Filters my .bash_history through a set of regexes, deduplicates, and sorts it
by most recently used.
Based on https://github.com/naggie/dotfiles/blob/master/scripts/cleanup-history
.bash_historyFormat:
```plaintext
echo each command has a timestamp immediately before it
echo 'after multiple timestamp lines, history will show the timestamp 1593575811'
echo after multiple timestamp lines, this will show the timestamp 3
echo 'when you run history this will show up with a timestamp long ago but still at the end of the list'
echo this will have the same timestamp as others above, duplicates don\'t matter
echo 'once you history -w all these extra timestamps will get removed'
for ((i=0;i<5;i++)); do echo $i; done
echo ^^ that was written on multiple lines
echo 'foo bar'
echo ^^ that was also written on multiple lines, cmdhist=on, lithist=off ```
If a line starts with #\d+, it will be interpreted as a timestamp.
```console $ export HISTFILE=./foo $ history -c $ echo 'this
that' $ history -w $ cat foo
echo 'this
that'
history -w $ history -c $ history -r $ history 1 2020-07-06 08:16.14 | history -r 2 2020-07-06 08:15.15 | echo 'this 3 1969-12-31 17:20.34 | that' 4 2020-07-06 08:15.25 | history -w 5 2020-07-06 08:16.16 | history ```
```console $ history -c $ echo 'foo
baz' $ history -w $ history # correct in memory 1 2020-07-06 08:24.30 | echo 'foo
baz' 2 2020-07-06 08:24.38 | history -w 3 2020-07-06 08:24.41 | history $ cat foo
echo 'foo
baz'
history -w
$ history -c # clear in-memory history
$ history -r # reread from file
$ history # now incorrectly interprets #1234 bar as a timestamp
1 2020-07-06 08:19.49 | history -r
2 2020-07-06 08:19.09 | echo 'foo
3 1969-12-31 17:20.34 | baz'
4 2020-07-06 08:19.31 | history -w
5 2020-07-06 08:19.51 | history
```
The deduplicated line count is a little different due to slightly different regexes ¯\(ツ)/¯. I think it's close enough to be informational.
console
$ wc -l bash_history.bak
86636 bash_history.bak
$ hyperfine --warmup=5 --prepare='cp bash_history.bak bash_history_python' \
--export-markdown=bash-history-python.txt \
--time-unit=millisecond \
'python3 cleanup-history.py bash_history_python'
$ wc -l bash_history_python
73149 bash_history_python
$ hyperfine --warmup=5 --prepare='cp bash_history.bak bash_history_rust' \
--export-markdown=bash-history-rust.txt \
--time-unit=millisecond \
'cleanup-history-rs/target/release/cleanup-history bash_history_rust'
$ wc -l bash_history_rust
64638 bash_history_rust
| Command | Mean [ms] | Min [ms] | Max [ms] |
|:---|---:|---:|---:|
| python3 cleanup-history.py bash_history_python | 2069.9 ± 112.4 | 1935.1 | 2356.4 |
| cleanup-history-rs/target/release/cleanup-history bash_history_rust | 653.5 ± 22.1 | 631.2 | 698.9 |