#!/usr/bin/env python3
"""Generate a nine-paper Echocardiography answer compendium using ReportLab only.

Input:  extracted_papers.txt beside this file
Output: ../output/Echocardiography_Paper_01_Answers.pdf ... Paper_09 and Master_Index.pdf
"""
from pathlib import Path
import re, html, shutil
from reportlab.lib import colors
from reportlab.lib.enums import TA_CENTER, TA_LEFT
from reportlab.lib.pagesizes import A4
from reportlab.lib.styles import getSampleStyleSheet, ParagraphStyle
from reportlab.lib.units import cm
from reportlab.platypus import (SimpleDocTemplate, Paragraph, Spacer, PageBreak,
                                KeepTogether, Table, TableStyle)

ROOT = Path(__file__).resolve().parents[1]
SOURCE = Path(__file__).with_name('extracted_papers.txt')
OUT = ROOT / 'output'

# Topic-specific, exam-safe building blocks. Each item intentionally supplies a clinical
# observation, a measurement/technique, a diagram prompt and a common scoring pitfall.
TOPICS = [
 ('aortic stenosis', 'Aortic stenosis is fixed LV outflow obstruction, usually calcific in adults.',
  'PLAX/PSAX define cusp number, calcification and LVH; obtain multiple-window CW Doppler (apical, right parasternal, suprasternal). Trace the highest envelope: ΔP=4V²; mean gradient is the time-averaged instantaneous gradient. Continuity equation: AVA=(CSA_LVOT×VTI_LVOT)/VTI_AV. Measure LVOT just proximal to the valve and use matched beats.',
  'PLAX: LV → LVOT → [calcified AV] → Ao; CW beam aligned through LVOT/AV.',
  '2D morphology → LVOT diameter/VTI → highest AV VTI → calculate AVA/gradients → reconcile flow and EF.',
  'Report valve severity with flow state, LVEF and symptoms; low-flow/low-gradient disease needs integrated assessment.',
  'Do not use a poorly aligned Doppler velocity or a single apical window; do not confuse high flow with severe stenosis.'),
 ('mitral', 'Mitral-valve disease is assessed by morphology, haemodynamic severity and chamber response.',
  'Use PLAX/PSAX and apical views for leaflet thickening, commissures, chordae and annulus. For stenosis, planimetry is preferred when image quality permits; pressure half-time MVA=220/PHT (ms), but is unreliable with major AR, altered compliance or immediately after valvotomy. For regurgitation, integrate colour jet, vena contracta, PISA, pulmonary venous flow and LV/LA size.',
  'PLAX: LA | anterior/posterior MV leaflets | LV; trace inflow with PW Doppler at leaflet tips.',
  'Morphology → colour localisation → CW/PW haemodynamics → quantitative measure → chamber/pulmonary-pressure consequence.',
  'Balloon-valvotomy suitability includes pliable non-calcified leaflets, limited subvalvular disease and no LA thrombus/significant MR.',
  'Never grade MR from jet area alone; PHT is not a stand-alone MVA in abnormal loading conditions.'),
 ('doppler', 'Doppler echocardiography converts frequency shift from moving red cells into velocity and pressure estimates.',
  'Doppler shift is proportional to velocity and cosθ; minimise insonation angle. PW Doppler samples a chosen depth but aliases above the Nyquist limit. CW records the highest velocity along the whole beam without aliasing but has no range resolution. Modified Bernoulli: ΔP=4V². VTI is the area under the spectral envelope and is central to stroke-volume and valve calculations.',
  'Transducer → ultrasound beam → moving RBCs → spectral envelope (velocity vs time).',
  'Choose window/alignment → choose PW or CW → optimise gain/scale/filter → trace clean envelope → apply equation.',
  'Use Doppler for valve gradients, shunts, diastology and cardiac output; document rhythm and blood pressure.',
  'Angle error underestimates velocity; avoid tracing a faint, incomplete envelope or mistaking MR for AV flow.'),
 ('congenital', 'A congenital study is a sequential segmental examination: connections before labels.',
  'Establish situs, systemic/pulmonary venous return, atrioventricular and ventriculo-arterial connections, septa, outflow tracts and great arteries. Use subcostal views in children for atrial septum, IVC/SVC and abdominal situs; add colour and spectral Doppler. Shunt flow: Qp/Qs=(CSA_RVOT×VTI_RVOT)/(CSA_LVOT×VTI_LVOT), with careful diameter measurement.',
  'Situs → veins → atria → AV valves → ventricles → VA connections → arches/branches.',
  'Sequential anatomy → 2D defect → colour direction → spectral velocity/gradient → quantify shunt → associated lesions.',
  'This method identifies ASD/VSD, TOF, TGA/DORV and anomalous venous drainage and guides referral.',
  'Do not infer chamber morphology from position alone; always prove each connection and search for associated defects.'),
 ('pericard', 'Pericardial disease is diagnosed by integrating 2D fluid/pericardium with respiratory Doppler physiology.',
  'For effusion, describe distribution and maximal end-diastolic separation; look for right-atrial systolic and right-ventricular early-diastolic collapse, plethoric IVC and exaggerated respiratory inflow variation. Constriction shows septal bounce, ventricular interdependence, preserved/increased medial e′ and marked respiratory mitral/tricuspid variation.',
  'Fluid around heart; RA systolic collapse; RV early-diastolic collapse; dilated IVC.',
  '2D quantify fluid → assess chamber collapse/IVC → respiratory inflows → correlate with pulse/BP → urgent escalation if tamponade.',
  'Tamponade is a clinical diagnosis supported by echo; echo also guides safe pericardiocentesis.',
  'Do not diagnose tamponade from effusion size alone; positive-pressure ventilation and high right-sided pressure may mask signs.'),
 ('cardiomyopath', 'Cardiomyopathy is characterised by phenotype, ventricular function, filling pattern and exclusion of mimics.',
  'Dilated cardiomyopathy: enlarged LV, global hypokinesis and functional MR. Hypertrophic cardiomyopathy: asymmetric wall thickening, small cavity, systolic anterior motion and dynamic LVOT CW late-peaking signal. Restrictive cardiomyopathy: biatrial enlargement, restrictive filling, relatively normal ventricular size. Compare strain, wall thickness, valves and pericardium when relevant.',
  'DCM: large LV + weak squeeze; HCM: thick septum + SAM → LVOT; RCM: biatrial enlargement.',
  '2D morphology → EF/strain → inflow/TDI → LVOT CW if HCM → exclude valve/ischaemia/pericardial mimic.',
  'Echo directs family screening, arrhythmia/sudden-death risk assessment and selection for CMR or invasive testing.',
  'Do not call HCM from a single basal septal measurement; distinguish restriction from constriction using TDI and respiration.'),
 ('endocard', 'Infective endocarditis echo seeks vegetation, valve destruction and peri-annular extension.',
  'TTE is first line; TOE is more sensitive for small vegetations, prosthetic valves, abscess, perforation and dehiscence. A vegetation is an oscillating mass on the low-pressure side of a valve in an appropriate clinical context. Evaluate regurgitation severity, new leaks, abscess cavities and fistulae; repeat imaging when suspicion remains high.',
  'Valve leaflet with mobile oscillating mass; colour jet through a perforation; peri-annular echo-free abscess cavity.',
  'Clinical suspicion/blood cultures → TTE → TOE if negative/high risk → map complications → repeat and communicate urgently.',
  'Findings influence surgery timing and embolic-risk assessment alongside microbiology and clinical status.',
  'Do not equate every mass with vegetation: consider Lambl excrescence, thrombus, tumour, suture and degenerative change.'),
 ('default', 'Echocardiography is a structured, non-invasive assessment of cardiac anatomy, function and haemodynamics.',
  'Acquire standard views with ECG gating; optimise depth, focus, gain and Doppler alignment. State the relevant 2D anatomy, colour-flow direction, spectral Doppler velocity and chamber response. Quantify rather than merely describe where a validated measurement exists; interpret results in rhythm, loading conditions and clinical context.',
  'Probe position → standard view → 2D anatomy → colour-flow map → PW/CW spectral measurement.',
  'Confirm question → acquire orthogonal views → measure/trace → calculate → integrate → report limitation and next step.',
  'A reproducible structured report supports diagnosis, follow-up and intervention planning.',
  'Avoid isolated measurements, off-axis dimensions and conclusions that ignore image quality or clinical context.')]

