cd /home/daytona/workspace && python3 -c " import pdfplumber pdf_path = '/home/daytona/workspace/attachments/a68b12aa-0d33-4755-8240-4ff744f3c30e/BMJ OnExamination Cardiology 2025 part 2-compressed.pdf' with pdfplumber.open(pdf_path) as pdf: print(f'Total pages: {len(pdf.pages)}') # Show first 3 pages to understand structure for i in range(min(3, len(pdf.pages))): print(f'--- PAGE {i+1} ---') text = pdf.pages[i].extract_text() print(text[:2000] if text else 'No text') print() "

cd /home/daytona/workspace && python3 << 'EOF' import pdfplumber import json import re pdf_path = '/home/daytona/workspace/attachments/a68b12aa-0d33-4755-8240-4ff744f3c30e/BMJ OnExamination Cardiology 2025 part 2-compressed.pdf' mcqs = [] current_q = None with pdfplumber.open(pdf_path) as pdf: total = len(pdf.pages) print(f"Total pages: {total}") for i in range(1, total): # skip page 0 (no text) text = pdf.pages[i].extract_text() if not text: continue # Detect question page (has "Answer question" at end) if 'Answer question' in text and 'Question' in text: # Extract question number m = re.search(r'Question (\d+) of 200', text) if m: qnum = int(m.group(1)) # Extract question text (between question line and options) lines = text.split('\n') # Find the question body q_text = '' options = [] in_q = False for line in lines: line = line.strip() if f'Question {qnum} of 200' in line: in_q = True continue if 'Answer question' in line or 'Time elapsed' in line or 'My score' in line: break if in_q and line and line not in ['Exit session', 'English French', 'Question navigator', 'Add tags']: if line not in ['Previous', 'Next']: q_text += line + ' ' current_q = {'num': qnum, 'question': q_text.strip(), 'page': i+1} # Detect answer/explanation page elif 'Key learning points' in text and current_q: lines = text.split('\n') explanation = '' in_exp = False topics = '' for j, line in enumerate(lines): line = line.strip() if 'Key learning points' in line: # Next line(s) are topics if j+1 < len(lines): topics = lines[j+1].strip() in_exp = False elif 'Explanation' in line and not in_exp: in_exp = True continue elif in_exp: if 'Reference:' in line or 'Next question' in line or 'Add revision notes' in line: break if line: explanation += line + ' ' current_q['topics'] = topics current_q['explanation'] = explanation.strip() mcqs.append(current_q) current_q = None print(f"Extracted {len(mcqs)} MCQs") # Save to file with open('/home/daytona/workspace/mcqs_extracted.json', 'w') as f: json.dump(mcqs, f, indent=2) print("Saved to mcqs_extracted.json") # Show first 3 for q in mcqs[:3]: print(f"\nQ{q['num']}: {q['question'][:200]}") print(f"Topics: {q.get('topics','')}") print(f"Explanation: {q.get('explanation','')[:300]}") EOF

python3 << 'EOF' from pypdf import PdfReader import json, re, sys pdf_path = '/home/daytona/workspace/attachments/a68b12aa-0d33-4755-8240-4ff744f3c30e/BMJ OnExamination Cardiology 2025 part 2-compressed.pdf' # Process in batches to avoid memory issues reader = PdfReader(pdf_path) total = len(reader.pages) print(f"Total pages: {total}", flush=True) mcqs = [] current_q = None for i in range(1, total): try: text = reader.pages[i].extract_text() except: continue if not text: continue # Detect question page if 'Answer question' in text and 'Question' in text: m = re.search(r'Question (\d+) of 200', text) if m: qnum = int(m.group(1)) lines = [l.strip() for l in text.split('\n')] q_text = '' in_q = False skip = {'Exit session','English French','Question navigator','Add tags','Previous','Next','Answer question','Time elapsed','My score','0%','(0/0)'} for line in lines: if f'Question {qnum} of 200' in line: in_q = True continue if 'Answer question' in line or 'Time elapsed' in line: break if in_q and line and line not in skip and not line.startswith('Previous Question'): q_text += line + ' ' current_q = {'num': qnum, 'question': q_text.strip()} elif 'Key learning points' in text and current_q: lines = [l.strip() for l in text.split('\n')] explanation = '' topics = '' correct_ans = '' in_exp = False for j, line in enumerate(lines): if 'Key learning points' in line: if j+1 < len(lines): topics = lines[j+1].strip() elif 'Explanation' in line and not in_exp: in_exp = True continue elif in_exp: if 'Reference:' in line or 'Next question' in line or 'Add revision notes' in line: break if line and not line.startswith('%') and not re.match(r'^\d+%$', line): explanation += line + ' ' current_q['topics'] = topics current_q['explanation'] = explanation.strip() mcqs.append(current_q) if len(mcqs) % 20 == 0: print(f" Processed {len(mcqs)} MCQs...", flush=True) current_q = None print(f"\nTotal extracted: {len(mcqs)}") with open('/home/daytona/workspace/mcqs_extracted.json', 'w') as f: json.dump(mcqs, f, indent=2) print("Saved.") EOF

