few devs are interested to supplement an existing repo or build a new one and share ideas. Please email me and I'll follow up.

Hi there. Thank you. Happy to follow up, if you could email me at max at mujynya.net, with a brief intro.

Thank you for the feedback!

Thank you for pointing out the error. I confirm that the issue is with newer versions of Python such as 3.13.0 are not compatible with the list of requirements. Users for now should use version 3.12.3

Thank you Marco! I think the reason it’s not working in your example is that I probably filtered out extensions that are not recognized by the Supernote. In other words, all you may need to do is renaming your md file with txt extension. I’d have to check it myself. But I implemented the logic to import to note format for md files and docx files are first converted to md. I’ll confirm.

Yes, the detection is made on the vectors themselves and is very fast. I am then replacing the original vectors by the “ideal” vectors, using the exact same amount of points, because I don’t want to rebuild the entire binary file. What I haven’t replaced yet are the “contours” vectors that are also stored in the binary file. They play a role in the selection. Agree that it would be nice to have the shapes available natively.

That’s correct. I usually present new features in a beta branch that is ahead of the main branch. But this video is a few months old and since then the main and stable version is what should install. What matters is the version #. At this time the main is running version 1.27

Even after a new clean install of Python - thanks to a Reddit user and activation of the virtual environment, which I guess is shown through the prefix (pysnvenv) in my Terminal, I get this error on setting my input folders. The error is: File "/Users/myname/pysn-digest/digest.py", line 7, in <module> import fitz ModuleNotFoundError: No module named 'fitz' So I Googled this error and installed fitz. Now that fitz is installed, I get two errors: 1) chmod: /Users/myname/pysn-digest/platform-tools/adb: No such file or directory 2) an error on line 11 in <module> regarding Pil which says is not installed. But in fact Pillow version 11.0.0 arm is installed as says the Terminal. The same error keeps appearing again. Could you @Max please indicated me at well level is the issue, and how I can fix it?

Hello there. I saw that you started an old fork of the project that is not maintained by me. Please install instead the official branch following the link in the description.

Thank you... It's just a teaser, for no. But you know that, right?

Glad you think so!

Thank you. Yes it would be cool to have some features natively!

Glad you enjoy it!

Thank you, hope it helps!

Thank you for the feedback and kind comments.

I was expecting Boox to be more open. Is there an unofficial file structure around?

Sorry for the late reply. Yes rooting allowed me to learn more about the note format. You may want to take a look at my latest videos. Just to clarify: at this time PySN library runs scripts out of the device. It does edit the .note binary format and I can create notebooks from scratch. Hopefully will post soon 2 new features: basic shapes detection and, more importantly, importing pdf or images to notebook format. That would allow you to import to Supernote your notebooks from RM or Boox as editable pen strokes. Stay tune!

@@mujmmb you're doing fantastic work, thanks!

I can see some advantages to have the way Boox implemented it (besides saving space). Since their solution will come for both text and pictures, I think the odds are that both the text and image will come as different embedded objects in a page.

@@mujmmb lack of shape generation and text boxes was what made me regretfully send my A6X2 back and get a Tab Mini C instead. I didn't even want color; fast B&W is fine! as an experiment I'm trying out a BigMe S6 but WOW the software on it feels 1/4 baked at best.

Thank you for the feedback and I hope it helps!

Glad it was helpful!

Thank you. An application from a commenter here is to create some sort of a reverse text recognition: the input is a text file, and the output is an appended page on a notebook with the text converted into pen strokes of a given characters set (an ASCII table of characters in a notebook page format). So the script reads the text, then for each character retrieves its ASCII code, then pulls the corresponding pens strokes from the characters set and keep appending the pen strokes, before adding the result to a new page. And you guessed it: using keyword settings :) Easier said than done, but I think I have all the information needed.

You're very welcome, thank you for the kind comments.

Pretty incredible work!! I'm excited to try integratimg a supernote to NotePlan workflow using PySN! Also on mac myself, so I really appreciate the specific walkthrough to getting it installed

Thank you! Are you on a Mac? The example given if for windows, requesting double backslashes for paths. Happy to look into your user_settings.json, if you email it to me.

@@mujmmb No, using the latest versions of windows and python (and pip) and the << \\ >> symbol. It has worked before, pdf and other files were copied to another target folder (as defined in the copy-cloud.json file). Then I changed the name of the targetfolder in this copy-cloud.json file and then it stopped working. I changed the target folder name back to original one, but to no avail. The user_settings.json file was the same (no changes made), before and after it stopped working.

@@martinbijlsma8507 The script expects the json to be well formed i.e. if it doesn't adhere to json format, it will raise an error. You may use an online json validator like jsonlint.com/ to detect what is wrong. Also please note as a general rule that the code will be case sensitive, even if Windows isn't. You can also email me the jsons for me to take a look, thank you.

@@mujmmb Thanks for your response. I've send you an e-mail, maybe you could have a look.

@@mujmmb Ok, thanks for your help, everything works now, exactly as shown in the video, GREAT!

Will you please share what is not working? Keep in mind that this is an “old” video. See the installation process on a newer video: th-cam.com/video/fKnpdr5G1qU/w-d-xo.html PySN expanded the features to bulk export all notes and one important change is that the “digest” feature in PySN now requires that the path of the documents to digest be in a folder containing a keyword. By default this is “digest_in” (see the “digest” key in the settings). If you don’t want to have a special path for docs to digest, you could for example set the digest key in the settings to “Document”. This would ensure that all pdfs under the “Document” root folder be processed.