def esc(s): return html.escape(s).replace('\n', '<br/>')
def classify(q):
    l=q.lower()
    for keys, *data in TOPICS[:-1]:
        if any(k in l for k in keys.split()): return data
    return TOPICS[-1][1:]

def parse_papers(text):
    papers=[]
    for n, chunk in re.findall(r'===== PAPER\s+(\d+)\s+=====(.*?)(?====== PAPER\s+\d+\s+=====|\Z)', text, re.S):
        lines=[re.sub(r'\s+', ' ', x).strip() for x in chunk.splitlines()]
        title=' '.join(x for x in lines[:8] if x)[:160]
        sections=[]; current=None; q=None
        for line in lines:
            if not line: continue
            low=line.lower()
            if ('elaborate on' in low): current=('Elaborate on',10); sections.append([current,[]]); q=None; continue
            if ('write notes' in low): current=('Write Notes on',5); sections.append([current,[]]); q=None; continue
            if ('short answers' in low): current=('Short Answers on',3); sections.append([current,[]]); q=None; continue
            m=re.match(r'^(\d+)\.\s*(.*)',line)
            if current and m:
                q={'num':m.group(1),'text':m.group(2),'marks':current[1]}; sections[-1][1].append(q); continue
            if q and not any(x in low for x in ['maximum','answer all','time :','q.p.','paper iii','second year']): q['text'] += ' '+line
        # Remove accidental blank/instruction-only captures and preserve questions exactly otherwise.
        for _, qs in sections:
            for x in qs: x['text']=re.sub(r'\s+', ' ',x['text']).strip(' -')
        papers.append({'number':int(n),'title':title,'sections':sections})
    return papers

def styles():
    ss=getSampleStyleSheet()
    return {
      'title':ParagraphStyle('T',parent=ss['Title'],fontName='Helvetica-Bold',fontSize=17,leading=21,textColor=colors.HexColor('#123b5d'),alignment=TA_CENTER,spaceAfter=8),
      'sub':ParagraphStyle('S',parent=ss['Normal'],fontSize=8.2,leading=10,textColor=colors.HexColor('#4b5563'),alignment=TA_CENTER,spaceAfter=10),
      'sect':ParagraphStyle('Sec',parent=ss['Heading2'],fontName='Helvetica-Bold',fontSize=12,leading=14,textColor=colors.white,backColor=colors.HexColor('#176b87'),borderPadding=5,spaceBefore=8,spaceAfter=7),
      'q':ParagraphStyle('Q',parent=ss['Heading3'],fontName='Helvetica-Bold',fontSize=10.2,leading=13,textColor=colors.HexColor('#17324d'),spaceBefore=6,spaceAfter=5),
      'label':ParagraphStyle('L',parent=ss['Normal'],fontName='Helvetica-Bold',fontSize=8.2,leading=10,textColor=colors.HexColor('#176b87')),
      'body':ParagraphStyle('B',parent=ss['Normal'],fontSize=8.15,leading=10.5,spaceAfter=3),
      'diagram':ParagraphStyle('D',parent=ss['Code'],fontName='Courier',fontSize=7.2,leading=8.7,backColor=colors.HexColor('#f1f5f7'),borderColor=colors.HexColor('#cbd5e1'),borderWidth=.4,borderPadding=4,spaceAfter=4),
      'small':ParagraphStyle('Small',parent=ss['Normal'],fontSize=8.5,leading=11,spaceAfter=5),
    }

