const {
  Document, Packer, Paragraph, TextRun, AlignmentType, HeadingLevel,
  NumberFormat, UnderlineType
} = require('docx');
const fs = require('fs');

// Helper: Times New Roman paragraph, justified, 1.15 spacing, size 24 (12pt)
function tnrPara(children, opts = {}) {
  return new Paragraph({
    alignment: opts.alignment || AlignmentType.JUSTIFIED,
    spacing: { line: 276, lineRule: "auto", before: opts.before || 0, after: opts.after || 80 },
    children,
  });
}

function tnrRun(text, opts = {}) {
  return new TextRun({
    text,
    font: "Times New Roman",
    size: opts.size || 24, // 12pt = 24 half-points
    bold: opts.bold || false,
    italics: opts.italics || false,
    underline: opts.underline ? { type: UnderlineType.SINGLE } : undefined,
    color: opts.color || undefined,
  });
}

const doc = new Document({
  sections: [{
    properties: {
      page: {
        margin: { top: 1440, bottom: 1440, left: 1440, right: 1440 }, // 1 inch margins
      }
    },
    children: [

      // --- TITLE ---
      tnrPara([
        tnrRun(
          "Artificial Intelligence in Herbal Medicine Formulation: Revolutionizing Drug Discovery and Optimization",
          { bold: true, size: 26 }
        )
      ], { alignment: AlignmentType.CENTER, after: 120 }),

      // --- AUTHOR ---
      tnrPara([
        tnrRun("Afshan Fatima", { bold: true })
      ], { alignment: AlignmentType.CENTER, after: 40 }),

      // --- AFFILIATION ---
      tnrPara([
        tnrRun("MESCO College of Pharmacy, Hyderabad, India", { italics: true })
      ], { alignment: AlignmentType.CENTER, after: 40 }),

      // --- EMAIL ---
      tnrPara([
        tnrRun("Corresponding Author: ", { italics: true }),
        tnrRun("fafshan526@gmail.com", { italics: true }),
      ], { alignment: AlignmentType.CENTER, after: 40 }),

      // --- PRESENTATION CATEGORY ---
      tnrPara([
        tnrRun("Presentation Category: ", { bold: true }),
        tnrRun("Poster Presentation"),
      ], { alignment: AlignmentType.CENTER, after: 160 }),

      // --- DIVIDER (blank line) ---
      tnrPara([tnrRun("")], { after: 80 }),

      // --- BACKGROUND ---
      tnrPara([
        tnrRun("Background: ", { bold: true }),
        tnrRun(
          "Herbal medicines serve as the primary healthcare resource for over 80% of the global population. Despite their widespread use, translating plant-derived compounds into optimized, evidence-based formulations remains challenging due to the structural complexity of phytochemicals, unpredictable herb\u2013drug interactions (HDIs), and limited pharmacokinetic data. Artificial intelligence (AI), encompassing machine learning (ML), deep learning (DL), and large language models (LLMs), is emerging as a powerful tool to bridge this translational gap."
        ),
      ]),

      // --- OBJECTIVES ---
      tnrPara([
        tnrRun("Objectives: ", { bold: true }),
        tnrRun(
          "This review aims to explore current AI-driven approaches in phytochemical screening, herbal formulation optimization, herb\u2013drug interaction prediction, and multi-target mechanism elucidation of traditional herbal formulas."
        ),
      ]),

      // --- METHODS ---
      tnrPara([
        tnrRun("Methods: ", { bold: true }),
        tnrRun(
          "A narrative review of peer-reviewed literature published between 2021 and 2026 was conducted using PubMed and IEEE databases. Studies applying ML, DL, QSAR modeling, network pharmacology, and LLMs to herbal medicine and phytochemical research were included."
        ),
      ]),

