Read, organize, archive

I heavily use getpocket to collect articles and save them for later reading. While on my mobile devices I use the official app, on my desktop systems (mostly tmux + Emacs) I use pocket-reader.el quite heavily to collect/organize/archive my list of articles. After my transition from Spacemacs to Doom Emacs I still love evil mode and like to do most of the stuff in normal mode. That’s why I’ve added some keybindings to pocket-reader:

(map! :map pocket-reader-mode-map
      :after pocket-reader
      :nm "d" #'pocket-reader-delete
      :nm "a" #'pocket-reader-toggle-archived
      :nm "TAB" #'pocket-reader-open-url
      :nm "tr" #'pocket-reader-remove-tags
      :nm "ta" #'pocket-reader-add-tags
      :nm "gr" #'pocket-reader-refresh
      :nm "p" #'pocket-reader-search
      :nm "y" #'pocket-reader-copy-url
      :nm "Y" #'dorneanu/pocket-reader-copy-to-scratch)

This way I can add/remove tags, archive articles, open links without leaving normal mode in evil.

Read later on E-Reader

Since I tend to read on my e-reader most of the time, I also wanted to have an almost automated way of saving articles to Epub and send these to my device. Fortunately PocketBook devices can sync with Dropbox which made my life quite easy in the past. I just had to copy e-books, PDFs, Epubs to a specific folder in Dropbox and these will eventually sync with my device once WiFi is activated. That’s first step of automation.

As for the Epub conversion I’ve used pandoc in the past which still does its job great. Initially I’ve used monolith to save complete web pages as HTML but I’ve realized the HTML also contained useless ads, images, text. That’s why I’ve searched for ways how to make the content more readable and discovered rdrview. It’s written in C and applies Firefox’s reader view to web pages. Here are some examples:

$ rdrview -M
Title: Note taking in 2021 -
Excerpt: [Update 2021-06-22] If you’re more interested in the Tiddlywiki aspect of this post you can also check this Tiddlywiki Google Groups thread. [Update 2021-06-18] This post caught some attention on this Hackernews thread. You might want to check the comments. Based on the recommendations in the thread I’ve put together a list of (digital) solutions (besides the preferred ones) Almost 6 years ago I was blogging about organizing and visualizing knowledge.
Readerable: Yes

And now I’m using some xpath to extract the title:

$ rdrview -T title,body -H | xmllint --html --xpath "//div/h1/text()" -
Note taking in 2021 -

Below I’ve glued everything together in order to:


DROPBOX_DIR="dropbox:Apps/Dropbox PocketBook/articles/2021/"

add_link_to_dropbox() {
    # Create tmp file

    # Make link readable
    rdrview -T title,body -H $1 > $TEMP_FILE

    # Extract title
    TITLE=$(xmllint --html --xpath "//div/h1/text()" 2>/dev/null ${TEMP_FILE})
    echo "[-] Converting $TITLE"

    # Convert to PDF
    pandoc --pdf-engine=xelatex --metadata title="${TITLE}" -f html -t epub -o ${OUTPUT_FILE} ${TEMP_FILE}

    # Copy to dropbox
    rclone copy ${OUTPUT_FILE} "${DROPBOX_DIR}"

    # Log
    echo "[-] Successfully added ${OUTPUT_FILE} to dropbox."

    # Clean up
    rm $TEMP_FILE

add_link_to_dropbox $1

Let’s give it a try using as an URL:

$ ~/work/dropbox/add_links_to_dropbox
[-] Converting Pandoc - About pandoc
[-] Successfully added /home/victor/work/dropbox/rdrview/Pandoc_-_About_pandoc.epub to dropbox.

Emacs for Everything

Now I’d like to be able to call that script from Emacs without copy/paste URLs and hand them over to my script. The most difficult part here was to come up with some valid Elisp code. What I wanted was:

And this is what I’ve got (my first elisp function ever):

;; Copy current url to scratch buffer
(defun dorneanu/pocket-reader-copy-to-scratch ()
  "Copy URL of current item to kill-ring/clipboard."
  (when-let ((id (tabulated-list-get-id))
             (item (ht-get pocket-reader-items id))
             (url (pocket-reader--get-url item)))
    (with-current-buffer "*scratch*"
      (insert url)
     (message "Added: %s to scratch buffer" url)))

This will insert the URL into the scratch buffer and add a new line. Once you have your URLs in the buffer you can use shell-command-on-region with xargs -n1 <script> as shown in the asciinema below.