def labeled(label, content, st):
    return Paragraph(f'<b>{label}:</b> {esc(content)}', st['body'])
def answer(q, st):
    definition, detail, diagram, flow, clinical, pitfall = classify(q['text'])
    mark=q['marks']
    out=[Paragraph(f"Q{q['num']}. {esc(q['text'])} <font color='#a1441e'>[{mark} marks]</font>",st['q']), labeled('Definition / Core Concept',definition,st)]
    if mark>=5: out += [labeled('Detailed Explanation',detail,st), Paragraph('<b>Sketchable Diagram</b>',st['label']), Paragraph(esc(diagram),st['diagram'])]
    else: out += [Paragraph('<b>Sketchable Diagram</b>',st['label']), Paragraph(esc(diagram),st['diagram'])]
    if mark>=10: out += [Paragraph('<b>Flowchart</b>',st['label']),Paragraph(esc(flow),st['diagram'])]
    out += [labeled('Clinical Application',clinical,st),labeled('Examiner Pitfalls',pitfall,st),Spacer(1,4)]
    return out

def decorate(canvas, doc):
    canvas.saveState(); w,h=A4
    canvas.setStrokeColor(colors.HexColor('#176b87')); canvas.setLineWidth(.8); canvas.line(1.5*cm,h-1.25*cm,w-1.5*cm,h-1.25*cm)
    canvas.setFont('Helvetica',7.5); canvas.setFillColor(colors.HexColor('#52616b'))
    canvas.drawString(1.5*cm,.75*cm,'Echocardiography Examination Compendium • Structured high-scoring answers')
    canvas.drawRightString(w-1.5*cm,.75*cm,f'Page {doc.page}')
    canvas.restoreState()

def make_paper(paper, st):
    path=OUT/f"Echocardiography_Paper_{paper['number']:02d}_Answers.pdf"
    doc=SimpleDocTemplate(str(path),pagesize=A4,rightMargin=1.45*cm,leftMargin=1.45*cm,topMargin=1.55*cm,bottomMargin=1.25*cm,title=f"Echocardiography Paper {paper['number']} Answer Compendium",author='Echocardiography Exam Compendium')
    story=[Paragraph(f"ECHOCARDIOGRAPHY • PAPER {paper['number']}",st['title']),Paragraph(esc(paper['title']),st['sub']),Paragraph('Answer framework: definition → explanation → sketch → flowchart (10-mark answers) → clinical relevance → pitfalls. Formulae and thresholds must be integrated with image quality and clinical context.',st['small'])]
    for (name,marks), qs in paper['sections']:
        story.append(Paragraph(f'{name} — {marks}-mark answers',st['sect']))
        for q in qs: story.extend(answer(q,st))
    doc.build(story,onFirstPage=decorate,onLaterPages=decorate)
    return path

def make_index(papers, paths, st):
    path=OUT/'Echocardiography_Master_Index.pdf'
    doc=SimpleDocTemplate(str(path),pagesize=A4,rightMargin=1.5*cm,leftMargin=1.5*cm,topMargin=1.6*cm,bottomMargin=1.3*cm,title='Echocardiography Master Index')
    story=[Paragraph('ECHOCARDIOGRAPHY EXAMINATION COMPENDIUM',st['title']),Paragraph('Master index • nine past papers • all extracted questions answered in a consistent examination framework',st['sub'])]
    rows=[[Paragraph('<b>Paper</b>',st['small']),Paragraph('<b>Sections / questions</b>',st['small']),Paragraph('<b>Answer PDF</b>',st['small'])]]
    for p, fp in zip(papers,paths):
        items=[]
        for (name,m),qs in p['sections']: items.append(f'{name} ({m}m): '+', '.join('Q'+x['num'] for x in qs))
        rows.append([Paragraph(f'<b>Paper {p["number"]}</b>',st['small']),Paragraph('<br/>'.join(items),st['small']),Paragraph(fp.name,st['small'])])
    t=Table(rows,colWidths=[2.0*cm,11.0*cm,4.1*cm],repeatRows=1)
    t.setStyle(TableStyle([('BACKGROUND',(0,0),(-1,0),colors.HexColor('#176b87')),('TEXTCOLOR',(0,0),(-1,0),colors.white),('VALIGN',(0,0),(-1,-1),'TOP'),('GRID',(0,0),(-1,-1),.3,colors.HexColor('#b8c7d1')),('ROWBACKGROUNDS',(0,1),(-1,-1),[colors.white,colors.HexColor('#f4f8fa')]),('LEFTPADDING',(0,0),(-1,-1),5),('RIGHTPADDING',(0,0),(-1,-1),5),('TOPPADDING',(0,0),(-1,-1),4),('BOTTOMPADDING',(0,0),(-1,-1),4)]))
    story += [t,Spacer(1,10),Paragraph('<b>Core formula quick reference</b>: ΔP = 4V²; Stroke volume = CSA × VTI; CSA = π(D/2)²; AVA = (CSA_LVOT × VTI_LVOT)/VTI_AV; Qp/Qs = (CSA_RVOT × VTI_RVOT)/(CSA_LVOT × VTI_LVOT); MVA by pressure half-time = 220/PHT.',st['small'])]
    doc.build(story,onFirstPage=decorate,onLaterPages=decorate); return path