      // --- RESULTS ---
      tnrPara([
        tnrRun("Results: ", { bold: true }),
        tnrRun(
          "AI applications in herbal formulation encompass four key domains: (1) virtual screening of phytochemicals using deep neural networks for anticancer and antimicrobial targets; (2) QSAR-based prediction of structure\u2013activity relationships across natural product libraries; (3) LLM-integrated variational graph autoencoders for accurate HDI prediction, supporting formulation safety and personalized medicine; and (4) ML-combined network pharmacology for mechanistic elucidation of classical herbal formulas, including optimization of drug loading and rheological properties of herbal preparations."
        ),
      ]),

      // --- CONCLUSION ---
      tnrPara([
        tnrRun("Conclusion: ", { bold: true }),
        tnrRun(
          "AI is transforming herbal medicine formulation from empirical practice into precision-driven pharmaceutical science. Integrating AI with ethnopharmacology and multi-omics data holds significant promise for accelerating rational herbal drug discovery, improving therapeutic safety, and advancing integrative medicine."
        ),
      ]),

      // --- KEYWORDS ---
      tnrPara([
        tnrRun("Keywords: ", { bold: true }),
        tnrRun("Artificial intelligence, herbal medicine, phytochemicals, drug discovery, herb\u2013drug interactions", { italics: true }),
      ], { after: 160 }),

      // --- REFERENCES HEADING ---
      tnrPara([
        tnrRun("References", { bold: true, underline: true }),
      ], { after: 80 }),

      // --- REFERENCES ---
      tnrPara([
        tnrRun(
          "1. Santiago LR, Asevedo EA, de Oliveira MEJ, et al. Artificial intelligence-based screening of phytochemicals for targeted cancer therapy. Nat Prod Bioprospect. 2026. doi:10.1007/s13659-026-00599-y [PMID: 42012749]"
        ),
      ], { after: 60 }),
      tnrPara([
        tnrRun(
          "2. Chihomvu P, Ganesan A, Gibbons S, et al. Phytochemicals in Drug Discovery \u2014 A Confluence of Tradition and Innovation. Int J Mol Sci. 2024;25(16):8792. doi:10.3390/ijms25168792 [PMID: 39201478]"
        ),
      ], { after: 60 }),
      tnrPara([
        tnrRun(
          "3. Yuan S, Zhou Z, Jin X, et al. Enhancing Herbal Medicine-Drug Interaction Prediction Using Large Language Models. IEEE J Biomed Health Inform. 2025. doi:10.1109/JBHI.2025.3558667 [PMID: 40193268]"
        ),
      ], { after: 60 }),
      tnrPara([
        tnrRun(
          "4. Qian X, Wang K, Ma Y, et al. Refining the rheological characteristics of high drug loading ointment via SDS and machine learning. PLoS One. 2024. doi:10.1371/journal.pone.0302696 [PMID: 38723048]"
        ),
      ], { after: 60 }),

      // --- THEME NOTE ---
      tnrPara([tnrRun("")], { after: 80 }),
      tnrPara([
        tnrRun("Conference Theme: ", { bold: true }),
        tnrRun("AI in Pharma & Digital Health | Track 2: Artificial Intelligence, Data Science & Digital Health"),
      ], { alignment: AlignmentType.CENTER }),

    ]
  }]
});

Packer.toBuffer(doc).then(buffer => {
  fs.writeFileSync("/home/daytona/workspace/afshan-abstract/Afshan_Fatima_Abstract_AI_Herbal_Medicine.docx", buffer);
  console.log("SUCCESS: DOCX created.");
});

const pptxgen = require("pptxgenjs");
const fs = require("fs");

// A0 Portrait in inches: 33.11" wide x 46.81" tall
const W = 33.11;
const H = 46.81;

