import json, re, pdfplumber
from reportlab.lib.pagesizes import A4
from reportlab.lib.styles import getSampleStyleSheet, ParagraphStyle
from reportlab.lib.units import inch
from reportlab.lib import colors
from reportlab.platypus import SimpleDocTemplate, Paragraph, Spacer, Table, TableStyle, HRFlowable
from reportlab.lib.enums import TA_CENTER, TA_LEFT, TA_JUSTIFY

PDF_PATH = '/tmp/workspace/attachments/759bcff1-1d22-466a-a83b-7816e7d6daf7/10. FMT Hyperrevision.pdf'

# ── helpers ──────────────────────────────────────────────────────────────────
def clean(t):
    t = re.sub(r'\s+', ' ', t or '').strip()
    return t

def extract_key_terms(text):
    """Extract important terms / acronyms / names from the page text."""
    # acronyms like IPC, BNS, CrPC …
    acronyms = re.findall(r'\b[A-Z]{2,6}\b', text)
    # Capitalised words that look like proper terms
    cap_words = re.findall(r'\b([A-Z][a-z]{3,})\b', text)
    # Numbers that appear to be sections / years
    numbers  = re.findall(r'\b(\d{3,4})\b', text)
    return acronyms, cap_words, numbers

# ── Q&A generation rules per page ────────────────────────────────────────────
def make_questions_for_page(page_num, text):
    """Return list of (question, one-word-answer) tuples — up to 5."""
    qa = []
    lines = [l.strip() for l in text.split('\n') if l.strip()]
    
    # ---------- pattern bank ----------
    # 1. "X → Y"  (definition arrow)
    for line in lines:
        if len(qa) >= 5: break
        m = re.search(r'([A-Z][A-Za-z\s\-]{2,40}?)\s*[→\-]+\s*([A-Z][A-Za-z]+)', line)
        if m:
            term = m.group(1).strip().rstrip(' ')
            ans  = m.group(2).strip()
            if len(ans.split()) == 1 and ans.isalpha() and len(ans) > 2:
                q = f"The legal/medical term '{term}' is specifically associated with which concept?"
                qa.append((q, ans))

    # 2. "Section XXX → topic"
    for line in lines:
        if len(qa) >= 5: break
        m = re.search(r'[Ss]ection\s+(\d+[A-Za-z]?)\s+[→\-:]+\s*([A-Z][A-Za-z]+)', line)
        if m:
            sec = m.group(1)
            ans = m.group(2).strip()
            if len(ans.split()) == 1:
                qa.append((f"Section {sec} of the Act deals with which offence/topic?", ans))

    # 3. Numbered definitions table: "Term   Definition"
    for i, line in enumerate(lines):
        if len(qa) >= 5: break
        # lines that start with a number then a term
        m = re.match(r'^\d+\s+([A-Z][A-Za-z\s\-]{2,30}?)\s{2,}(.+)', line)
        if m:
            term = m.group(1).strip()
            rest = m.group(2).strip()
            # first capitalised word in rest is the answer
            ans_m = re.match(r'([A-Z][a-z]{2,})', rest)
            if ans_m:
                ans = ans_m.group(1)
                qa.append((f"In the context of Forensic Medicine, what is '{term}' primarily associated with?", ans))

    # 4. "X is called / known as Y"
    for line in lines:
        if len(qa) >= 5: break
        m = re.search(r'([A-Z][A-Za-z\s]{2,30}?)\s+is\s+(?:called|known as|termed)\s+([A-Z][A-Za-z]+)', line)
        if m:
            desc = m.group(1).strip()
            ans  = m.group(2).strip()
            if len(ans.split()) == 1:
                qa.append((f"What is '{desc}' called?", ans))

    # 5. Key-value lines like "Term   Meaning"  (two tokens separated by spaces)
    for line in lines:
        if len(qa) >= 5: break
        parts = re.split(r'\s{2,}|\t', line)
        if len(parts) >= 2:
            key = parts[0].strip()
            val = parts[1].strip()
            # answer should be a single word
            if (2 < len(key) < 40 and len(val.split()) == 1
                    and val.isalpha() and val[0].isupper()):
                qa.append((f"What is the one-word term associated with '{key}'?", val))