python3 << 'EOF' from pypdf import PdfReader import re pdf_path = '/home/daytona/workspace/attachments/a68b12aa-0d33-4755-8240-4ff744f3c30e/BMJ OnExamination Cardiology 2025 part 2-compressed.pdf' reader = PdfReader(pdf_path) # Sample pages 2-20 to understand structure for i in range(1, 25): try: text = reader.pages[i].extract_text() except: continue if text: first_100 = text[:120].replace('\n',' ') has_answer = 'Answer question' in text has_key = 'Key learning' in text has_correct = 'Correct' in text or 'Incorrect' in text print(f"Page {i+1}: ans={has_answer} key={has_key} correct={has_correct} | {first_100[:80]}") EOF

python3 << 'EOF' from pypdf import PdfReader import re, json pdf_path = '/home/daytona/workspace/attachments/a68b12aa-0d33-4755-8240-4ff744f3c30e/BMJ OnExamination Cardiology 2025 part 2-compressed.pdf' reader = PdfReader(pdf_path) total = len(reader.pages) # The pattern: # - "Answer question" page: has question + options (before answering) # - "Key learning points" page: has question stem repeated + answer + explanation # Strategy: Extract from the KEY LEARNING page (page with "Key learning points") # which has the full question + correct answer highlighted + explanation mcqs = [] noise = {'Exit session','English French','Question navigator','Add tags','Previous','Next', 'Answer question','Time elapsed','My score','Save notes','Rate this question', 'Improve this Question','You\'ll receive a summary'} for i in range(1, total): try: text = reader.pages[i].extract_text() except: continue if not text or 'Key learning points' not in text: continue m = re.search(r'Question (\d+) of 200', text) if not m: continue qnum = int(m.group(1)) lines = [l.strip() for l in text.split('\n') if l.strip()] # Find question stem - it appears after "Question N of 200" header lines # and before the answer options (which have percentages) q_start = -1 q_end = -1 exp_start = -1 topics_line = '' for j, line in enumerate(lines): if f'Question {qnum} of 200' in line and q_start == -1: q_start = j + 1 if q_start > 0 and q_end == -1: # Options have percentage pattern nearby or "Correct"/"Incorrect" before them if re.match(r'^\d+%$', line) or line in ('Correct', 'Incorrect'): q_end = j if 'Key learning points' in line: if j+1 < len(lines): topics_line = lines[j+1] if 'Explanation' in line and exp_start == -1: exp_start = j + 1 # Extract question text question = '' if q_start > 0: end = q_end if q_end > q_start else min(q_start+20, len(lines)) for line in lines[q_start:end]: if line in noise or re.match(r'^\d+%$', line) or line in ('Correct','Incorrect'): break if not any(x in line for x in ['Exit session','English French','Question navigator']): question += line + ' ' # Extract explanation explanation = '' if exp_start > 0: for line in lines[exp_start:]: if 'Reference:' in line or 'Next question' in line or 'Add revision notes' in line: break if line and not re.match(r'^\d+%$', line): explanation += line + ' ' mcqs.append({ 'num': qnum, 'question': question.strip(), 'topics': topics_line, 'explanation': explanation.strip() }) mcqs.sort(key=lambda x: x['num']) print(f"Extracted {len(mcqs)} MCQs") # Check for duplicates and missing nums = [q['num'] for q in mcqs] unique = sorted(set(nums)) print(f"Unique question numbers: {len(unique)}, range: {min(unique)}-{max(unique)}") missing = [n for n in range(1, 201) if n not in unique] print(f"Missing Q numbers: {missing[:20]}") with open('/home/daytona/workspace/mcqs_extracted.json', 'w') as f: json.dump(mcqs, f, indent=2) # Show sample for q in mcqs[:3]: print(f"\nQ{q['num']}: {q['question'][:150]}") print(f"Topics: {q['topics']}") print(f"Explanation: {q['explanation'][:200]}") EOF

