# -*- coding: utf-8 -*-
"""Klärt die verbleibenden Layout-Fragen vor dem Bau des Parsers.

Prüft: exakte Spaltengrenzen, Kreuz-in-Checkbox-Erkennung, Unterstreichungen
(Kernelemente der Musterantworten), Kapitelkopfzeilen und seitenübergreifende
Fragen.
"""
import io
import re
import sys
from collections import Counter, defaultdict

import fitz

sys.stdout = io.TextIOWrapper(sys.stdout.buffer, encoding="utf-8")
PDF = r"C:\Antigravity\Waffensachkundeprüfung\Fragenkatalog_sachkunde_mitAntworten.pdf"
FIRST_PAGE = 5  # 0-basiert: PDF-Seite 6

doc = fitz.open(PDF)

# ---------------------------------------------------------------- Spalten
vlines = Counter()
hlines_full = Counter()
box_x = Counter()
for pno in range(FIRST_PAGE, doc.page_count):
    for d in doc[pno].get_drawings():
        for item in d["items"]:
            if item[0] == "l":
                p1, p2 = item[1], item[2]
                if abs(p1.x - p2.x) < 0.7 and abs(p1.y - p2.y) > 3:
                    vlines[round(p1.x)] += 1
            elif item[0] == "re":
                r = item[1]
                if r.width < 1.6 and r.height > 3:
                    vlines[round(r.x0)] += 1
                if 6 < r.width < 20 and abs(r.width - r.height) < 4:
                    box_x[round(r.x0)] += 1

print("Vertikale Tabellenlinien (x -> Häufigkeit), Top 12:")
for x, n in vlines.most_common(12):
    print(f"   x={x:4d}  {n}")
print(f"\nCheckbox-x-Positionen: {box_x.most_common(6)}")

# ------------------------------------------------- Kreuz-in-Checkbox-Test
page = doc[FIRST_PAGE]
boxes, diagonals = [], []
for d in page.get_drawings():
    for item in d["items"]:
        if item[0] == "re":
            r = item[1]
            if 6 < r.width < 20 and abs(r.width - r.height) < 4 and r.x0 > 480:
                boxes.append(r)
        elif item[0] == "l":
            p1, p2 = item[1], item[2]
            if abs(p1.x - p2.x) > 2 and abs(p1.y - p2.y) > 2 and p1.x > 480:
                diagonals.append((p1, p2))

print(f"\nSeite 6: {len(boxes)} Checkboxen, {len(diagonals)} Diagonalsegmente")
for b in sorted(boxes, key=lambda r: r.y0):
    inside = sum(1 for p1, p2 in diagonals
                 if b.x0 - 2 <= p1.x <= b.x1 + 2 and b.y0 - 2 <= p1.y <= b.y1 + 2)
    words = page.get_text("words", clip=fitz.Rect(290, b.y0 - 4, 500, b.y0 + 14))
    label = " ".join(w[4] for w in words)[:46]
    print(f"   y={b.y0:6.1f}  Diagonalen={inside}  ->  {'ANGEKREUZT' if inside >= 2 else 'leer      '}  {label!r}")

# --------------------------------------------------------- Unterstreichung
print("\n--- Unterstreichungs-Kandidaten (dünne Füllrechtecke) ---")
thin = Counter()
for pno in range(FIRST_PAGE, doc.page_count):
    for d in doc[pno].get_drawings():
        if d.get("fill") is None:
            continue
        for item in d["items"]:
            if item[0] == "re":
                r = item[1]
                if 0.3 < r.height < 1.6 and r.width > 4:
                    thin[round(r.height, 2)] += 1
print(f"Höhenverteilung dünner Rechtecke: {sorted(thin.items())}")

# Beispielseite mit Musterantwort-Hervorhebung (Frage 1.01 auf Seite 6)
p6 = doc[FIRST_PAGE]
for d in p6.get_drawings():
    for item in d["items"]:
        if item[0] == "re":
            r = item[1]
            if 0.6 < r.height < 1.6 and r.width > 4 and r.x0 > 280:
                above = p6.get_text("words", clip=fitz.Rect(r.x0 - 2, r.y0 - 12, r.x1 + 2, r.y0))
                print(f"   Unterstrichen y={r.y0:.1f} x={r.x0:.0f}-{r.x1:.0f}: {' '.join(w[4] for w in above)!r}")

# ------------------------------------------------------------ Kapitelköpfe
print("\n--- Kapitelkopfzeilen (aus dem Seitenkopf) ---")
heads = defaultdict(list)
for pno in range(FIRST_PAGE, doc.page_count):
    head = doc[pno].get_text("text", clip=fitz.Rect(60, 30, 540, 95)).strip()
    lines = [l.strip() for l in head.splitlines() if l.strip()]
    heads[tuple(lines[:3])].append(pno + 1)
for key, pages in heads.items():
    print(f"   Seiten {pages[0]}-{pages[-1]} ({len(pages)}): {key}")

# ------------------------------------------- Fragennummern und Seitenlage
print("\n--- Fragennummern je Seite (Spalte 1) ---")
num_re = re.compile(r"^\d{1,3}(?:\.\d{2,3})?$")
total = 0
per_page_first = {}
for pno in range(FIRST_PAGE, doc.page_count):
    words = doc[pno].get_text("words", clip=fitz.Rect(60, 92, 103, 800))
    nums = [w[4] for w in sorted(words, key=lambda w: w[1]) if num_re.match(w[4])]
    total += len(nums)
    if nums:
        per_page_first[pno + 1] = (nums[0], nums[-1], len(nums))
print(f"Gefundene Fragennummern gesamt: {total}")
sample = list(per_page_first.items())
for pg, v in sample[:5] + sample[-5:]:
    print(f"   Seite {pg}: {v[0]} … {v[1]} ({v[2]} Fragen)")