    # 6. Abbreviation expansions  "BNS = Bharatiya …"
    for line in lines:
        if len(qa) >= 5: break
        m = re.search(r'\b([A-Z]{2,6})\s*[=:]\s*([A-Z][a-z]+(?:\s+[A-Z][a-z]+){0,3})', line)
        if m:
            abbr = m.group(1)
            full = m.group(2).strip()
            first_word = full.split()[0]
            if len(first_word) > 3:
                qa.append((f"'{abbr}' stands for — what is its full form's first word?", first_word))

    # 7. Numbers / years in context
    for line in lines:
        if len(qa) >= 5: break
        m = re.search(r'(\d{4})\s+[→\-:]+\s*([A-Z][A-Za-z]+)', line)
        if m:
            year = m.group(1)
            topic = m.group(2)
            if len(topic.split()) == 1:
                qa.append((f"The year {year} is associated with which act/event in the text?", topic))

    # 8. Mnemonic lines  "PYT Gold → …"
    for line in lines:
        if len(qa) >= 5: break
        if 'PYT' in line or 'MNEMONIC' in line:
            # pull the answer word that follows
            m = re.search(r'MNEMONIC\s*[:\-]?\s*([A-Z][A-Za-z]+)', line)
            if m:
                ans = m.group(1)
                qa.append((f"What is the mnemonic key word highlighted on page {page_num}?", ans))

    # 9. Fallback: grab any unique capitalised single-word answers from the text
    fallback_terms = re.findall(r'\b([A-Z][a-z]{3,12})\b', text)
    seen = set(a for _, a in qa)
    for term in dict.fromkeys(fallback_terms):   # preserve order, deduplicate
        if len(qa) >= 5: break
        if term not in seen and term not in ('Page', 'Course', 'Gold', 'Memory'):
            qa.append((f"Name the key term/concept '{term}' found on this page.", term))
            seen.add(term)

    return qa[:5]


# ── Load pages ────────────────────────────────────────────────────────────────
print("Extracting pages …")
pages_text = {}
with pdfplumber.open(PDF_PATH) as pdf:
    for i, page in enumerate(pdf.pages):
        pages_text[i + 1] = clean(page.extract_text() or '')

print(f"Loaded {len(pages_text)} pages")

# ── Generate all QAs ──────────────────────────────────────────────────────────
all_qa = {}   # page_num → [(q, a), …]
for pn, text in pages_text.items():
    all_qa[pn] = make_questions_for_page(pn, text)

# stats
total_q = sum(len(v) for v in all_qa.values())
print(f"Generated {total_q} questions across {len(all_qa)} pages")

# ── Build PDF ─────────────────────────────────────────────────────────────────
OUT = '/tmp/workspace/FMT_OneWord_QA.pdf'
doc = SimpleDocTemplate(
    OUT,
    pagesize=A4,
    rightMargin=0.75*inch,
    leftMargin=0.75*inch,
    topMargin=0.75*inch,
    bottomMargin=0.75*inch,
)

styles = getSampleStyleSheet()

title_style = ParagraphStyle(
    'TitleStyle',
    parent=styles['Title'],
    fontSize=20,
    textColor=colors.HexColor('#1a3a6b'),
    spaceAfter=6,
    alignment=TA_CENTER,
)
sub_style = ParagraphStyle(
    'SubStyle',
    parent=styles['Normal'],
    fontSize=11,
    textColor=colors.HexColor('#555555'),
    alignment=TA_CENTER,
    spaceAfter=20,
)
page_header_style = ParagraphStyle(
    'PageHeader',
    parent=styles['Heading2'],
    fontSize=13,
    textColor=colors.white,
    backColor=colors.HexColor('#1a3a6b'),
    spaceBefore=14,
    spaceAfter=6,
    leftIndent=6,
    borderPad=4,
)
q_style = ParagraphStyle(
    'QStyle',
    parent=styles['Normal'],
    fontSize=10.5,
    textColor=colors.HexColor('#222222'),
    leftIndent=10,
    spaceBefore=4,
)
a_style = ParagraphStyle(
    'AStyle',
    parent=styles['Normal'],
    fontSize=11,
    textColor=colors.HexColor('#1a7a1a'),
    leftIndent=24,
    spaceBefore=2,
    spaceAfter=8,
    fontName='Helvetica-Bold',
)

story = []