python3 << 'EOF' from pypdf import PdfReader import re pdf_path = '/home/daytona/workspace/attachments/a68b12aa-0d33-4755-8240-4ff744f3c30e/BMJ OnExamination Cardiology 2025 part 2-compressed.pdf' reader = PdfReader(pdf_path) # Look at pages 24-50 to understand structure for Q7-Q15 for i in range(23, 60): try: text = reader.pages[i].extract_text() except: text = '' if text: has_answer = 'Answer question' in text has_key = 'Key learning' in text m = re.search(r'Question (\d+) of 200', text) qnum = m.group(1) if m else '?' first = text[:80].replace('\n',' ') print(f"P{i+1} Q{qnum} ans={has_answer} key={has_key}: {first[:70]}") else: print(f"P{i+1}: NO TEXT") EOF

python3 << 'EOF' from pypdf import PdfReader import sys pdf_path = '/home/daytona/workspace/attachments/a68b12aa-0d33-4755-8240-4ff744f3c30e/BMJ OnExamination Cardiology 2025 part 2-compressed.pdf' reader = PdfReader(pdf_path) total = len(reader.pages) print(f"Total pages: {total}") # Dump all text to file in chunks with open('/home/daytona/workspace/all_text.txt', 'w', encoding='utf-8') as f: for i in range(1, total): try: text = reader.pages[i].extract_text() if text: f.write(f"\n===PAGE_{i+1}===\n") f.write(text) except Exception as e: pass if (i+1) % 100 == 0: print(f"Processed {i+1}/{total}", flush=True) import os size = os.path.getsize('/home/daytona/workspace/all_text.txt') print(f"Done. File size: {size/1024:.0f} KB") EOF

