#+TITLE: Program Usage History
#+PROPERTY: header-args:python :session *history*
#+PROPERTY: header-args:python+ :exports both
#+PROPERTY: header-args:python+ :tangle yes
#+PROPERTY: header-args:python+ :async yes

#+begin_src elisp :exports none
(setq-local org-image-actual-width '(1024))
#+end_src

#+RESULTS:
| 1024 |

* Init
#+begin_src python
from matplotlib import pyplot as plt
from matplotlib import dates as mdates
from pprint import pprint
from datetime import date, datetime

import copy
import json
import matplotlib as mpl
import numpy as np
import pandas as pd

mpl.rcParams['figure.dpi'] = 125
mpl.rcParams['figure.facecolor'] = '0.1'
mpl.rcParams['hatch.linewidth'] = 4.0
#+end_src

#+RESULTS:

#+begin_src python
plt.style.use('./palenight.mplstyle')
#+end_src

#+RESULTS:

** Colors
#+begin_src python
COLORS = {
    'k': '#292d3e',
    'r': '#f07178',
    'g': '#c3e88d',
    'y': '#ffcb6b',
    'b': '#82aaff',
    'm': '#c792ea',
    'c': '#89ddff',
    'w': '#d0d0d0',
    'k-l': '#434758',
    'r-l': '#ff8b92',
    'g-l': '#ddffa7',
    'y-l': '#ffe585',
    'b-l': '#9cc4ff',
    'm-l': '#e1acff',
    'c-l': '#a3f7ff',
    'w-l': '#ffffff'
}
#+end_src

#+RESULTS:

* Load data
** Read file
#+begin_src python
DATA = './history.json'

with open(DATA, 'r') as f:
    data_o = json.load(f)

# pprint(data)
#+end_src

#+RESULTS:
** Some preprocessing
#+begin_src python
def preprocess_data(data):
    data = copy.deepcopy(data)
    for category in data:
        for elem in category['elements']:
            for state in elem['states']:
                if 'TODO' in state['startDate']:
                    print(f'TODO in {elem["name"]} (startDate)')
                    state['startDate'] = state['startDate'].replace('TODO', '').strip()
                state['startDate'] = datetime.strptime(state['startDate'], '%Y-%m-%d')
                try:
                    if 'TODO' in state['endDate']:
                        print(f'TODO in {elem["name"]} (endDate)')
                        state['endDate'] = state['endDate'].replace('TODO', '').strip()
                    state['endDate'] = datetime.strptime(state['endDate'], '%Y-%m-%d')
                except KeyError:
                    state['endDate'] = datetime.today()
                try:
                    state['hatchColor'] = COLORS.get(state['hatchColor'], state['hatchColor'])
                except KeyError:
                    pass
            elem['states'].sort(key=lambda k: k['startDate'])
            try:
                elem['color'] = COLORS.get(elem['color'], elem['color'])
            except KeyError:
                pass
        category['elements'].sort(key=lambda k: k['states'][0]['startDate'])

    return data

data = preprocess_data(data_o)
#+end_src

#+RESULTS:
: TODO in Linux Mint (startDate)
: TODO in Windows (startDate)
: TODO in Windows (endDate)
: TODO in bash (startDate)
: TODO in Mailspring (startDate)
** Crop data
#+begin_src python
def crop_data(data, start_date):
    data = copy.deepcopy(data)
    for category in data:
        for elem in category['elements']:
            elem['states'] = [
                state for state in elem['states']
                if state['endDate'] >= start_date
            ]
            for state in elem['states']:
                if state['startDate'] <= start_date:
                    state['startDate'] = start_date

        category['elements'] = [
            element for element in category['elements']
            if len(element['states']) > 0
        ]

    data = [
        category for category in data
        if len(category['elements']) > 0
    ]
    return data

data = crop_data(data, datetime(2017, 1, 1))
#+end_src

#+RESULTS:

* Plot one category
#+begin_src python
CAT_W = 0.3

def process_category(category, ax, is_last=True, notes=None):
    ticks = list(range(0, -len(category['elements']), -1))
    if notes is None:
        notes = []
    min_start_date = 10**9
    for elem, k in zip(category['elements'], ticks):
        for state in elem['states']:
            start_date, end_date = mdates.date2num(state['startDate']), mdates.date2num(state['endDate'])
            min_start_date = min(min_start_date, start_date)
            kwargs = {}
            kwargs['color'] = elem.get('color', None)

            if state.get('state', None) == 'dashed':
                kwargs['hatch'] = '//'
                kwargs['edgecolor'] = state.get('hatchColor', 'y')
                kwargs['lw'] = 0

                note = state.get('note', None)
                if note is not None:
                    notes.append(note)
                    stars = '*' * len(notes)
                    ax.text(end_date, k + CAT_W * 0.7, stars, size=15)

            bars = ax.broken_barh(
                [(start_date, end_date - start_date)],
                (k - CAT_W, CAT_W * 2),
                ,**kwargs
            )

    ax.set_yticks(ticks)
    ax.set_yticklabels([elem['name'] for elem in category['elements']])
    ax.set_axisbelow(True)
    ax.grid(True, alpha=0.25)
    if not is_last:
        ax.tick_params(axis='x', which='both', labelbottom=False, length=0)
    else:
        ax.xaxis.set_major_formatter(mdates.DateFormatter('%Y-%m'))

    return notes

def plot_notes(fig, ax, notes, x=0.9, y = 0.03):
    if len(notes) > 0:
        notes_text = ''
        for i, note in enumerate(notes):
            notes_text += '*' * (i + 1) + ' ' + note + '\n'
        ax.text(x, y, notes_text, transform=fig.transFigure, va='top', ha='right')

fig, ax = plt.subplots(figsize=(12, 3))
notes = process_category(data[0], ax)
ax.set_title(data[0]['title'])
plot_notes(fig, ax, notes)
notes
#+end_src

#+RESULTS:
:RESULTS:
| Dual boot, rarely used |
[[file:./.ob-jupyter/c1be31055726c333300f950d82c2c9d15f9b0205.png]]
:END:

* Plot all separately
#+begin_src python :display plain
def plot_category(category):
    fig, ax = plt.subplots(figsize=(12, len(category['elements']) * 0.7 + 1))
    notes = process_category(category, ax)
    ax.set_title(category['title'])
    plot_notes(fig, ax, notes)
    return fig

def plot_separate(data):
    for category in data:
        fig = plot_category(category)
        # fig.tight_layout()
        fig.savefig(f'./img/{category["title"].replace("/", "-")}.png')

# plot_separate(data)
#+end_src

#+RESULTS:

* Plot all in one chart
#+begin_src python :file img/all.png
def plot_all(data):
    fig, axes = plt.subplots(
        len(data),
        gridspec_kw={
            'height_ratios': [len(datum['elements']) for datum in data]
        },
        figsize=(14, 11),
        sharex=True
    )
    notes = []
    for i, [datum, ax] in enumerate(zip(data, axes)):
        is_last = i == len(data) - 1
        notes = process_category(datum, ax, is_last=is_last, notes=notes)
        ax.yaxis.set_label_position("right")
        ax.set_ylabel(datum['title'], labelpad=16, rotation=270)
    plot_notes(fig, ax, notes, y=0.09)
    # fig.tight_layout()
    fig.subplots_adjust(hspace=0.15)
    ax.text(0.075, 0.08, f'upd. {datetime.now().strftime("%Y-%m-%d")}', transform=fig.transFigure, va='top', ha='left')

plot_all(data)
#+end_src

#+RESULTS:
[[file:img/all.png]]