@@mujmmb Yeah, sorry for that, this was a very useless post of mine... Believe it or not, last night I did get the pdf-digest to work after all. Now I am back to the real problem, trying to make the ks-keywords work. But that is another topic. In spite of the troubles I am having, it remains great fun playing around with python and Supernote

@@martinbijlsma8507 , No worries. I replied to you on the other video: feel free to email me the user_settings.json and perhaps a copy and paste of messages showing in the terminal. I am happy to fix bugs.

Thank you!

Thank you!

ˈpaɪ-sən/ (Python for SN, pun intended :) ) gitlab.com/mmujynya/pysn-digest

@@mujmmb Gotcha! I was sitting there scratching my head. I'm tracking now! That was pretty clever! Am I correct in assuming that you also worked in a double entendre with a Lean Six Sigma reference there with the "pai-sən?" (The Poisson Distribution).

@@wheelch0ck Glad you liked the name! The pun was mainly intended as a play on 'Python' and 'SN' for Supernote, leading to 'paɪ-sən.' I didn't initially have the Poisson distribution or Six Sigma in mind, but it's interesting that you made that connection! Thanks for the thoughtful interpretation-it's always fun to see how different layers of meaning can emerge

Glad it helped!

Thank you for the kind comments. Adding keyword settings features is relatively easy so feedback from users is appreciated. See for instance in this beta version a relatively more complex keyword setting: th-cam.com/video/vSZ8fj2UhmU/w-d-xo.htmlsi=mZggxekM3QgbGeJD

Trust me, I almost cried when I broke that ceramic nib. Glad your DYI worked for you. Lamy is now my preferred pen and in retrospect, it's the original wight that I hate.

Hello there, thank you for your comment and application suggestion. For now, I can modify existing strokes or simulate pen strokes. Simulating pen strokes using a template, in your case a picture of verses from the Bible, is very slow and would require to have the Supernote open on a blank page. But technically speaking, you are right: even without having full transparency on the Supernote format, one could compile for instance the pen strokes for a given font of all the letters of the alphabet. The script would then need to combine these letters into words, while ensuring that they fit within a page frame. I anticipate this would be much faster than simulation. It's a very interesting idea and I may check the feasibility. What font do you think would be good?

​@@mujmmb I 'assume' Supernote actually records something at the moment of writing, whereas I'm asking it to in someway _'trace typed text'_ and then record it was hand-written text. The difference them might be better understood if I... A) Was able to brilliantly copy, to the pixel a verse of typed text then... B) Imported the typed text I had copied so magnificently from. Supernote would have recorded the pen-strokes of [A], and I would be able to copy and paste it into any note whenever I needed that verse, but my Supernote would have no data on [B]. It would have to trace [B] in some way... and although my Mac converts text in images to usable text in seconds, it might be asking allot of a Supernote to trace typed text so that it can then be copied and pasted directly into a Note as with [A]. Boox does it, but text copied from documents has to be pasted into a 'text box' within a note, which isn't the same and is fidgety and full of issues. I only asked the Supernote team to consider it as I assume I'm not the only one who is constantly hand writing text copied from documents I already have on the Supernote. Yes I can link to them, but if you are building an argument... jumping to links can be distracting if not impossible once you've exported the document as a PDF with live links. Anyway, thank you for your reply and for taking the time to consider the issue... appreciated.

@@karlmunford7000 , thank you for the feedback and use-case mention. Please take a look at the following screen cast, that shows a full import from text to pen strokes (in this case I used my own "typed" characters, but will probably implement with a font like Arial or Helvetica. Should come to PySN in the near future for everybody to use: app.screencast.com/UxqdvLhoK45fx?conversation=FLfDBYsiuscQIDOmggYhzf&tab=Details

Thank you!

Sorry for the late reply, Sahitya. I find the A6x2 is too small for this task. I am planning to purchase the A5x2 when it eventually comes out.

Dear MR N.A._McBee, Thank you for your thoughtful comment and for taking the time to try out the repo. I'm glad to hear that the "light gray marker feature" is working well for you and that it aligns with Supernote's original digest feature. I appreciate your feedback regarding the video and the suggestion to create a more step-by-step guide. I that the initial setup could be see as a bit complex. I will update the ReadMe to clarify some issues and reassure less technical users that they should be able to run the script easily by utilizing the Supernote folders they have set up and simply clicking on the shortcut. Your support and insights are invaluable, and I'm pleased to know that you find the project useful. Looking forward to any further feedback you might have. PS: I did fix a bug issue in the table of contents and I added handling bulk export to pdf of notebooks (pure .note files) contained in one of the script "input" folders. While this was a minor add-on, it makes sharing notebooks more usable and takes advantage of Supernote built-in text recognition to create a table of content and make the pdf searchable for your recognized handwritten notes. You'd have to reinstall (or simply replace the "digest.py") to have this features. Short video: th-cam.com/video/p3mXuRzog8M/w-d-xo.html Best regards, Max

The more challenging part was to handle the annotation so I left aside the note conversion to pdf. There are already tools like the Supernote tool library on GitHub that allow conversion. I’d like to go a bit beyond that and have automatic table of contents generated from the headers, as well as integrating the recognized text in the pdf + re uploading on a given folder of the supernote. Should post an update in the coming days, stay tuned.