# Cover / title
story.append(Spacer(1, 0.5*inch))
story.append(Paragraph("FMT Hyperrevision", title_style))
story.append(Paragraph("One-Word Answer Questions &amp; Answers", sub_style))
story.append(Paragraph("5 Questions per Page | Page Numbers Indicated", sub_style))
story.append(HRFlowable(width='100%', thickness=2, color=colors.HexColor('#1a3a6b')))
story.append(Spacer(1, 0.3*inch))

for pn in sorted(all_qa.keys()):
    qa_list = all_qa[pn]
    if not qa_list:
        continue
    
    # Section header
    story.append(Paragraph(f"  Page {pn}", page_header_style))

    for idx, (q, a) in enumerate(qa_list, 1):
        story.append(Paragraph(f"<b>Q{idx}.</b> {q}", q_style))
        story.append(Paragraph(f"Ans: {a}", a_style))
    
    story.append(HRFlowable(width='100%', thickness=0.5, color=colors.HexColor('#cccccc')))

doc.build(story)
print(f"PDF saved to {OUT}")

import json, re, pdfplumber
from reportlab.lib.pagesizes import A4
from reportlab.lib.styles import getSampleStyleSheet, ParagraphStyle
from reportlab.lib.units import inch
from reportlab.lib import colors
from reportlab.platypus import SimpleDocTemplate, Paragraph, Spacer, HRFlowable, KeepTogether
from reportlab.lib.enums import TA_CENTER, TA_LEFT

PDF_PATH = '/tmp/workspace/attachments/759bcff1-1d22-466a-a83b-7816e7d6daf7/10. FMT Hyperrevision.pdf'

SKIP_WORDS = {
    'Cerebellum','Hyperrevision','Course','Page','Gold','Memory','Also','Then',
    'This','That','Both','When','Note','Only','Each','From','With','Have',
    'They','Into','More','Such','Some','Other','After','Before','Under','Over',
    'These','Those','Very','Most','About','Between','Through','During','Against',
    'Because','However','Therefore','Thus','Hence','Also','Even','Just','Much',
    'Same','Last','Next','First','Second','Third','Follow','Given','Used','Made',
    'Well','Like','Than','Then','Than','Been','Were','What','Which','Where','While',
    'Conditions','Related','Common','Legal','Medical','Section','Example','Table',
    'Mnemonic','Following','Important','Above','Below','Around','Across','Within',
    'Aid','Def','Definition','Term','Type','Kind','Form','Part','Area','Part','Role',
    'Here','There','Their','Often','Always','Usually','Never','Every','Each',
    'Make','Made','Show','Shows','Shown','Found','Shows','Cases','Case',
}

def clean(t):
    return re.sub(r'\s+', ' ', t or '').strip()

# ── smart Q&A generator ───────────────────────────────────────────────────────
def make_questions_for_page(page_num, text):
    qa = []
    seen_answers = set()
    lines = [l.strip() for l in text.split('\n') if len(l.strip()) > 5]

    def add(q, a):
        a = a.strip().rstrip('.,;:')
        if a and len(a.split()) == 1 and a not in seen_answers and a not in SKIP_WORDS and len(a) > 2:
            seen_answers.add(a)
            qa.append((q, a))
            return True
        return False

    # ─ 1. "X → Y" or "X - Y" definition arrows (very common in this text)
    for line in lines:
        if len(qa) >= 5: break
        m = re.search(r'"?([A-Z][A-Za-z\s\-]{2,35}?)"?\s*[→\-]{1,2}\s*([A-Z][A-Za-z]+)', line)
        if m:
            term = m.group(1).strip().strip('"')
            ans  = m.group(2).strip()
            if ans not in SKIP_WORDS:
                add(f"What does '{term}' refer to in Forensic Medicine and Toxicology?", ans)

