plot.py

import numpy as np
import calendar
import datetime
import matplotlib.pyplot as plt
from typing import List, Any, Optional, Union, Tuple
from matplotlib.pyplot import Axes
from matplotlib.colors import LinearSegmentedColormap, ListedColormap
from july.helpers import (
    date_grid,
    cal_heatmap,
    get_month_outline,
    get_calendar_title,
)
from july.utils import preprocess_inputs, preprocess_month, unique
from july.rcmod import update_rcparams


def heatmap(
    dates: List[Union[str, datetime.date, datetime.datetime]],
    data: List[float],
    horizontal: bool = True,
    cmap: Union[str, LinearSegmentedColormap, ListedColormap] = "july",
    value_label: bool = False,
    date_label: bool = False,
    weekday_label: bool = True,
    month_label: bool = True,
    year_label: bool = True,
    month_grid: bool = False,
    colorbar: bool = False,
    frame_on: bool = False,
    value_format: str = "int",
    title: Optional[str] = None,
    cmin: Optional[int] = None,
    cmax: Optional[int] = None,
    ax: Optional[Axes] = None,
    **kwargs
) -> Axes:
    """Create heatmap of input dates and data.

    Args:
        dates: List like data structure with dates.
        data: List like data structure with numeric data.
        horizontal: Whether to plot heatmap horizontally. Grid shape (7, n_weeks)
            if True, (n_weeks, 7) if False.
        cmap: Colormap. Any matplotlib colormap works.
        value_label: Whether to add value label inside grid.
        date_label: Whether to add date label inside grid.
        weekday_label: Whether to label the short axis with weekday abbreviations.
        month_label: Whether to add month label(s) along the long axis.
        year_label: Whether to add year label(s) along the long axis.
        month_grid: Whether to outline each month in the grid.
        colorbar: Whether to add colorbar.
        frame_on: Whether to turn frame on.
        value_format: Format of value_label: 'int' or 'decimal'. Only relevant if
            `value_label` is True.
        title: Title of the plot.
        cmin: Minimum value of the colorbar. Defaults to minimum value of `data`.
            Only relevant if `colorbar` is True.
        cmax: Maximum value of the colorbar. Defaults to maximum value of 'data'.
            Only relevant if 'colorbar' is True.
        ax: Matplotlib Axes object.
        kwargs: Parameters passed to `update_rcparams`. Figure aesthetics. Named
            keyword arguments as defined in `update_rcparams` or a dict with any
            rcParam as key(s).
    Returns:
        Matplotlib Axes object.
    """
    update_rcparams(**kwargs)
    dates_clean, data_clean = preprocess_inputs(dates, data)
    cal = date_grid(dates_clean, data_clean, horizontal)
    ax = cal_heatmap(
        cal=cal,
        dates=dates_clean,
        horizontal=horizontal,
        cmap=cmap,
        value_label=value_label,
        date_label=date_label,
        weekday_label=weekday_label,
        month_label=month_label,
        year_label=year_label,
        month_grid=month_grid,
        colorbar=colorbar,
        frame_on=frame_on,
        value_format=value_format,
        title=title,
        cmin=cmin,
        cmax=cmax,
        ax=ax,
    )

    return ax


def month_plot(
    dates: List[Union[str, datetime.date, datetime.datetime]],
    data: List[Any],
    horizontal: bool = False,
    cmap: Union[str, LinearSegmentedColormap, ListedColormap] = "july",
    value_label: bool = False,
    date_label: bool = False,
    weeknum_label: bool = True,
    month_label: bool = True,
    colorbar: bool = False,
    value_format: str = "int",
    cal_mode: bool = False,
    title: Optional[str] = None,
    month: Optional[int] = None,
    year: Optional[int] = None,
    cmin: Optional[int] = None,
    cmax: Optional[int] = None,
    ax: Optional[Axes] = None,
    **kwargs
) -> Axes:
    """Create calendar shaped heatmap of one month in input dates and data.

    Args:
        dates: List like data structure with dates.
        data: List like data structure with numeric data.
        horizontal: Whether to plot heatmap horizontally. Grid shape (7, n_weeks)
            if True, (n_weeks, 7) if False.
        cmap: Colormap. Any matplotlib colormap works.
        value_label: Whether to add value label inside grid.
        date_label: Whether to add date label inside grid.
        weeknum_label: Whether to label the short axis with week numbers.
        month_label: Whether to add month label(s) along the long axis.
        colorbar: Whether to add colorbar.
        value_format: Format of value_label: 'int' or 'decimal'. Only relevant if
            `value_label` is True.
        cal_mode: Whether to pad the month to be six weeks.
        title: Title of the plot.
        month: Which month in 'dates' to plot.
        year: Which year in 'dates' to plot 'month' for. Only required if 'month'
            is ambiguous.
        cmin: Minimum value of the colorbar. Defaults to minimum value of `data`.
            Only relevant if `colorbar` is True.
        cmax: Maximum value of the colorbar. Defaults to maximum value of 'data'.
            Only relevant if 'colorbar' is True.
        ax: Matplotlib Axes object.
        kwargs: Parameters passed to `update_rcparams`. Figure aesthetics. Named
            keyword arguments as defined in `update_rcparams` or a dict with any
            rcParam as key(s).
    Returns:
        Matplotlib Axes object.
    """
    update_rcparams(**kwargs)
    dates_mon, data_mon = preprocess_month(dates, data, month=month, year=year)
    month = dates_mon[0].month
    month_grid = date_grid(dates_mon, data_mon, horizontal=horizontal)
    weeknum_grid = date_grid(
        dates_mon, [d.isocalendar()[1] for d in dates_mon], horizontal=horizontal
    )
    weeknum_labels: List[Any] = [int(x) for x in unique(weeknum_grid) if np.isfinite(x)]

