#!/usr/bin/env python3 """Extract full structure of template1.pptx: slides, layouts, masters, theme.""" import re, json, glob, os from xml.etree import ElementTree as ET NS = { 'a': 'http://schemas.openxmlformats.org/drawingml/2006/main', 'p': 'http://schemas.openxmlformats.org/presentationml/2006/main', 'r': 'http://schemas.openxmlformats.org/officeDocument/2006/relationships', } EMU_IN = 914400 def qn(prefix, tag): return '{%s}%s' % (NS[prefix], tag) def emu2in(v): try: return round(int(v) / EMU_IN, 2) except: return None def get_fill(sp_el): """Get solid fill color from shape properties.""" colors = [] for sf in sp_el.iter(qn('a', 'solidFill')): srgb = sf.find(qn('a', 'srgbClr')) sch = sf.find(qn('a', 'schemeClr')) if srgb is not None: colors.append('#' + srgb.get('val')) elif sch is not None: colors.append('scheme:' + sch.get('val')) return colors def get_text_runs(txbody): """Extract paragraphs with runs: text, size, bold, color, font.""" paras = [] if txbody is None: return paras for p in txbody.findall(qn('a', 'p')): runs = [] for r in p.findall(qn('a', 'r')): t = r.find(qn('a', 't')) rpr = r.find(qn('a', 'rPr')) run = {'text': t.text if t is not None and t.text else ''} if rpr is not None: if rpr.get('sz'): run['size'] = int(rpr.get('sz')) / 100 if rpr.get('b') == '1': run['bold'] = True sf = rpr.find(qn('a', 'solidFill')) if sf is not None: srgb = sf.find(qn('a', 'srgbClr')) sch = sf.find(qn('a', 'schemeClr')) if srgb is not None: run['color'] = '#' + srgb.get('val') elif sch is not None: run['color'] = 'scheme:' + sch.get('val') lat = rpr.find(qn('a', 'latin')) if lat is not None and lat.get('typeface') and not lat.get('typeface').startswith('+'): run['font'] = lat.get('typeface') runs.append(run) # paragraph alignment ppr = p.find(qn('a', 'pPr')) para = {'runs': runs} if ppr is not None and ppr.get('algn'): para['align'] = ppr.get('algn') if runs: paras.append(para) return paras def parse_shape(sp_el, rels): """Parse a single shape (sp, pic, graphicFrame, grpSp).""" out = {} cNvPr = sp_el.find('.//' + qn('p', 'cNvPr')) if cNvPr is None: cNvPr = sp_el.find('.//' + qn('a', 'cNvPr')) if cNvPr is not None: out['name'] = cNvPr.get('name') xfrm = sp_el.find('.//' + qn('a', 'xfrm')) if xfrm is not None: off = xfrm.find(qn('a', 'off')) ext = xfrm.find(qn('a', 'ext')) if off is not None: out['x'] = emu2in(off.get('x')) out['y'] = emu2in(off.get('y')) if ext is not None: out['w'] = emu2in(ext.get('cx')) out['h'] = emu2in(ext.get('cy')) # geometry geom = sp_el.find('.//' + qn('a', 'prstGeom')) if geom is not None: out['geometry'] = geom.get('prst') fill = get_fill(sp_el) if fill: out['fill'] = fill # line ln = sp_el.find(qn('a', 'ln')) if ln is not None: lfill = get_fill(ln) if lfill: out['line'] = lfill # text txbody = sp_el.find(qn('p', 'txBody')) if txbody is None: txbody = sp_el.find(qn('a', 'txBody')) if txbody is not None: paras = get_text_runs(txbody) if paras: out['text'] = paras # picture rel blip = sp_el.find('.//' + qn('a', 'blip')) if blip is not None and rels: rid = blip.get(qn('r', 'embed')) if rid and rid in rels: out['image'] = rels[rid] # chart chart = sp_el.find('.//' + qn('a', 'chart')) if chart is not None and rels: rid = chart.get(qn('r', 'id')) if rid and rid in rels: out['chart'] = rels[rid] return out def parse_slide(path, rels_path): rels = {} if os.path.exists(rels_path): tree = ET.parse(rels_path) for rel in tree.getroot(): rels[rel.get('Id')] = rel.get('Target').replace('../', '') tree = ET.parse(path) root = tree.getroot() shapes = [] def walk(el, depth=0): tag = el.tag.split('}')[-1] if tag in ('sp', 'pic', 'graphicFrame'): s = parse_shape(el, rels) s['tag'] = tag if depth > 0: s['depth'] = depth shapes.append(s) elif tag == 'grpSp': shapes.append({'tag': 'grpSp', **parse_shape(el, rels)}) for child in el: walk(child, depth + 1) else: for child in el: walk(child, depth) spTree = root.find('.//' + qn('p', 'spTree')) if spTree is not None: for child in spTree: walk(child) return shapes def parse_all(): base = '_extract' result = {'slides': []} for i in range(1, 17): path = f'{base}/ppt/slides/slide{i}.xml' rels = f'{base}/ppt/slides/_rels/slide{i}.xml.rels' result['slides'].append({'slide': i, 'shapes': parse_slide(path, rels)}) # layouts result['layouts'] = [] for lf in sorted(glob.glob(f'{base}/ppt/slideLayouts/slideLayout*.xml')): num = re.search(r'slideLayout(\d+)', lf).group(1) rels = f'{base}/ppt/slideLayouts/_rels/slideLayout{num}.xml.rels' r = {} if os.path.exists(rels): tree = ET.parse(rels) for rel in tree.getroot(): r[rel.get('Id')] = rel.get('Target').replace('../', '') # layout name tree = ET.parse(lf) name = None for cNvPr in tree.getroot().findall('.//' + qn('p', 'cNvPr')): name = cNvPr.get('name'); break shapes = parse_slide(lf, rels) result['layouts'].append({'file': num, 'name': name, 'shapes': shapes}) # masters result['masters'] = [] for mf in sorted(glob.glob(f'{base}/ppt/slideMasters/slideMaster*.xml')): num = re.search(r'slideMaster(\d+)', mf).group(1) rels = f'{base}/ppt/slideMasters/_rels/slideMaster{num}.xml.rels' r = {} if os.path.exists(rels): tree = ET.parse(rels) for rel in tree.getroot(): r[rel.get('Id')] = rel.get('Target').replace('../', '') shapes = parse_slide(mf, rels) result['masters'].append({'file': num, 'shapes': shapes}) return result if __name__ == '__main__': data = parse_all() with open('_extract/deck_structure.json', 'w', encoding='utf-8') as f: json.dump(data, f, ensure_ascii=False, indent=1) # summary for s in data['slides']: texts = [] for sh in s['shapes']: if 'text' in sh: for p in sh['text']: texts.append(''.join(r['text'] for r in p['runs'])) print(f"--- Slide {s['slide']}: {len(s['shapes'])} shapes") for t in texts[:12]: if t.strip(): print(f" {t[:80]}")