    # ─ 2. Abbreviation = Full form  (e.g. IPC = Indian Penal Code)
    for line in lines:
        if len(qa) >= 5: break
        m = re.search(r'\b([A-Z]{2,5})\s*[=:]\s*([A-Z][a-z]+(?:\s+[A-Z][a-z]+){0,4})', line)
        if m:
            abbr = m.group(1)
            full = m.group(2).strip()
            first_word = full.split()[0]
            if first_word not in SKIP_WORDS and len(first_word) > 3:
                add(f"'{abbr}' is an abbreviation — what is the first word of its full form?", first_word)

    # ─ 3. Section number → topic
    for line in lines:
        if len(qa) >= 5: break
        m = re.search(r'[Ss]ec(?:tion)?\.?\s*(\d{1,4}[A-Za-z]?)\s*[→\-:]+\s*([A-Z][A-Za-z]+)', line)
        if m:
            sec = m.group(1)
            ans = m.group(2).strip()
            if ans not in SKIP_WORDS:
                add(f"Section {sec} in the relevant Act primarily covers which topic?", ans)

    # ─ 4. "X is Y" or "called X"
    for line in lines:
        if len(qa) >= 5: break
        m = re.search(r'([A-Z][A-Za-z\s]{3,30}?)\s+is\s+(?:called|known as|defined as|termed)\s+([A-Z][A-Za-z]+)', line)
        if m:
            desc = m.group(1).strip()
            ans  = m.group(2).strip()
            add(f"What is '{desc}' officially called or defined as?", ans)

    # ─ 5. Paired table rows (two or more uppercase-starting words separated by lots of space)
    for line in lines:
        if len(qa) >= 5: break
        parts = re.split(r'\s{3,}|\t', line)
        if len(parts) >= 2:
            key = parts[0].strip()
            val = parts[1].strip()
            val_words = val.split()
            if (3 < len(key) < 40 and len(val_words) >= 1 and
                    val_words[0][0].isupper() and len(val_words[0]) > 3 and
                    val_words[0] not in SKIP_WORDS and key not in SKIP_WORDS):
                first = val_words[0].rstrip('.,;:')
                add(f"In the FMT table, what is the key concept paired with '{key}'?", first)

    # ─ 6. Latin / medico-legal terms (italicised / all-caps phrases)
    for line in lines:
        if len(qa) >= 5: break
        # Latin phrases like "Res Ipsa Loquitur", "Post Mortem", etc.
        m = re.search(r'\b((?:[A-Z][a-z]{1,12}\s+){1,3}[A-Z][a-z]{1,12})\b.*?([A-Z][a-z]{4,})', line)
        if m:
            phrase = m.group(1).strip()
            ans = m.group(2).strip()
            if phrase.count(' ') >= 1 and ans not in SKIP_WORDS:
                add(f"What is the key concept associated with the medico-legal term '{phrase}'?", ans)

    # ─ 7. Numbers / years associated with specific topics
    for line in lines:
        if len(qa) >= 5: break
        m = re.search(r'\b(\d{4})\b.*?([A-Z][A-Za-z]{4,})', line)
        if m:
            year = m.group(1)
            ans  = m.group(2).strip()
            if ans not in SKIP_WORDS and 1800 < int(year) < 2030:
                add(f"The year {year} in FMT context is associated with which term/act?", ans)

    # ─ 8. Toxicology / pharmacology words
    tox_pattern = re.findall(r'\b([A-Z][a-z]{3,}(?:ine|ide|ate|ane|ene|ium|sis|tic|ism|oid|oma|ase|uria|emia|itis|osis|pathy))\b', text)
    for ans in dict.fromkeys(tox_pattern):
        if len(qa) >= 5: break
        if ans not in SKIP_WORDS and ans not in seen_answers:
            add(f"Name the compound/condition ending in a medical suffix that appears on page {page_num}.", ans)