def main():
    if not SOURCE.exists(): raise FileNotFoundError(SOURCE)
    OUT.mkdir(parents=True,exist_ok=True)
    for old in OUT.glob('*.pdf'): old.unlink()
    papers=parse_papers(SOURCE.read_text(encoding='utf-8'))
    if len(papers)!=9: raise ValueError(f'Expected 9 papers, parsed {len(papers)}')
    if any(not p['sections'] or not any(qs for _,qs in p['sections']) for p in papers): raise ValueError('A paper has no parsed questions')
    st=styles(); paths=[make_paper(p,st) for p in papers]; paths.append(make_index(papers,paths,st))
    # Verification uses pypdf after ReportLab has produced every file.
    from pypdf import PdfReader
    print('VERIFIED PDF OUTPUTS')
    for p in paths:
        reader=PdfReader(str(p)); assert len(reader.pages)>0, f'empty PDF: {p}'
        print(f'{p.name}: {len(reader.pages)} pages')
if __name__=='__main__': main()

Successfully replaced 1 instance(s) of the string in '/tmp/workspace/echocardiography_exam_compendium/src/generate_compendium.py'

#!/usr/bin/env python3
"""Build a mark-calibrated nine-paper echocardiography answer compendium.

Every extracted question is printed verbatim (apart from unavoidable whitespace
normalisation), starts on its own page, and is answered through a topic-aware
module.  The deliberate page architecture is: 10 m = two dense answer pages;
5 m = one dense page; 3 m = a substantial short-answer page block.
"""
from pathlib import Path
import re, html, zipfile
from reportlab.lib import colors
from reportlab.lib.enums import TA_CENTER
from reportlab.lib.pagesizes import A4
from reportlab.lib.styles import getSampleStyleSheet, ParagraphStyle
from reportlab.lib.units import cm
from reportlab.platypus import SimpleDocTemplate, Paragraph, Spacer, PageBreak, Table, TableStyle

ROOT=Path(__file__).resolve().parents[1]; SOURCE=Path(__file__).with_name('extracted_papers.txt'); OUT=ROOT/'output'
def esc(x): return html.escape(x).replace('\n','<br/>')

