path: root/scripts/gen_map/gen_map.py

#!/usr/bin/env python3

import argparse
from datetime import datetime, timedelta
from pathlib import Path
import os
import re
import sys
import urllib

from bs4 import BeautifulSoup
import icalevents
from icalevents.icalevents import events
import requests

BAYERN = 1
ICAL_BASE_URL = 'https://www.schulferien.eu/downloads/ical4.php'

TEMPLATE_DIR = Path(
    os.environ['HOME']
) / 'Documents/MyDocuments/Work/Schule/Unterricht/templates'
TEMPLATE_FILE = 'map.md'

LPP_BASE_URL = 'https://www.lehrplanplus.bayern.de/fachlehrplan/gymnasium'

WEEKLY_HOURS = {
    'mathematik': {
        5: 4,
        6: 4,
        7: 4,
        8: 3,
        9: 4,
        10: 3,
        11: 4
    },
    'informatik': {
        9: 2,
        10: 2,
        11: 2
    },
    'nt_gym': {
        6: 2,
        7: 2
    }
}


def default_year() -> int:
    """return the default year"""
    now = datetime.now()
    return now.year if now.month < 8 else now.year + 1


def get_holidays(land: int, year: int) -> list[icalevents.icalparser.Event]:
    """Get list of all holidays in a school year"""
    start_year = datetime(year=year - 1, month=1, day=1)
    end_year = datetime(year=year + 1, month=12, day=31)

    params = {'land': land, 'year': year, 'type': 1}
    uri_first_year = ICAL_BASE_URL + '?' + urllib.parse.urlencode(params)

    first_holidays = events(url=uri_first_year, start=start_year, end=end_year)

    params['year'] = year + 1
    uri_second_year = ICAL_BASE_URL + '?' + urllib.parse.urlencode(params)

    second_holidays = events(url=uri_second_year,
                             start=start_year,
                             end=end_year)

    first_summer = next(
        filter(lambda e: "Sommerferien" in e.summary, first_holidays))

    second_summer = next(
        filter(lambda e: "Sommerferien" in e.summary, second_holidays))

    return [
        h for h in first_holidays + second_holidays
        if h.start >= first_summer.start and h.start <= second_summer.start
    ]


def fmt_holiday(holiday: icalevents.icalparser.Event) -> str:
    """Return a formatted holiday description"""
    h_start = f'{holiday.start.date().strftime("%d.%m")}'
    h_end = f'{holiday.end.date().strftime("%d.%m")}'
    return f'{h_start}-{h_end} {holiday.summary}'


def get_public_holidays(land: int, start: datetime,
                        end: datetime) -> list[icalevents.icalparser.Event]:
    """Get list of all public holidays in the school year"""

    params = {'land': land, 'year': start.year, 'type': 0}
    uri_first_year = ICAL_BASE_URL + '?' + urllib.parse.urlencode(params)

    params['year'] = end.year
    uri_second_year = ICAL_BASE_URL + '?' + urllib.parse.urlencode(params)

    all_public_holidays = events(url=uri_first_year, start=start,
                                 end=end) + events(
                                     url=uri_second_year, start=start, end=end)

    return [
        e for e in all_public_holidays if e.start >= start and e.start <= end
    ]


class LehrplanSection:

    def __init__(self, num, summary, duration, subsections=None):
        self.num = num
        self.summary = summary
        self.duration = duration
        self.subsections = subsections or []

    def has_subsections(self) -> bool:
        return bool(self.subsections)

    def __str__(self):
        return f'{self.num}: {self.summary} ({self.duration}h) {self.subsections}'

    def __repr__(self):
        return f'LehrplanSection({self.num}, {self.summary}, {self.duration}h, {self.subsections})'


Lehrplan = list[LehrplanSection]


def get_lehrplan(subject: str, form: str) -> Lehrplan:
    """Retrieve the sections and expected duration from lehrplanplus"""

    lpp_url = '/'.join([LPP_BASE_URL, form, subject])
    resp = requests.get(lpp_url, timeout=10)
    html = BeautifulSoup(resp.text)

    summary_re = re.compile(r'(.*) \(ca\. ([1-9][0-9]*) Std\.\)')

    lehrplan: Lehrplan = []
    top_level_sections = html.find_all(**{'class': 'headline_lvl1'})
    for section in top_level_sections:
        num = list(section.find(**{
            'class': 'head-absatz-num'
        }).children)[2].strip()

        summary = section.find(**{
            'class': 'head-absatz-title-short'
        }).string.strip()
        summary = summary.replace('\xa0', ' ')