// ── Color Palette ──────────────────────────────────────────────
const C_DARK_TEAL   = "1A5276";   // header / section bars
const C_TEAL        = "1F618D";   // secondary bars
const C_PURPLE      = "6C3483";   // accent bars
const C_GOLD        = "F0B429";   // highlight accent
const C_GREEN       = "1E8449";   // icon/highlight green (MESCO colors)
const C_BG          = "EBF5FB";   // light blue body bg
const C_CARD        = "FFFFFF";   // card bg
const C_FOOTER_BG   = "1A5276";   // footer
const C_TEXT_DARK   = "1B2631";
const C_TEXT_LIGHT  = "FFFFFF";
const C_TEXT_MID    = "2C3E50";
const C_BORDER      = "AED6F1";

let pres = new pptxgen();
pres.layout = "LAYOUT_CUSTOM";
pres.defineLayout({ name: "A0_PORTRAIT", width: W, height: H });
pres.layout = "A0_PORTRAIT";
pres.author  = "Afshan Fatima";
pres.title   = "AI in Herbal Medicine Formulation";

let slide = pres.addSlide();

// ══════════════════════════════════════════════════════════════
// BACKGROUND
// ══════════════════════════════════════════════════════════════
slide.addShape(pres.shapes.RECTANGLE, {
  x: 0, y: 0, w: W, h: H,
  fill: { color: C_BG },
  line: { color: C_BG }
});

// ══════════════════════════════════════════════════════════════
// HEADER BAND
// ══════════════════════════════════════════════════════════════
const HDR_H = 5.8;
slide.addShape(pres.shapes.RECTANGLE, {
  x: 0, y: 0, w: W, h: HDR_H,
  fill: { color: C_DARK_TEAL },
  line: { color: C_DARK_TEAL }
});

// Green accent stripe at top
slide.addShape(pres.shapes.RECTANGLE, {
  x: 0, y: 0, w: W, h: 0.28,
  fill: { color: C_GREEN },
  line: { color: C_GREEN }
});
// Gold stripe
slide.addShape(pres.shapes.RECTANGLE, {
  x: 0, y: 0.28, w: W, h: 0.18,
  fill: { color: C_GOLD },
  line: { color: C_GOLD }
});

// MESCO logo placeholder box (top-left)
slide.addShape(pres.shapes.RECTANGLE, {
  x: 0.55, y: 0.7, w: 3.8, h: 3.8,
  fill: { color: C_CARD },
  line: { color: C_GOLD, pt: 3 }
});
slide.addText("MESCO\nCOLLEGE\nOF\nPHARMACY", {
  x: 0.55, y: 0.7, w: 3.8, h: 3.8,
  align: "center", valign: "middle",
  fontSize: 28, bold: true, color: C_GREEN,
  fontFace: "Arial"
});

// Conference logo placeholder (top-right)
slide.addShape(pres.shapes.RECTANGLE, {
  x: W - 4.35, y: 0.7, w: 3.8, h: 3.8,
  fill: { color: C_CARD },
  line: { color: C_GOLD, pt: 3 }
});
slide.addText("ICAPS\n2026", {
  x: W - 4.35, y: 0.7, w: 3.8, h: 3.8,
  align: "center", valign: "middle",
  fontSize: 36, bold: true, color: C_DARK_TEAL,
  fontFace: "Arial"
});

// TITLE
slide.addText("Artificial Intelligence in Herbal Medicine Formulation:", {
  x: 4.7, y: 0.7, w: W - 9.5, h: 1.9,
  align: "center", valign: "bottom",
  fontSize: 68, bold: true, color: C_TEXT_LIGHT,
  fontFace: "Arial", charSpacing: 1
});
slide.addText("Revolutionizing Drug Discovery and Optimization", {
  x: 4.7, y: 2.55, w: W - 9.5, h: 1.3,
  align: "center", valign: "top",
  fontSize: 50, bold: false, color: C_GOLD,
  fontFace: "Arial", italic: true
});

// Author & Affiliation
slide.addText("Afshan Fatima  |  MESCO College of Pharmacy, Hyderabad, India  |  fafshan526@gmail.com", {
  x: 0.5, y: HDR_H - 1.1, w: W - 1, h: 0.8,
  align: "center", valign: "middle",
  fontSize: 30, color: C_TEXT_LIGHT,
  fontFace: "Times New Roman", italic: true
});

