TL;DR: This converts a file like this (config file at ~/.config/ttally.py
):
from datetime import datetime
from typing import NamedTuple, Optional
class Weight(NamedTuple):
when: datetime
pounds: float
class Food(NamedTuple):
when: datetime
calories: int
food: str
quantity: float
water: int # how much ml of water was in this
# specify a special way to prompt for quantity
@staticmethod
def attr_validators() -> dict:
# https://purarue.xyz/d/ttally_types.py?redirect
from my.config.pura.ttally_types import prompt_float_default # type: ignore
# if I don't supply a quantity, default to 1
return {"quantity": lambda: prompt_float_default("quantity")}
class Event(NamedTuple):
"""e.g. a concert or something"""
event_type: str
when: datetime
description: str
score: Optional[int]
comments: Optional[str]
@staticmethod
def attr_validators() -> dict:
from my.config.pura.ttally_types import edit_in_vim # type: ignore
return {"comments": edit_in_vim}
import os
from enum import Enum
# dynamically create an enum using each line of the file as an option
with open(os.path.join(os.environ["HPIDATA"], ".self_types.txt")) as f:
SelfT = Enum("SelfT", [s.rstrip().upper() for s in f])
class Self(NamedTuple):
when: datetime
what: SelfT # type: ignore
to (shell aliases)...
alias event='python3 -m ttally prompt event'
alias event-now='python3 -m ttally prompt-now event'
alias event-recent='python3 -m ttally recent event'
alias food='python3 -m ttally prompt food'
alias food-now='python3 -m ttally prompt-now food'
alias food-recent='python3 -m ttally recent food'
alias self='python3 -m ttally prompt self'
alias self-now='python3 -m ttally prompt-now self'
alias self-recent='python3 -m ttally recent self'
alias weight='python3 -m ttally prompt weight'
alias weight-now='python3 -m ttally prompt-now weight'
alias weight-recent='python3 -m ttally recent weight'
Whenever I run any of those aliases, it inspects the model in the config file, and on-the-fly creates and runs an interactive interface like this:
... which saves what I enter to a file:
- when: 1598856786,
glasses": 2.0
ttally
is an interactive module using autotui
to save things I do often to YAML/JSON
Currently, I use this to store info like whenever I eat something/drink water/my current weight/thoughts on concerts
Given a NamedTuple
defined in ~/.config/ttally.py
, this creates interactive interfaces which validates my input and saves it to a file
The {tuple}-now
aliases set the any datetime
values for the prompted tuple to now
This also gives me {tuple}-recent
aliases, which print recent items I've logged. For example:
$ water-recent 5
2021-03-20 18:23:24 2.0
2021-03-20 01:28:27 1.0
2021-03-19 23:34:12 1.0
2021-03-19 22:49:05 1.5
2021-03-19 16:05:34 1.0
The -recent
aliases can accept all
to print all items, or a duration like 1d
or 6h
to print data from the last few hours/days.
- validates my user input to basic types
- stores it as a user-editable format (YAML)
- can be loaded into python as typed objects
- minimal boilerplate to add a new model
- can be synced across multiple machines without conflicts
- allow completely custom types or prompts - see autotui docs, my custom prompts
This intentionally uses YAML and doesn't store the info into a single "merged" database. That way:
- you can just open the YAML file and quickly change/edit some item, no need to re-invent a CRUD interface (though
ttally edit-recent
does exist) - files can be synced across machines and to my phone using syncthing without file conflicts
- prevents issues with trying to merge multiple databases from different machines together (I've tried)
The YAML files are versioned with the date/OS/platform, so I'm able to add items on my linux, mac, or android (using termux
) and sync them across all my devices using SyncThing
. Each device creates its own file it adds items to, like:
food-darwin-mbp.localdomain-2021-03.yaml
food-linux-bastion-2021-03.yaml
food-linux-localhost-2021-04.yaml
... which can then be combined back into python, like:
>>> from more_itertools import take # just to grab a few items
>>> from ttally.__main__ import ext
>>> from ttally.config import Food
>>> take(3, ext.glob_namedtuple(Food))
[Food(when=datetime.datetime(2020, 9, 27, 6, 49, 34, tzinfo=datetime.timezone.utc), calories=440, food='ramen, egg'),
Food(when=datetime.datetime(2020, 9, 27, 6, 52, 16, tzinfo=datetime.timezone.utc), calories=160, food='2 eggs'),
Food(when=datetime.datetime(2020, 9, 27, 6, 53, 44, tzinfo=datetime.timezone.utc), calories=50, food='ginger chai')]
... or into JSON using ttally export food
The from-json
command can be used to send this JSON which matches a model, i.e. providing a non-interactive interface to add items, in case I want to call this from a script
hpi query
from HPI
can be used with the ttally.__main__
module, like:
# how many calories in the last day
$ hpi query ttally.__main__.food --recent 1d -s | jq -r '(.quantity)*(.calories)' | datamash sum 1
2252
If you'd prefer to use JSON files, you can set the TTALLY_EXT=json
environment variable.