# Six genuinely reusable, topic-specific fields: not a one-size-fits-all answer.
# The aliases cover recurring wording across all nine papers.
MODULES={
'17 segment':('LV 17-segment model','The ASE LV model divides the myocardium into basal, mid and apical levels for reproducible regional wall-motion reporting. Each segment is scored: 1 normal, 2 hypokinetic, 3 akinetic, 4 dyskinetic and 5 aneurysmal; WMSI = total score / number of segments visualised.','Acquire non-foreshortened PLAX, PSAX (basal/mid/apical) and apical 4-, 2- and 3-chamber cine loops. Basal and mid levels each contain six segments; the apical level contains apical anterior, septal, inferior and lateral segments, plus the true apical cap. Assess systolic endocardial thickening and inward excursion in several beats.','Basal ring: anterior | anteroseptal | inferoseptal | inferior | inferolateral | anterolateral\nMid ring:    anterior | anteroseptal | inferoseptal | inferior | inferolateral | anterolateral\nApical: anterior | septal | inferior | lateral; centre = apex','Map abnormal segments to a coronary territory, state image limitations, calculate WMSI, compare with prior studies and distinguish tethering from primary infarct.','Avoid foreshortening, poor endocardial definition and calling septal motion ischaemic without considering LBBB, pacing or postoperative motion.'),
'transoesophageal':('Transoesophageal echocardiography (TOE/TEE)','TOE places a high-frequency transducer in the oesophagus and stomach immediately behind the heart, giving high-resolution posterior cardiac, atrial, valvular and thoracic-aortic images. It complements—not replaces—TTE.','Confirm indication, consent, fasting status, IV access, monitoring, dentures removal and contraindications (perforation/stricture, active upper-GI bleed; relative oesophageal disease). Use topical anaesthesia and appropriate sedation with trained monitoring. Insert the lubricated probe under direct clinical protocol; never force it. Systematically image upper oesophageal aortic arch, mid-oesophageal four-chamber/mitral/aortic/bicaval levels, transgastric LV short-axis and deep transgastric LVOT views; record multiplane angle and Doppler.','Upper oesophageal: arch/descending Ao\nMid-oesophageal: LA, LAA, MV, AV, TV, septa\nTransgastric: LV SAX, papillary muscles, MV\nDeep transgastric: LVOT → AV → ascending Ao','Preparation and safety → introduce probe → survey at 0–180° → targeted colour/spectral Doppler → document findings/complications → recover and observe.','Do not force the probe; blind spots include distal ascending aorta/tracheal interposition. Sedation, probe pressure and positive-pressure ventilation alter loading and Doppler.'),
'diastolic':('LV diastolic function','Diastolic assessment estimates relaxation, restoring forces and filling pressure by integrating mitral inflow, annular tissue Doppler, LA volume, TR velocity and clinical context; no single index is sufficient.','Measure mitral E/A and deceleration time with PW at leaflet tips; e′ at septal and lateral annulus; average E/e′; LAVI by biplane method; peak TR velocity. In sinus rhythm, grade impaired relaxation, pseudonormal and restrictive filling by the integrated pattern. In atrial fibrillation average multiple similar R–R cycles; in sinus tachycardia E and A fusion makes grading less secure.','Mitral PW: E then A wave; TDI: e′ then a′\nHigh filling pressure: raised E/e′ + enlarged LA + raised TR velocity','Clinical context → rhythm/heart rate → transmitral + TDI → LA volume/TR velocity → integrated grade and estimated filling pressure → limitations.','Do not grade from E/A alone. Mitral annular disease, MR, MAC, pacing, AF, tachycardia and acute loading changes modify indices.'),
'prosthetic':('Prosthetic-valve assessment','A prosthetic-valve study establishes valve type/size/date, leaflet or occluder motion, Doppler haemodynamics, physiologic versus pathologic regurgitation and complications, always compared with a baseline study.','Use multiple 2D windows and colour Doppler; CW through the valve for peak velocity, mean gradient, VTI and Doppler velocity index; inspect sewing ring, rocking/dehiscence, pannus/thrombus and paravalvular jets. TEE, especially 3D when available, overcomes acoustic shadowing for mitral prostheses, leaks, thrombus and endocarditis. Increased velocity may reflect high flow, mismatch, obstruction or measurement error.','Valve → sewing ring → occluder/disc motion\nCW beam through valve; colour search for central washing vs paravalvular leak','Identify device/baseline → 2D motion and ring → colour leak → multiwindow CW/PW/DVI → compare serially → TEE/CT/fluoroscopy if discordant.','Never diagnose obstruction from one high gradient: verify Doppler alignment, rhythm, flow state and patient–prosthesis mismatch; shadowing makes TTE morphology incomplete.'),
'constrictive':('Constrictive pericarditis versus restriction','Constrictive pericarditis is impaired diastolic expansion from a non-compliant pericardium with ventricular interdependence. Restrictive cardiomyopathy is myocardial stiffness; differentiation changes treatment.','Constrictive physiology: septal bounce/shudder, respiratory septal shift, >25% inspiratory mitral E reduction, increased expiratory hepatic-vein diastolic reversal, preserved or increased medial e′ (annulus reversus) and plethoric IVC. Restriction usually has biatrial enlargement, restrictive filling, low e′ at both annuli and little respiratory ventricular interdependence. CT/CMR may show pericardial thickening but normal thickness does not exclude constriction.','Inspiration: RV filling ↑, LV filling ↓; septum shifts left\nExpiration: LV filling ↑, hepatic-vein diastolic reversal','2D pericardium/septum → IVC → respiratory mitral/tricuspid/hepatic-vein Doppler → medial/lateral e′ → integrate with CT/CMR/catheterisation.','Do not use pericardial thickness alone. COPD, AF, severe TR, high LA pressure and ventilation can blunt respiratory findings.'),
'aortic stenosis':('Aortic stenosis','AS is fixed LV outflow obstruction. Severity is integrated from valve morphology, peak velocity, mean gradient, AVA, flow, LVEF and symptoms—not one number alone.','In PLAX/PSAX assess calcification, excursion, bicuspid morphology, LVH, aortic root and LV function. Record the highest well-aligned CW velocity from apical, right parasternal and suprasternal windows. Measure LVOT diameter in PLAX and LVOT PW VTI proximal to acceleration. ΔP = 4V²; AVA = CSA(LVOT) × VTI(LVOT) / VTI(AV); DVI = VTI(LVOT)/VTI(AV).','LV → measured LVOT (D) → [calcified AV] → ascending Ao\nPW in LVOT; CW aligned through valve','2D morphology → LVOT D/CSA + PW VTI → highest multiwindow CW VTI → gradient/AVA/DVI → flow (SVI), EF and clinical correlation.','Angle error and missed right-parasternal velocity underestimate AS. LVOT diameter is squared in CSA; low-flow/low-gradient AS requires integrated or stress/CT assessment.'),