        m = summary_re.match(summary)
        summary = m.group(1)
        duration = int(m.group(2))
        lehrplan.append(LehrplanSection(num, summary, duration))

    second_level_sections = html.find_all(**{'class': 'headline_lvl2'})
    for section in second_level_sections:
        summary = section.find(**{
            'class': 'head-absatz-title-long'
        }).string.strip()
        summary = summary.replace('\xa0', ' ')

        m = summary_re.match(summary)
        summary = m.group(1)
        duration = int(m.group(2))
        num = list(section.find(**{
            'class': 'head-absatz-num'
        }).children)[2].strip()
        lehrplan[int(num.split('.')[0]) - 1].subsections.append(
            LehrplanSection(num, summary, duration))

    return lehrplan


def gen_section_iter(lehrplan):
    sections = []
    for s in lehrplan:
        if not s.has_subsections():
            sections.append(s)
        else:
            sections.extend(s.subsections)

    next_section_in = sections[0].duration
    cur = 0

    def _section_iter(weekly_hours, total_hours) -> list[LehrplanSection]:
        nonlocal cur
        if cur == len(sections) - 1:
            return [sections[-1]]

        cur_sections = []

        nonlocal next_section_in
        # advance section
        if total_hours > next_section_in:
            if total_hours - weekly_hours < next_section_in:
                cur_sections.append(sections[cur])

            cur += 1
            next_section_in += sections[cur].duration

        cur_sections.append(sections[cur])

        return cur_sections

    return _section_iter


def gen_latex_table(weekly_hours: int, lehrplan: Lehrplan, school_holidays,
                    public_holidays) -> list[str]:
    """generate a latex table matching lehrplan section on kalendarweeks"""

    rows: list[str] = []

    lehrplan.append(LehrplanSection('', 'Puffer', 0))

    start = school_holidays[0].end
    end = school_holidays[-1].start

    holiday_iter = iter(school_holidays[1:])
    next_holiday = next(holiday_iter)

    section_iter = gen_section_iter(lehrplan)
    total_hours = 0

    week_iter = start + timedelta(days=1)
    while week_iter < end:
        if week_iter > next_holiday.start:
            rows.append((
                f'\\rowcolor{{blue!30}}{week_iter.isocalendar().week} & '
                f' \\multicolumn{{2}}{{c}}{{{fmt_holiday(next_holiday)}}} \\\\'
            ))
            week_iter = next_holiday.end + timedelta(days=1)
            next_holiday = next(holiday_iter)
            continue

        lp_sections = section_iter(weekly_hours, total_hours)

        lp_level = ' \\& '.join([s.num for s in lp_sections])
        lp_summary = ' \\& '.join([s.summary for s in lp_sections])

        rows.append(
            f'{week_iter.isocalendar().week} & {lp_level} & {lp_summary} \\\\')

        week_iter += timedelta(days=7)
        total_hours += weekly_hours

    return rows


def print_map(rows: list[str], template: Path, outfile):
    with open(template, 'r', encoding='utf-8') as template_file:
        template = template_file.readlines()

    template_prefix = template[:-6]
    template_suffix = template[-2:]

    print(''.join(template_prefix), file=outfile)
    print('\n'.join(rows), file=outfile)
    print(''.join(template_suffix), file=outfile)


def main():
    """Create a rough map template for a certain school year"""
    parser = argparse.ArgumentParser()
    parser.add_argument('subject', type=str)
    parser.add_argument('form', type=int)
    parser.add_argument('--template-dir')
    parser.add_argument('--year', default=default_year())
    parser.add_argument('--land', default=BAYERN)
    parser.add_argument('-o', '--output', default=sys.stdout)

    args = parser.parse_args()

    school_holidays = get_holidays(args.land, args.year)
    public_holidays = get_public_holidays(args.land, school_holidays[0].end,
                                          school_holidays[-1].start)

    weekly_hours = WEEKLY_HOURS[args.subject][args.form]
    lehrplan = get_lehrplan(args.subject, str(args.form))

    rows = gen_latex_table(weekly_hours, lehrplan, school_holidays,
                           public_holidays)

    if args.template_dir:
        template_path = Path(args.template_dir) / TEMPLATE_FILE
    else:
        template_path = TEMPLATE_DIR / TEMPLATE_FILE

    if args.output != sys.stdout:
        with open(args.output, 'w', encoding='utf-8') as outfile:
            print_map(rows, template_path, outfile)
    else:
        print_map(rows, template_path, args.output)


if __name__ == '__main__':
    main()