"""Diagnose sparse-page errors: crop both MISSED (FN) and FALSE-POSITIVE (FP) boxes.
Usage: python _analyze_sparse.py <version> [355|dean]"""
import csv, glob, re, sys
from pathlib import Path
import fitz
from PIL import Image
Image.MAX_IMAGE_PIXELS = None

HERE = Path(__file__).resolve().parent
ROOT = HERE.parent
ver = sys.argv[1] if len(sys.argv) > 1 else "v12_fastocr"
grp = sys.argv[2] if len(sys.argv) > 2 else "355"
PDF = (ROOT / "raw_pdfs" / ("355 E 86th St Mechanical 100% CD Set_flattened.pdf" if grp == "355"
       else "2023.10.12_953 Dean Street_M-marked.pdf"))
GT = {2: HERE/"ground_truth"/"355_p02_cellar.csv", 13: HERE/"ground_truth"/"355_p13_15th.csv",
      4: HERE/"ground_truth"/"dean_p04.csv", 6: HERE/"ground_truth"/"dean_p06.csv"}
RUN = HERE / "runs" / ver / grp
OUT = HERE / "sparse_crops"; OUT.mkdir(exist_ok=True)
DPI = 300

def load(p, is_gt):
    out = []
    for r in csv.DictReader(open(p, encoding="utf-8")):
        if not r.get("x"): continue
        act = (r.get("action") or "").strip()
        if is_gt and act.lower() == "delete": continue
        tag = r.get("tag", "")
        if is_gt and re.match(r"(?i)^ac[-\s]", act): tag = act
        try: x,y,w,h = int(float(r["x"])),int(float(r["y"])),int(float(r["w"])),int(float(r["h"]))
        except (KeyError, ValueError): continue
        out.append({"tag": tag.upper().replace(" ",""), "x":x,"y":y,"w":w,"h":h,
                    "status": (r.get("status") or "")})
    return out

def iou(a,b):
    ix0,iy0=max(a["x"],b["x"]),max(a["y"],b["y"])
    ix1=min(a["x"]+a["w"],b["x"]+b["w"]); iy1=min(a["y"]+a["h"],b["y"]+b["h"])
    iw,ih=max(0,ix1-ix0),max(0,iy1-iy0); inter=iw*ih
    return 0.0 if inter<=0 else inter/(a["w"]*a["h"]+b["w"]*b["h"]-inter)

doc = fitz.open(PDF)
pages = [2,13] if grp=="355" else [4,6]
for page in pages:
    gts = load(GT[page], True)
    pf = [f for f in glob.glob(str(RUN/"floor_*_review.csv"))
          if any(str(r.get("page_num"))==str(page) for r in csv.DictReader(open(f,encoding="utf-8")))]
    preds = load(pf[0], False) if pf else []
    fn = [g for g in gts if max((iou(g,p) for p in preds), default=0) < 0.3]
    fp = [p for p in preds if max((iou(p,g) for g in gts), default=0) < 0.3]
    print(f"\n=== {ver} {grp} p{page}: GT {len(gts)}, pred {len(preds)}, "
          f"MISSED {len(fn)}, FALSE-POS {len(fp)} ===")
    for m in fn: print(f"  MISS {m['tag']:8s} at ({m['x']},{m['y']}) {m['w']}x{m['h']}")
    for m in fp: print(f"  FP   {m['tag']:8s} at ({m['x']},{m['y']}) {m['w']}x{m['h']} status={m['status']}")
    pg = doc.load_page(page-1); mat = fitz.Matrix(DPI/72,DPI/72)
    pix = pg.get_pixmap(matrix=mat, alpha=False)
    img = Image.frombytes("RGB", [pix.width, pix.height], pix.samples)
    for kind, items in (("miss", fn), ("fp", fp)):
        for i,m in enumerate(items):
            cx,cy=m["x"]+m["w"]//2,m["y"]+m["h"]//2; pad=420
            crop=img.crop((max(0,cx-pad),max(0,cy-pad),min(img.width,cx+pad),min(img.height,cy+pad)))
            crop.save(OUT/f"{grp}_p{page}_{kind}{i+1}_{m['tag']}.png")
print(f"\nCrops -> {OUT}")