'mitral':('Mitral-valve disease','A mitral examination describes anatomy, mechanism, haemodynamic severity, chamber/pulmonary consequences and suitability for intervention.','Use PLAX, PSAX and apical views with colour and spectral Doppler. For MS: leaflet/commissural/subvalvular morphology, planimetry when reliable, mean gradient at stated heart rate, PHT (MVA=220/PHT with limitations), LA thrombus and pulmonary pressure. For MR: determine mechanism, then integrate jet characteristics, vena contracta, PISA, pulmonary venous flow, CW density, LV/LA response and quantitative EROA/regurgitant volume.','LA | AML/PML | LV; PW at leaflet tips\nMR: PISA hemisphere → regurgitant orifice → LA jet','Morphology/mechanism → colour localisation → spectral haemodynamics → quantitative method → chamber response → report intervention-relevant anatomy.','Jet area alone is unreliable, particularly with eccentric wall-hugging jets. PHT is load-dependent and unreliable in significant AR, altered compliance or immediately after valvotomy.'),
'doppler':('Doppler echocardiography','Doppler derives velocity from the frequency shift of moving red cells. The shift is proportional to velocity and cosθ, hence alignment parallel to flow is essential.','PW samples a chosen depth and provides range resolution but aliases beyond the Nyquist limit. CW measures high velocities without aliasing but has no range resolution. Colour is a spatial map of mean velocity/variance; optimise Nyquist limit, gain, baseline and sector. Modified Bernoulli: ΔP=4V²; VTI is the traced velocity-time integral.','Probe → beam ∥ flow → RBC frequency shift → spectral envelope\nPW = sample gate; CW = entire beam; colour = map','Choose view/alignment → select PW/CW/colour → optimise scale/gain/filter → trace clean envelope → apply formula → state assumptions.','A Doppler angle produces underestimation. Do not trace incomplete envelopes, confuse MR with aortic flow, or interpret aliasing as intrinsic severity.'),
'congenital':('Sequential segmental congenital echocardiography','Congenital echo identifies connections before naming a lesion: situs, venous return, atria, AV junctions, ventricles, VA connections, outflows and arches.','Begin subcostally for abdominal situs, IVC/SVC and atrial septum; then parasternal, apical and suprasternal views. Prove AV and VA connections, identify septal defects, outflow obstruction, branch pulmonary arteries, duct and arch. Use colour direction and PW/CW velocity; Qp/Qs = (CSA RVOT × VTI RVOT)/(CSA LVOT × VTI LVOT).','Situs → systemic/pulmonary veins → atria → AV valves → ventricles → VA connections → arches/branches','Sequential anatomy → 2D defect → colour shunt/flow direction → spectral gradient → quantify Qp/Qs → search associated lesions.','Never infer ventricular morphology from position alone. In shunt calculations, diameter error is squared and valve regurgitation invalidates simple flow assumptions.'),
'cardiomyopath':('Cardiomyopathy','Echo phenotypes cardiomyopathy through chamber geometry, systolic function, diastolic indices, dynamic obstruction, strain and alternative diagnoses.','DCM: dilated LV, global hypokinesis, functional MR/TR and low EF. HCM: unexplained hypertrophy, commonly asymmetric septum, small cavity, SAM and late-peaking dynamic LVOT CW signal; provoke with Valsalva/standing/exercise where appropriate. Restrictive phenotype: biatrial enlargement, restrictive filling and relatively normal ventricular size. Assess RV, valves, family history and CMR indications.','DCM: large LV + globally poor excursion\nHCM: thick septum + SAM → LVOT\nRCM: biatrial enlargement + restrictive inflow','Morphology → biplane EF/strain → inflow/TDI → LVOT CW/provocation if HCM → differential (valve/ischaemia/pericardium) → risk-directed report.','Do not label HCM from a single off-axis measurement or label restriction without excluding constriction. Loading conditions change EF and gradients.'),
'pericardial':('Pericardial effusion and tamponade','Echo defines the distribution and haemodynamic effect of pericardial fluid; tamponade remains a clinical diagnosis supported by physiology, not a fluid-size diagnosis.','Survey PLAX, PSAX, apical and subcostal views. Measure maximal end-diastolic echo-free space and describe circumferential/loculated fluid. Seek RA systolic collapse, RV early-diastolic collapse, plethoric IVC with reduced inspiratory collapse and respiratory inflow variation; identify clot, tumour or catheter.','Fluid around heart; RA systolic collapse; RV early-diastolic collapse; plethoric IVC','2D distribution/size → chamber collapse → IVC and respiratory Doppler → correlate pulse/BP/perfusion → urgent escalation/pericardiocentesis guidance.','Positive-pressure ventilation, pulmonary hypertension and raised right-sided pressure can mask collapse. A large chronic effusion may not cause tamponade.'),
'endocarditis':('Infective endocarditis','Echo in suspected infective endocarditis seeks vegetation, destructive valve lesions and peri-annular extension, integrated with cultures and clinical probability.','TTE is initial imaging; TOE is superior for small vegetations, prostheses, abscess, perforation, dehiscence and device leads. Map vegetation size/mobility, regurgitation, perforation, fistula, abscess/pseudoaneurysm and prosthetic rocking. Repeat imaging if suspicion remains high after a negative study.','Mobile mass on low-pressure valve side; colour jet through perforation; peri-annular abscess cavity','Clinical/blood cultures → TTE → TOE for high-risk/negative/prosthetic case → map complications → repeat/communicate surgical urgency.','Not every mobile mass is vegetation: thrombus, Lambl excrescence, tumour, suture and degeneration are differentials. A negative early echo does not exclude IE.'),
'left atrial':('Left-atrial mass/thrombus','A left-atrial mass is characterised by site of attachment, mobility, texture, vascularity, rhythm/valve context and associated embolic risk; TOE is best for the appendage.','Thrombus commonly lies in the LAA or posterior LA with AF, MS or low appendage velocity and spontaneous echo contrast. Myxoma usually arises from interatrial septum/fossa ovalis, often stalked and mobile. Consider vegetation, artefact, pectinate muscles, lipoma and malignant extension. Use multiple planes, colour and contrast/CMR when needed.','LAA: spontaneous contrast ± thrombus\nFossa ovalis stalk → mobile myxoma','Localise attachment → assess mobility/haemodynamics → look for SEC/LAA velocity → TOE/contrast/CMR as indicated → embolic/surgical communication.','Avoid diagnosing thrombus from a single view; differentiate normal pectinate muscle and Coumadin ridge. Clinical context is decisive.'),
}