    # ─ 9. Smarter fallback: rare capitalised single-word nouns only
    cap_words = re.findall(r'\b([A-Z][a-z]{4,15})\b', text)
    freq = {}
    for w in cap_words:
        freq[w] = freq.get(w, 0) + 1
    # prefer words that appear 1-3 times (content-specific, not structural)
    candidates = [w for w, c in freq.items() if 1 <= c <= 3 and w not in SKIP_WORDS]
    for ans in candidates:
        if len(qa) >= 5: break
        add(f"Identify the specific term '{ans}' encountered in this page of FMT Hyperrevision.", ans)

    return qa[:5]


# ── Load pages ────────────────────────────────────────────────────────────────
print("Extracting pages …")
pages_text = {}
with pdfplumber.open(PDF_PATH) as pdf:
    for i, page in enumerate(pdf.pages):
        pages_text[i + 1] = clean(page.extract_text() or '')

print(f"Loaded {len(pages_text)} pages")

# ── Generate all QAs ──────────────────────────────────────────────────────────
all_qa = {}
for pn, text in pages_text.items():
    all_qa[pn] = make_questions_for_page(pn, text)

total_q = sum(len(v) for v in all_qa.values())
print(f"Generated {total_q} questions across {len(all_qa)} pages")

# ── Build PDF ─────────────────────────────────────────────────────────────────
OUT = '/tmp/workspace/FMT_OneWord_QA.pdf'
doc = SimpleDocTemplate(
    OUT,
    pagesize=A4,
    rightMargin=0.8*inch,
    leftMargin=0.8*inch,
    topMargin=0.8*inch,
    bottomMargin=0.8*inch,
)

styles = getSampleStyleSheet()

title_style = ParagraphStyle('TitleS', parent=styles['Title'],
    fontSize=22, textColor=colors.HexColor('#1a3a6b'),
    spaceAfter=4, alignment=TA_CENTER)
sub_style = ParagraphStyle('SubS', parent=styles['Normal'],
    fontSize=11, textColor=colors.HexColor('#555555'),
    alignment=TA_CENTER, spaceAfter=4)
note_style = ParagraphStyle('NoteS', parent=styles['Normal'],
    fontSize=9, textColor=colors.HexColor('#888888'),
    alignment=TA_CENTER, spaceAfter=20)
page_hdr_style = ParagraphStyle('PageHdr', parent=styles['Normal'],
    fontSize=12, textColor=colors.white,
    fontName='Helvetica-Bold',
    backColor=colors.HexColor('#1a3a6b'),
    spaceBefore=12, spaceAfter=4,
    leftIndent=8, rightIndent=8, borderPad=5)
q_style = ParagraphStyle('QS', parent=styles['Normal'],
    fontSize=10.5, textColor=colors.HexColor('#1a1a1a'),
    leftIndent=14, spaceBefore=5, spaceAfter=1)
a_style = ParagraphStyle('AS', parent=styles['Normal'],
    fontSize=11, textColor=colors.HexColor('#1a7a1a'),
    fontName='Helvetica-Bold',
    leftIndent=28, spaceBefore=1, spaceAfter=6)

story = []

# Title block
story.append(Spacer(1, 0.4*inch))
story.append(Paragraph("FMT Hyperrevision", title_style))
story.append(Paragraph("One-Word Answer · Question &amp; Answer Bank", sub_style))
story.append(Paragraph("5 Questions per Page · Source Page Number Indicated", note_style))
story.append(HRFlowable(width='100%', thickness=2, color=colors.HexColor('#1a3a6b'), spaceAfter=10))
story.append(Spacer(1, 0.15*inch))

for pn in sorted(all_qa.keys()):
    qa_list = all_qa[pn]
    if not qa_list:
        continue

    block = []
    block.append(Paragraph(f"  Page {pn}", page_hdr_style))
    for idx, (q, a) in enumerate(qa_list, 1):
        block.append(Paragraph(f"<b>Q{idx}.</b> {q}", q_style))
        block.append(Paragraph(f"<b>Ans:</b> {a}", a_style))
    block.append(HRFlowable(width='100%', thickness=0.5,
                             color=colors.HexColor('#cccccc'), spaceAfter=4))

    story.append(KeepTogether(block))

doc.build(story)
print(f"PDF saved → {OUT}")
import os
size = os.path.getsize(OUT)
print(f"File size: {size:,} bytes")