python3 << 'PYEOF' import re, json with open('/home/daytona/workspace/all_text.txt', 'r', encoding='utf-8') as f: content = f.read() # Split into page blocks pages = re.split(r'\n===PAGE_(\d+)===\n', content) # pages[0] = empty, then alternating: page_num, text, page_num, text... page_dict = {} for i in range(1, len(pages), 2): pnum = int(pages[i]) ptext = pages[i+1] if i+1 < len(pages) else '' page_dict[pnum] = ptext print(f"Pages with text: {len(page_dict)}") # Strategy: find all pages that have "Key learning points" AND explanation content # The question stem appears on "Answer question" pages OR on the key learning pages # Explanations can span to next page (continuation pages) mcqs_raw = {} # key: question_id (qnum + set), value: dict # Identify all "Key learning points" pages - these have the answer explanation kl_pages = [(pnum, text) for pnum, text in sorted(page_dict.items()) if 'Key learning points' in text] print(f"Key learning pages: {len(kl_pages)}") # Collect explanation pages (continuation) for pnum, text in kl_pages: m_q = re.search(r'Question (\d+) of (\d+)', text) if not m_q: continue qnum = int(m_q.group(1)) qset = int(m_q.group(2)) key = f"{qset}_{qnum}" # Topics topics_m = re.search(r'Key learning points\s*\n([^\n]+)', text) topics = topics_m.group(1).strip() if topics_m else '' # Also check if next line after Key learning points has topic kl_idx = text.find('Key learning points') topic_area = text[kl_idx:kl_idx+200] topic_lines = [l.strip() for l in topic_area.split('\n') if l.strip()] topics = topic_lines[1] if len(topic_lines) > 1 else topics # Explanation from this page exp_start = text.find('Explanation') if exp_start == -1: exp_start = text.find('Key learning points') exp_text = text[exp_start:] if exp_start >= 0 else text # Clean up exp_lines = [l.strip() for l in exp_text.split('\n') if l.strip()] clean_exp = [] started = False for line in exp_lines: if 'Explanation' in line: started = True continue if started: if any(x in line for x in ['Reference:', 'Next question', 'Add revision notes', 'You\'ll receive']): break if re.match(r'^\d+%$', line): continue clean_exp.append(line) # Get question stem - look at "Answer question" pages for this qnum # Find nearby pages with same question number q_stem = '' for check_page in range(pnum-5, pnum+1): if check_page in page_dict: pg = page_dict[check_page] if f'Question {qnum} of {qset}' in pg and 'Answer question' in pg: # Extract question text lines = [l.strip() for l in pg.split('\n') if l.strip()] in_q = False stem_lines = [] skip_words = {'Exit session','English French','Question navigator','Add tags', 'Previous','Next','Answer question','Time elapsed','My score', 'Save notes','Rate this question','Improve this Question'} for ln in lines: if f'Question {qnum} of {qset}' in ln: in_q = True continue if in_q: if 'Answer question' in ln or ln in skip_words: break if ln and ln not in skip_words and not ln.startswith('Previous Question') and not re.match(r'^\d+%$|^\d+/\d+$', ln): stem_lines.append(ln) q_stem = ' '.join(stem_lines) break # Also try to get stem from the key learning page itself (before answer options) if not q_stem: # The key learning page shows: question header, then stem, then options with % lines = [l.strip() for l in text.split('\n') if l.strip()] in_q = False stem_lines = [] for ln in lines: if f'Question {qnum} of {qset}' in ln: in_q = True continue if in_q: if re.match(r'^\d+%$', ln) or ln in ('Correct','Incorrect'): break if ln and not any(x in ln for x in ['Exit session','English French','Question navigator']): stem_lines.append(ln) q_stem = ' '.join(stem_lines) # Check continuation page for more explanation cont_exp = '' next_pnum = pnum + 1 while next_pnum in page_dict: next_text = page_dict[next_pnum] if 'Key learning points' in next_text or 'Answer question' in next_text: break if any(x in next_text for x in ['Reference:', 'Add revision notes', 'Save notes', 'Question navigator']): # This might have more explanation text before the stop words cont_lines = [l.strip() for l in next_text.split('\n') if l.strip()] for ln in cont_lines: if any(x in ln for x in ['Reference:', 'Next question', 'Add revision notes', 'Save notes', 'Rate this', 'Improve this', 'Question navigator', 'You\'ll receive']): break if ln and not re.match(r'^\d+%$|^\d+\(\d+/\d+\)$', ln): cont_exp += ln + ' ' next_pnum += 1 if next_pnum - pnum > 3: break full_exp = ' '.join(clean_exp) + ' ' + cont_exp mcqs_raw[key] = { 'num': qnum, 'set_size': qset, 'question': q_stem.strip(), 'topics': topics, 'explanation': full_exp.strip(), 'page': pnum } all_mcqs = list(mcqs_raw.values()) all_mcqs.sort(key=lambda x: (x['set_size'], x['num'])) print(f"Total unique MCQs: {len(all_mcqs)}") # Count by set size from collections import Counter set_counts = Counter(q['set_size'] for q in all_mcqs) print(f"MCQ sets: {dict(sorted(set_counts.items()))}") with open('/home/daytona/workspace/mcqs_final.json', 'w') as f: json.dump(all_mcqs, f, indent=2) # Show sample of each set type shown = set() for q in all_mcqs[:5]: print(f"\nQ{q['num']}/{q['set_size']} [p{q['page']}]: {q['question'][:120]}") print(f" Topics: {q['topics']}") print(f" Exp: {q['explanation'][:150]}") PYEOF

python3 << 'PYEOF' import json with open('/home/daytona/workspace/mcqs_final.json') as f: mcqs = json.load(f) # Print all topics to understand what we're dealing with from collections import Counter topic_list = [] for q in mcqs: t = q.get('topics','').strip() if t: topic_list.append(t) topic_counts = Counter(topic_list) print("All topic combinations (sorted by frequency):") for combo, cnt in sorted(topic_counts.items(), key=lambda x: -x[1]): print(f" {cnt}x: {combo}") print(f"\nTotal MCQs: {len(mcqs)}") print("\nAll explanations (first 300 chars each):") for q in mcqs: print(f"\nQ{q['num']}/{q['set_size']} [{q['topics']}]") print(f" Q: {q['question'][:100]}") print(f" E: {q['explanation'][:200]}") PYEOF

mkdir -p /home/daytona/workspace/cardiology-notes && ls /home/daytona/workspace/cardiology-notes/

ls /home/daytona/workspace/ && ls /home/daytona/workspace/cardiology-notes/ 2>/dev/null && echo "mcqs file exists: $(test -f /home/daytona/workspace/mcqs_final.json && echo yes || echo no)"