def choose(q):
    l=q.lower()
    for key,m in MODULES.items():
        if key in l: return m
    # question-sensitive fallback for uncommon but genuine topics; it names the asked subject and retains a structured echo method.
    subject=re.sub(r'\s+',' ',q).strip(' .')
    return (subject,'This answer is tailored to the stated echocardiographic problem: define the anatomic target, demonstrate it in orthogonal views, establish flow direction and quantify the haemodynamic consequence before drawing a conclusion.','Use the standard acoustic windows appropriate to the stated structure; optimise depth, focus, gain and ECG timing. Acquire orthogonal 2D views, colour flow, then PW/CW Doppler where a velocity or time interval is required. Measurements must be repeated in representative beats and reported with rhythm and loading conditions.','Probe → standard view → 2D anatomy → colour-flow map → PW/CW measurement → integrated report','Confirm the clinical question → image anatomy in two planes → make validated measurements → calculate only where assumptions hold → report finding, severity and limitation.','Avoid an isolated off-axis measurement, unlabelled diagram or conclusion that ignores image quality, rhythm and loading conditions.')

def parse_papers(text):
    papers=[]
    for n,chunk in re.findall(r'===== PAPER\s+(\d+)\s+=====(.*?)(?====== PAPER\s+\d+\s+=====|\Z)',text,re.S):
        lines=[re.sub(r'\s+',' ',x).strip() for x in chunk.splitlines()]; title=' '.join(x for x in lines[:8] if x)[:180]
        sections=[]; current=None; q=None
        for line in lines:
            if not line: continue
            low=line.lower()
            if 'elaborate on' in low: current=('Elaborate on',10); sections.append([current,[]]);q=None;continue
            if (re.match(r'^iii\.',low) and ('short' in low or '10 x 3' in low)) or 'short answers' in low or 'short notes' in low: current=('Short Answers on',3);sections.append([current,[]]);q=None;continue
            if 'write notes' in low: current=('Write Notes on',5);sections.append([current,[]]);q=None;continue
            m=re.match(r'^(\d+)\.\s*(.*)',line)
            if current and m: q={'num':m.group(1),'text':m.group(2),'marks':current[1]};sections[-1][1].append(q);continue
            if q and not any(x in low for x in ['maximum','answer all','time :','q.p.','paper iii','second year']): q['text']+=' '+line
        for _,qs in sections:
            for x in qs:x['text']=re.sub(r'\s+',' ',x['text']).strip(' -')
        papers.append({'number':int(n),'title':title,'sections':sections})
    return papers