    if cal_mode:
        # Pad all grids to have six rows so weeks align when plotted side by side.
        while len(month_grid) < 6:
            month_grid = np.vstack([month_grid, 7 * [np.nan]])
            weeknum_labels.append("")

    if not ax:
        _, ax = plt.subplots(figsize=(5, 4))

    ax = cal_heatmap(
        cal=month_grid,
        dates=dates_mon,
        horizontal=horizontal,
        cmap=cmap,
        value_label=value_label,
        value_format=value_format,
        date_label=date_label,
        year_label=False,
        month_label=False,
        frame_on=False,
        colorbar=colorbar,
        cmin=cmin,
        cmax=cmax,
        ax=ax,
    )

    ax.tick_params(axis="y", pad=8)

    if weeknum_label:
        if horizontal:
            ax.set_xticks([i + 0.5 for i in range(month_grid.shape[1])])
            ax.set_xticklabels(weeknum_labels)
        else:
            ax.set_yticks([i + 0.5 for i in range(month_grid.shape[0])])
            ax.set_yticklabels(weeknum_labels)
    else:
        if horizontal:
            ax.set_xticklabels([])
        else:
            ax.set_yticklabels([])

    outline_coords = get_month_outline(dates_mon, month_grid, horizontal, month)
    ax.plot(outline_coords[:, 0], outline_coords[:, 1], color="black", linewidth=1)
    ax.set_xlim(ax.get_xlim()[0] - 0.1, ax.get_xlim()[1] + 0.1)
    ax.set_ylim(ax.get_ylim()[0] + 0.1, ax.get_ylim()[1] - 0.1)
    if month_label:
        ax.set_title(calendar.month_name[month])
    if title:
        plt.suptitle(title, y=1.07, size="x-large")

    return ax


def calendar_plot(
    dates: List[Union[str, datetime.date, datetime.datetime]],
    data: List[Any],
    cmap: Union[str, LinearSegmentedColormap, ListedColormap] = "july",
    value_label: bool = False,
    date_label: bool = False,
    weeknum_label: bool = True,
    month_label: bool = True,
    value_format: str = "int",
    title: bool = True,
    ncols: int = 4,
    figsize: Optional[Tuple[float, float]] = None,
    **kwargs
) -> Axes:
    """Create calendar shaped heatmap of all months im input dates and data.

    Args:
        dates: List like data structure with dates.
        data: List like data structure with numeric data.
        cmap: Colormap. Any matplotlib colormap works.
        value_label: Whether to add value label inside grid.
        date_label: Whether to add date label inside grid.
        weeknum_label: Whether to label the short axis with week numbers.
        month_label: Whether to add month label(s) along the long axis.
        value_format: Format of value_label: 'int' or 'decimal'. Only relevant if
            `value_label` is True.
        title: Title of the plot.
        ncols: Number of columns in the calendar plot.
        ax: Matplotlib Axes object.
        figsize: Figure size. Defaults to sensible values determined from 'ncols'.
        kwargs: Parameters passed to `update_rcparams`. Figure aesthetics. Named
            keyword arguments as defined in `update_rcparams` or a dict with any
            rcParam as key(s).
    Returns:
        Matplotlib Axes object.
    """
    update_rcparams(**kwargs)
    dates_clean, data_clean = preprocess_inputs(dates, data)
    # Get unique years in input dates.
    years = sorted(set([day.year for day in dates_clean]))
    # Get unique months (YYYY-MM) in input dates.
    year_months = sorted(set([day.strftime("%Y-%m") for day in dates_clean]))

    nrows = int(np.ceil(len(year_months) / ncols))
    if not figsize:
        if ncols == 6:
            figsize = (12, 0.5 + nrows * 2)
        elif ncols == 5:
            figsize = (12, 1 + nrows * 2)
        elif ncols == 4:
            figsize = (14, 2 + nrows * 2)
        elif ncols == 3:
            figsize = (12, 2 + nrows * 2)

    fig, axes = plt.subplots(nrows, ncols, figsize=figsize)

    for i, year_month in enumerate(year_months):
        month = [day for day in dates_clean if day.strftime("%Y-%m") == year_month]
        vals = [
            val
            for day, val in zip(dates_clean, data_clean)
            if day.strftime("%Y-%m") == year_month
        ]
        month_plot(
            month,  # type: ignore
            vals,
            cmap=cmap,
            date_label=date_label,
            weeknum_label=weeknum_label,
            month_label=month_label,
            value_label=value_label,
            value_format=value_format,
            ax=axes.reshape(-1)[i],
            cal_mode=True,
        )

    for ax in axes.reshape(-1)[len(year_months) :]:
        ax.set_visible(False)

    plt.subplots_adjust(wspace=0.75, hspace=0.5)
    if title:
        plt.suptitle(get_calendar_title(years), fontsize="x-large", y=1.03)

    return axes