extract.py

from bs4 import BeautifulSoup, Tag
import datetime, json, pathlib, re

# ------------------------------------------------------------------
# 1.  Parse one “week N” HTML file  ➜  { Monday: {meat/…}, … Friday }
# ------------------------------------------------------------------
def _extract_line(block, label):
    p = block.find(
        lambda t: t.name == "p" and t.find("strong") and label in t.find("strong").text
    )
    if not p:
        return None
    return p.get_text(" ", strip=True).replace(label, "").strip()

def parse_week(file_path: str):
    soup = BeautifulSoup(pathlib.Path(file_path).read_text(encoding="utf-8"),
                         "html.parser")

    menu = {}
    for h2 in soup.find_all("h2"):
        day = h2.get_text(strip=True)
        if day not in ("Monday", "Tuesday", "Wednesday", "Thursday", "Friday"):
            continue

        # the two blocks that follow each <h2> always have this class
        blocks = []
        candidate = h2
        while len(blocks) < 2:
            candidate = candidate.find_next("div", class_="well alternate-blue")
            if not candidate:
                break
            blocks.append(candidate)

        meatfree_block = next(b for b in blocks if "Meat-free" in b.h3.text)
        meat_block     = next(b for b in blocks if "Meat option" in b.h3.text)

        menu[day] = {
            "meat":      _extract_line(meat_block,     "Main course"),
            "meatfree":  _extract_line(meatfree_block, "Main course"),
            "desert":    _extract_line(meatfree_block, "To finish"),
        }
    return menu

# Parse every weekly pattern we have (order is important!)
WEEK_FILES      = ["week1.html", "week2.html", "week3.html"]           # adjust paths if needed
WEEK_PATTERNS   = [parse_week(f) for f in WEEK_FILES]

# -----------------------------------------------------------
# 2.  Build the calendar between 21 Apr 2025 and 24 Oct 2025
# -----------------------------------------------------------
START_DATE  = datetime.date(2025, 4, 21)   # first Monday on the menus
END_DATE    = datetime.date(2025,10,24)    # last Friday requested

# School / public-holiday breaks where NO menu is served (inclusive)
BREAKS = [
    (datetime.date(2025, 5, 26), datetime.date(2025, 5, 30)),  # May half-term
    (datetime.date(2025, 7, 24), datetime.date(2025, 8, 31)),  # Summer holidays
]

def in_break(day):
    return any(start <= day <= end for start, end in BREAKS)

def build_full_menu(start_idx=1):
    """
    start_idx tells us which WEEK_PATTERNS element applies to the very first
    teaching week (0 = week1.html, 1 = week2.html, 2 = week3.html).
    From the JSON you verified, the 21 Apr 2025 menu is Week-2, so offset 1.
    """
    full = {}
    week_idx = start_idx
    current  = START_DATE

    while current <= END_DATE:
        if in_break(current):                     # skip whole holiday weeks
            current += datetime.timedelta(days=7)
            continue

        pattern = WEEK_PATTERNS[week_idx % len(WEEK_PATTERNS)]

        for offset, weekday in enumerate(("Monday","Tuesday","Wednesday","Thursday","Friday")):
            day = current + datetime.timedelta(days=offset)
            if day > END_DATE or in_break(day):
                continue
            full[day.isoformat()] = pattern[weekday]

        week_idx += 1
        current  += datetime.timedelta(days=7)

    return full

TERM_MENU = build_full_menu(start_idx=1)          # <- matches your examples

# -----------------------------------------------------------
# 3.  Use it – print / write / import …
# -----------------------------------------------------------
if __name__ == "__main__":
    # Optionally dump to JSON file
    with open("term_menu.json", "w", encoding="utf-8") as f:
        json.dump(TERM_MENU, f, indent=2, ensure_ascii=False)