"""I1. Paint map: share of page pixels in each pigment class (green, blue, red/brown, yellow)
for every folio image, from Yale's photographs (2,000 px). Vellum vs background is found
by brightness; pigment classes by HSV thresholds tuned on known pages. Joined to the codex
map so paint use can be compared across quires, scribes and languages.
Outputs: out/paint_map.csv
"""
import os, sys
_here = os.path.dirname(os.path.abspath(__file__))  # data/analysis, or a subfolder of it
sys.path.insert(0, _here if os.path.exists(os.path.join(_here, "paths.py")) else os.path.dirname(_here))
from paths import *  # noqa: repository locations
import os, re, csv, json, sys
import numpy as np
from PIL import Image

IMG = IMG

rows = []
codex = {r["page"]: r for r in csv.DictReader(open(os.path.join(TABLES, "codex_map.csv")))}
files = sorted(os.listdir(IMG))


def classify(path):
    im = Image.open(path).convert("RGB"); im.thumbnail((700, 700))
    a = np.asarray(im).astype(float) / 255
    hsv = np.asarray(im.convert("HSV")).astype(float) / 255
    h, s, v = hsv[..., 0], hsv[..., 1], hsv[..., 2]
    page = v > 0.35                       # vellum (black background excluded)
    n = page.sum()
    green = page & (h > 0.20) & (h < 0.47) & (s > 0.22)
    blue = page & (h > 0.50) & (h < 0.72) & (s > 0.20)
    red = page & ((h < 0.06) | (h > 0.92)) & (s > 0.35) & (v > 0.25)
    yellow = page & (h > 0.09) & (h < 0.19) & (s > 0.40) & (v > 0.5)
    brown_ink = page & (v < 0.45) & (s < 0.6)     # ink/dark strokes (text + outlines)
    return dict(page_px=int(n), green=green.sum() / n, blue=blue.sum() / n, red=red.sum() / n, yellow=yellow.sum() / n, ink=brown_ink.sum() / n)


for f in files:
    m = re.match(r"\d{3}_(\d+[rv])(?:_part)?\.jpg", f)
    if not m: continue
    folio = m.group(1)
    # map to ZL page names: plain folio pages; fold-out panels get the whole image
    r = classify(os.path.join(IMG, f))
    cm = codex.get("f" + folio)
    rows.append(dict(file=f, folio=folio, quire=cm["quire"] if cm else "", section=cm["section"] if cm else "", lang=cm["lang"] if cm else "", hand=cm["hand"] if cm else "", **r))
    print(f, {k: round(v, 3) for k, v in r.items() if k != "page_px"}, flush=True)
with open(os.path.join(TABLES, "paint_map.csv"), "w", newline="") as fh:
    w = csv.DictWriter(fh, fieldnames=rows[0].keys()); w.writeheader(); w.writerows(rows)
print(len(rows))