This can load data from YAML or JSON (or both at the same time), every couple months I'll combine all the versioned files to a single merged file using the merge
command:
ttally merge food
pip install ttally # can use 'pip install ttally[optional]' for extra features
Usage: ttally [OPTIONS] COMMAND [ARGS]...
Tally things that I do often!
Given a few namedtuples, this creates serializers/deserializers and an
interactive interface using 'autotui', and aliases to:
prompt using default autotui behavior, writing to the ttally datafile, same
as above, but if the model has a datetime, set it to now, query the 10 most
recent items for a model
Options:
--help Show this message and exit.
Commands:
datafile print the datafile location
drop-last drop the last n items
edit edit the datafile
edit-recent fuzzy select/edit recent items
export export all data from a model
from-json add item by piping JSON
generate generate shell aliases
merge merge all data for a model into one file
models list models
prompt tally an item
prompt-now tally an item (now)
recent print recently tallied items
update-cache cache export data
You need to setup a ~/.config/ttally.py
file. You can use the block above as a starting point, or with mine:
curl -s 'https://purarue.xyz/d/ttally.py' > ~/.config/ttally.py
To setup aliases; You can do it each time you launch you terminal like:
eval "$(python3 -m ttally generate)"
Or, 'cache' the generated aliases by putting a block like this in your shell config:
TTALLY_ALIASES="${HOME}/.cache/ttally_aliases"
if [[ ! -e "${TTALLY_ALIASES}" ]]; then # alias file doesn't exist
python3 -m ttally generate >"${TTALLY_ALIASES}" # generate and save the aliases
fi
source "${TTALLY_ALIASES}" # make aliases available in your shell
i.e., it runs the first time I open a terminal, but then stays the same until I remove the file
You can set the TTALLY_DATA_DIR
environment variable to the directory that ttally
should save data to, defaults to ~/.local/share/ttally
. If you want to use a different path for configuration, you can set the TTALLY_CFG
to the absolute path to the file.
For shell completion to autocomplete options/model names:
eval "$(_TTALLY_COMPLETE=bash_source ttally)" # in ~/.bashrc
eval "$(_TTALLY_COMPLETE=zsh_source ttally)" # in ~/.zshrc
eval "$(_TTALLY_COMPLETE=fish_source ttally)" # in ~/.config/fish/config.fish
ttally update-cache
can be used to speedup the export
and recent
commands:
Usage: ttally update-cache [OPTIONS]
Caches data for 'export' and 'recent' by saving the current data and an
index to ~/.cache/ttally
exit code 0 if cache was updated, 2 if it was already up to date
Options:
--print-hashes print current filehash debug info
--help Show this message and exit.
I run this using entr whenever the data files change. In the background, like:
find ~/data/ttally -type f | entr -n ttally update-cache
Default cache directory can be overwritten with the TTALLY_CACHE_DIR
environment variable
The entire ttally
library/CLI can also be subclassed/extended for custom usage, by using ttally.core.Extension
class and wrap_cli
to add additional click commands. For an example, see flipflop.py
cz
lets me fuzzy select something I've eaten in the past using fzf
, like: