from pathlib import Path
from datetime import datetime, timezone
import zipfile,json,hashlib,subprocess,xml.etree.ElementTree as ET
import fitz
from PIL import Image,ImageOps,ImageDraw

out=Path(__file__).resolve().parent
extract=out/'first-source-extracted'; extract.mkdir(exist_ok=True)
render=out/'renders';render.mkdir(exist_ok=True)
sha=lambda b:hashlib.sha256(b).hexdigest()
with zipfile.ZipFile(out/'first-source.zip') as z:
    assert z.testzip() is None
    manifest=[]
    for i in z.infolist():
        target=(extract/i.filename).resolve()
        assert target.is_relative_to(extract.resolve())
        assert not i.is_dir()
        data=z.read(i)
        target.parent.mkdir(parents=True,exist_ok=True)
        target.write_bytes(data)
        manifest.append({'path':i.filename,'bytes':len(data),'sha256':sha(data)})
profile=json.loads((extract/'brand-system.json').read_text())
pdf=fitz.open(extract/'brandbook.pdf')
texts=[p.get_text() for p in pdf]
(out/'pdf-text.txt').write_text('\n\n'.join(f'PAGE {i+1}\n{t}' for i,t in enumerate(texts)),encoding='utf-8')
subprocess.run(['pdftoppm','-scale-to','1400','-png',str(extract/'brandbook.pdf'),str(render/'pdf')],check=True,capture_output=True)
svgs=[]
for p in sorted(extract.rglob('*.svg')):
    root=ET.fromstring(p.read_bytes())
    doc=fitz.open(p)
    page=doc[0]
    scale=1400/max(page.rect.width,page.rect.height)
    dest=render/('svg-'+p.parent.name+'-'+p.stem+'.png')
    page.get_pixmap(matrix=fitz.Matrix(scale,scale)).save(dest)
    svgs.append({'path':p.relative_to(extract).as_posix(),'validXml':True,'viewBox':root.attrib.get('viewBox'),'textNodes':sum(e.tag.endswith('}text') for e in root.iter()),'paths':sum(e.tag.endswith('}path') for e in root.iter()),'primaryColorPresent':'#243c35' in p.read_text().lower(),'render':dest.name})

def sheets(files,prefix,cols=3,rows=3):
    for k in range(0,len(files),cols*rows):
        batch=files[k:k+cols*rows]
        sheet=Image.new('RGB',(cols*480,rows*410),'#dddddd')
        d=ImageDraw.Draw(sheet)
        for n,p in enumerate(batch):
            im=Image.open(p).convert('RGB'); im.thumbnail((464,370))
            x=(n%cols)*480+8;y=(n//cols)*410+26
            sheet.paste(im,(x,y));d.text((x,y-20),p.stem,fill='black')
        sheet.save(render/f'{prefix}-{k//(cols*rows)+1}.jpg',quality=94)
sheets(sorted(render.glob('pdf-*.png')),'pdf-contact')
sheets(sorted(render.glob('svg-*.png')),'svg-contact')
report={'generatedAt':datetime.now(timezone.utc).isoformat(),'sourceZipSha256':sha((out/'first-source.zip').read_bytes()),'entries':manifest,'pdfPages':len(pdf),'pdfSha256':sha((extract/'brandbook.pdf').read_bytes()),'svgCount':len(svgs),'svgInspection':svgs,'identity':profile['identity'],'pdfTextExtractionCharacters':sum(map(len,texts)),'visualReview':'pending','separatePdfAttempt':{'downloadListener':'timeout_after_60_seconds','fileFound':False,'retryClicks':0,'packagedPdfPresent':True}}
(out/'artifact-inspection.json').write_text(json.dumps(report,ensure_ascii=False,indent=2)+'\n',encoding='utf-8')
print(json.dumps({k:report[k] for k in ['sourceZipSha256','pdfPages','pdfSha256','svgCount','pdfTextExtractionCharacters']},indent=2))