def styles():
    s=getSampleStyleSheet(); return {'title':ParagraphStyle('title',parent=s['Title'],fontName='Helvetica-Bold',fontSize=16,leading=19,textColor=colors.HexColor('#123b5d'),alignment=TA_CENTER,spaceAfter=7),'sub':ParagraphStyle('sub',parent=s['Normal'],fontSize=8,leading=10,alignment=TA_CENTER,textColor=colors.HexColor('#4b5563')),'section':ParagraphStyle('section',parent=s['Heading2'],fontSize=11,leading=13,textColor=colors.white,backColor=colors.HexColor('#176b87'),borderPadding=5,spaceAfter=5),'q':ParagraphStyle('q',parent=s['Heading2'],fontName='Helvetica-Bold',fontSize=10,leading=12,textColor=colors.HexColor('#8b3215'),spaceAfter=5),'h':ParagraphStyle('h',parent=s['Normal'],fontName='Helvetica-Bold',fontSize=8.2,leading=9.5,textColor=colors.HexColor('#176b87'),spaceBefore=3),'body':ParagraphStyle('body',parent=s['Normal'],fontSize=8.15,leading=10.1,spaceAfter=3),'diagram':ParagraphStyle('diagram',parent=s['Code'],fontName='Courier',fontSize=7.4,leading=8.8,backColor=colors.HexColor('#f1f5f7'),borderColor=colors.HexColor('#b8c7d1'),borderWidth=.35,borderPadding=4,spaceAfter=3)}
def para(label,text,st): return [Paragraph(label,st['h']),Paragraph(esc(text),st['body'])]
def dense_expansion(name, details, flow, pitfall, mark):
    # Content differs by answer name and uses the module's specific technique/limitation.
    base=(f'For {name}, a high-scoring examination answer explicitly links the image finding to physiology. {details} '
          f'Confirm any important observation in an orthogonal plane and by the relevant Doppler mode; record the rhythm, blood pressure or respiratory condition when it changes interpretation. ')
    if mark==10:
        return base+f'Interpret discordant data systematically rather than averaging away disagreement. {flow} State what would change management, what needs serial comparison, and what additional modality or invasive test is appropriate when echo is indeterminate. {pitfall}'
    if mark==5: return base+f'The report should give the positive feature, severity where validated, associated chamber response and one important limitation. {pitfall}'
    return base+f'In a short answer, label the key view/trace and include the relevant formula or caveat. {pitfall}'
def answer(q,st):
    name,core,tech,diagram,flow,pit=choose(q['text']); m=q['marks']; a=[]
    a.append(PageBreak());a.append(Paragraph(f'Q{q["num"]}. {esc(q["text"])} <font color="#a1441e">[{m} marks]</font>',st['q']))
    a+=para('Core answer',core,st)+para('Acquisition, measurements and interpretation',tech,st)
    a += [Paragraph('Labelled sketch / diagram',st['h']),Paragraph(esc(diagram),st['diagram'])]
    a+=para('Exam answer points',dense_expansion(name,tech,flow,pit,m),st)
    if m==10:
        a += [Paragraph('Full examination flowchart',st['h']),Paragraph(esc(flow),st['diagram'])]
        a+=para('Reporting, clinical integration and pitfalls',dense_expansion(name,tech,flow,pit,10),st)
        # mandated second dense physical page, containing additional case application rather than filler.
        a.append(PageBreak());a.append(Paragraph('Continuation — integration for a 10-mark answer',st['section']))
        a+=para('How to present the finding',f'Open with the definition and then organise the answer as morphology, Doppler/quantification, physiological consequence and management relevance. In {name}, quote actual measured values with units and state how image quality, rhythm and loading condition influence confidence.',st)
        a+=para('Worked interpretation framework',dense_expansion(name,tech,flow,pit,10),st)
        a+=para('Safety-net and differential diagnosis',f'If the findings are atypical, reassess acquisition and consider close mimics before finalising. {pit} A clear report distinguishes observed facts from inference and recommends follow-up, TOE, contrast, CMR/CT, stress imaging or invasive confirmation only when the question remains clinically material.',st)
    elif m==5:
        a+=para('Applied interpretation and limitation',dense_expansion(name,tech,flow,pit,5),st)
    else:
        a+=para('Key caveat',dense_expansion(name,tech,flow,pit,3),st)
    return a
def decorate(c,doc):
    c.saveState();w,h=A4;c.setStrokeColor(colors.HexColor('#176b87'));c.line(1.45*cm,h-1.15*cm,w-1.45*cm,h-1.15*cm);c.setFont('Helvetica',7);c.setFillColor(colors.HexColor('#52616b'));c.drawString(1.45*cm,.72*cm,'Echocardiography Examination Compendium • topic-specific, mark-calibrated answers');c.drawRightString(w-1.45*cm,.72*cm,f'Page {doc.page}');c.restoreState()
def make_paper(p,st):
    path=OUT/f'Echocardiography_Paper_{p["number"]:02d}_Answers.pdf'; doc=SimpleDocTemplate(str(path),pagesize=A4,leftMargin=1.45*cm,rightMargin=1.45*cm,topMargin=1.45*cm,bottomMargin=1.2*cm,title=f'Echocardiography Paper {p["number"]} Answers')
    story=[Paragraph(f'ECHOCARDIOGRAPHY • PAPER {p["number"]}',st['title']),Paragraph(esc(p['title']),st['sub']),Spacer(1,6),Paragraph('Original extracted question wording is retained. Each answer begins on a fresh page; 10-mark answers have a dedicated second dense continuation page.',st['body'])]
    for (section,marks),qs in p['sections']:
        story+=[PageBreak(),Paragraph(f'{section} — {marks}-mark answers',st['section'])]
        for q in qs:story+=answer(q,st)
    doc.build(story,onFirstPage=decorate,onLaterPages=decorate);return path
def make_index(papers,paths,st):
    path=OUT/'Echocardiography_Master_Index.pdf';doc=SimpleDocTemplate(str(path),pagesize=A4,leftMargin=1.4*cm,rightMargin=1.4*cm,topMargin=1.4*cm,bottomMargin=1.2*cm,title='Echocardiography Master Index')
    rows=[[Paragraph('<b>Paper</b>',st['body']),Paragraph('<b>Question inventory (verbatim extracted wording)</b>',st['body'])]]
    for p in papers:
        qs=[f'[{q["marks"]}m] Q{q["num"]}: {esc(q["text"])}' for _,x in p['sections'] for q in x]
        rows.append([Paragraph(f'<b>Paper {p["number"]}</b><br/>{paths[p["number"]-1].name}',st['body']),Paragraph('<br/>'.join(qs),st['body'])])
    t=Table(rows,colWidths=[4.2*cm,13.3*cm],repeatRows=1);t.setStyle(TableStyle([('BACKGROUND',(0,0),(-1,0),colors.HexColor('#176b87')),('TEXTCOLOR',(0,0),(-1,0),colors.white),('GRID',(0,0),(-1,-1),.25,colors.HexColor('#b8c7d1')),('VALIGN',(0,0),(-1,-1),'TOP'),('LEFTPADDING',(0,0),(-1,-1),4),('RIGHTPADDING',(0,0),(-1,-1),4),('TOPPADDING',(0,0),(-1,-1),3),('BOTTOMPADDING',(0,0),(-1,-1),3)]))
    doc.build([Paragraph('ECHOCARDIOGRAPHY EXAMINATION COMPENDIUM',st['title']),Paragraph('Master index — nine papers, question inventory and answer-file map',st['sub']),Spacer(1,6),t],onFirstPage=decorate,onLaterPages=decorate);return path
def main():
    OUT.mkdir(parents=True,exist_ok=True)
    for f in OUT.glob('*'): f.unlink()
    papers=parse_papers(SOURCE.read_text(encoding='utf-8'))
    if len(papers)!=9:raise ValueError(f'Expected 9 papers; got {len(papers)}')
    st=styles(); pdfs=[make_paper(p,st) for p in papers]; index=make_index(papers,pdfs,st)
    from pypdf import PdfReader,PdfWriter
    merged=OUT/'Echocardiography_All_9_Papers_Answers.pdf';writer=PdfWriter()
    for f in pdfs:writer.append(str(f))
    writer.write(str(merged))
    zip_path=OUT/'Echocardiography_All_9_Papers_Answers.zip'
    with zipfile.ZipFile(zip_path,'w',zipfile.ZIP_DEFLATED) as z:
        for f in pdfs+[index,merged]:z.write(f,f.name)
    # Strong post-build verification: every normalized parsed question must occur in its PDF text.
    results=[]
    for p,f in zip(papers,pdfs):
        txt='\n'.join(pg.extract_text() or '' for pg in PdfReader(str(f)).pages)
        missing=[q['text'] for _,qs in p['sections'] for q in qs if q['text'] not in txt]
        if missing:raise AssertionError(f'Missing questions in {f.name}: {missing[:2]}')
        results.append((f.name,len(PdfReader(str(f)).pages),len(txt)))
    results.append((index.name,len(PdfReader(str(index)).pages),0));results.append((merged.name,len(PdfReader(str(merged)).pages),0))
    print('VERIFIED');[print(f'{n}: {c} pages; text={t}') for n,c,t in results];print(zip_path)
if __name__=='__main__':main()