ShowMyLog

I log all my daily activities in fixed-format text files. The script showmylog.py helps me analyze them.

I use a text editor to create an input file (see example.mylog). Then I run showmylog.py on the input file. It prints a summary of my activities on screen (see output.txt for example) and outputs an HTML report (see report.html for example). The report contains a timeline-diagram where you can hover over intervals to see details.

How logging helps

I started logging all my activities a few years ago. This has these advantages:

• I wanted to increase my productivity. If I can find out how much productive and non-productive work I do everyday, and how much time different activities take, I can make changes to my routine accordingly to optimize productive work.

• Maintaining a log has become a habit and now I'm often constantly aware of what I'm doing. This has helped me become more productive.

• I have a record of my past activities. If I can't recall what I did a few days ago, or I can't recall when I last did something, my log helps me find out.

Log file format considerations

Initially I used notebooks made of paper to do this, but I soon switched to text files. This was better since edits were less messy and I could grep those files.

I needed a machine-readable format so that I could do automated or semi-automated analysis on it. But I also needed a format which is not tied to a particular program, so these files must be human-readable and editable with a simple text editor.

Log file format details

A file is made up of records. Each record should span just one line. Anything after a # will be treated as a comment and ignored. There can be empty lines in the file.

Each record should be of this form:

work_type start_time end_time penalty duration label sublabel description

None of these parts should have spaces, except description.

• work_type can be any string, but I prefer keeping it a single-character string. Currently, I use these strings:
  • +: Positive work, like study or exercise.
  • -: Negative work, like wasting time on the internet or talking to friends about useless stuff.
  • !: Partially negative work.
  • :: Non-productive but necessary work, like eating, bathing, shopping for groceries.
  • s: Sleeping.
  • u: Unaccounted, i.e. time intervals not covered by a log file. I never use it, but showmylog.py reports unaccounted time under this work_type.
• start_time, end_time, penalty and duration should either be -:-- or --:-- or it should match the regex [\d?]{1,2}:[\d?]{2}. The symbol ? denotes uncertainty. duration should be less than or equal to end_time minus start_time.
• label and sublabel are used to identify a certain type of work. sublabel can be empty.
• description can be anything. It can also be empty.

I use the mylog file extension for these files.

You can see example.mylog for an example of a log file.

Additional constraints imposed on format by showmylog.py

• No two activities should overlap.
• All activities should be in chronological order.
• duration should be equal to end_time - start_time.

What does showmylog.py do?

• Does basic validation of log files.
• Displays duration sums after grouping on work_type and (work_type, label).
• Creates a timeline visualization of the log in HTML.

Run ./showmylog.py --help for more info.

Currently showmylog.py has no way of handling uncertainty. It replaces all ? by 0 in times.

Code

This code is written for python 3.5 and above. It contains PEP 484 type annotations, which makes it possible to use mypy for static type-checking. showmylog.py passes the --strict mypy test.

Usage tips

Add this line to your ~/.vimrc to enable syntax highlighting for mylog files:

au BufRead,BufNewFile *.mylog set filetype=diff

I store my files in ~/mylog/. Each file has a name of the form YYYY-MM-DD.mylog.

These are some handy bash aliases:

alias today='date "+%F"'
alias yesterday='date -d "yesterday" "+%F"'
alias logit='vim ~/mylog/$(today).mylog'
alias logyes='vim ~/mylog/$(yesterday).mylog'