I’ve got a way that has been helpful for beeminding my journaling and I’d like to share it with you. This requires a Day One client on a machine and running a chron job to update a txt file.
The flow is like this:
Day One writes to a it’s SQL database after a journal entry is saved.
A script is kicked off by chron that reads records from the database and pipes the “ZCREATIONOSNAME” values from each journal entry into a txt file. (ZCREATIONOSNAME is just what OS was used to write the journal entry to the database - so it usually just says MacOS - unless I journal on my phone). So after this, what is essentially happens if you have a text file that is a list that looks like this:
MacOS
MacOS
MacOS
iOS
And the number of times a OS entry occurs = the number of journal entries that you have.
3. This file is saved on Dropbox.
4. URL enters the words into beeminder - which happens to be the same number as your journal entries.
Here is the script that is used:
#!/bin/sh
/usr/bin/sqlite3 /Users/USERNAME/Library/Group\ Containers/5U8NS4GX82.dayoneapp2/Data/Documents/DayOne.sqlite -cmd “SELECT ZCREATIONOSNAME FROM ZENTRY;” > dayoneentrycount.txt
I used Lingon X to set up the script to point at a dropbox directory.
Clever! I also Beemind “journaling” more broadly, although practically it’s almost always in Day One. This might be handy.
I decided to play around with it a bit; here’s a query that outputs the actual text content of my journal IDs 3, 6, and 9, replaces all the characters with x before being saved:
sqlite3 ~/Library/Group\ Containers/5U8NS4GX82.dayoneapp2/Data/Documents/DayOne.sqlite "SELECT ZMARKDOWNTEXT FROM ZENTRY WHERE ZJOURNAL IN (3, 6, 9);" | perl -0777 -i -pe 's/\w/x/g' > dayoneentries.txt
That gets me something like this, which might be anonymized enough to use for a word count goal:
# xxx <> xxxx xxxxx\-xx
xxx xxxxx xxxxxxxx
xxxx xxxxxxxx xx xx xxxx xx xxxxxxxx
xxxxxx xxxx xxxxx xxxx xxxx
\- xxx, xxxx xxxxxx x xxxxxx xxxxxxxxx xxx xxx xxx xxx xxx xxxxx xxxxx
\- xxxx xxxx xxxxxxxxx xx xxxxxxx xx xxxx
xxxxx xxxx xxxxxx xxxxxxx
xxxx xxxxx xxxxxxxx
xxxx xxxxx xx xx%, xxx xx%, xxxx xxxxxxx xxxx x xxx
xxxx xxx \-
...
…on second thought, I’m not sure I’d want even the above in the wild – I’d bet a machine learning algo could slurp up all the text content you’ve ever written, then make a pretty good recreation of your journal based on the word length combinations here.
Adding a second replacement turns this into something I’m a bit more comfortable sharing with the world:
sqlite3 ~/Library/Group\ Containers/5U8NS4GX82.dayoneapp2/Data/Documents/DayOne.sqlite "SELECT ZMARKDOWNTEXT FROM ZENTRY WHERE ZJOURNAL IN (3, 6, 9);" | perl -0777 -i -pe 's/\w/x/g' | perl -0777 -i -pe 's/(x+)/buffalo/g' > dayoneentries.txt
Awesome! I thought about doing word counts, but I didn’t want to burden Beeminder’s server with counting words on huge files. Also I don’t know perl very well at all.
How big does that file turn out to be? I imagine Beeminder’s URL minder would take a bit to count words from all of my journal entries. so I try to hit around around 1k words per day - so that would turn out to be around 365k words by the end of the year.
BTW, do you know of any way to get a time length value from a directory of mp3s? I would like to beemind the # of seconds of songs I make, but I can’t figure out how to make it work.
Somewhat aside: @dreev et al, is there a better way “url-mind” a single data point? This example works by creating a text document with the appropriate number of words, and uploading that to Dropbox.
But it seems like it would be more elegant to simply output a file with a single integer. Is there a simple way to do that?