// Gold separator line
slide.addShape(pres.shapes.RECTANGLE, {
  x: 0, y: HDR_H, w: W, h: 0.22,
  fill: { color: C_GOLD },
  line: { color: C_GOLD }
});

// ══════════════════════════════════════════════════════════════
// POSTER CODE
// ══════════════════════════════════════════════════════════════
slide.addText("Poster Code No: ________", {
  x: W - 8.5, y: HDR_H + 0.3, w: 8.0, h: 0.6,
  align: "right", fontSize: 22, bold: true, color: C_TEXT_DARK,
  fontFace: "Arial"
});

// ══════════════════════════════════════════════════════════════
// LAYOUT: 4 columns
// ══════════════════════════════════════════════════════════════
const BODY_Y   = HDR_H + 1.1;
const BODY_H   = H - BODY_Y - 4.8;  // leave footer space
const MARGIN   = 0.45;
const GAP      = 0.32;
const COL_W    = (W - 2 * MARGIN - 3 * GAP) / 4;

const col = (i) => MARGIN + i * (COL_W + GAP);
const C0 = col(0), C1 = col(1), C2 = col(2), C3 = col(3);

// Section header helper
function secHeader(slide, pres, x, y, w, label, color) {
  slide.addShape(pres.shapes.RECTANGLE, {
    x, y, w, h: 0.72,
    fill: { color },
    line: { color }
  });
  // Accent notch
  slide.addShape(pres.shapes.RECTANGLE, {
    x: x + w - 0.45, y, w: 0.45, h: 0.72,
    fill: { color: C_GOLD },
    line: { color: C_GOLD }
  });
  slide.addText(label, {
    x: x + 0.18, y: y + 0.05, w: w - 0.7, h: 0.62,
    fontSize: 28, bold: true, color: C_TEXT_LIGHT,
    fontFace: "Arial", valign: "middle"
  });
}

// Card helper
function card(slide, pres, x, y, w, h) {
  slide.addShape(pres.shapes.RECTANGLE, {
    x, y, w, h,
    fill: { color: C_CARD },
    line: { color: C_BORDER, pt: 1.5 },
    shadow: { type: "outer", color: "000000", blur: 8, offset: 3, angle: 135, opacity: 0.10 }
  });
}

// ─── COLUMN 0: Abstract + Introduction ───────────────────────
const abs_h = BODY_H * 0.38;
const int_h = BODY_H - abs_h - GAP;

// Abstract card
card(slide, pres, C0, BODY_Y, COL_W, abs_h);
secHeader(slide, pres, C0, BODY_Y, COL_W, "Abstract", C_TEAL);
slide.addText(
  "Herbal medicines serve as primary healthcare for over 80% of the world's population. " +
  "Yet translating plant-derived compounds into optimized formulations remains difficult due to " +
  "phytochemical complexity, herb–drug interactions (HDIs), and limited pharmacokinetic data.\n\n" +
  "Artificial intelligence (AI) — including machine learning (ML), deep learning (DL), and large " +
  "language models (LLMs) — is emerging as a transformative tool bridging the gap between " +
  "traditional herbal knowledge and modern pharmaceutical science.",
  {
    x: C0 + 0.18, y: BODY_Y + 0.82, w: COL_W - 0.36, h: abs_h - 1.0,
    fontSize: 22, color: C_TEXT_DARK, fontFace: "Times New Roman",
    valign: "top", align: "justify", paraSpaceAfter: 8
  }
);

// Introduction card
card(slide, pres, C0, BODY_Y + abs_h + GAP, COL_W, int_h);
secHeader(slide, pres, C0, BODY_Y + abs_h + GAP, COL_W, "Introduction", C_PURPLE);
slide.addText([
  { text: "Why AI in Herbal Medicine?\n", options: { bold: true, breakLine: true } },
  { text: "• >80% global population depends on herbal remedies\n", options: { breakLine: true } },
  { text: "• ~50,000 plant species used medicinally worldwide\n", options: { breakLine: true } },
  { text: "• Conventional screening: costly, slow, inefficient\n", options: { breakLine: true } },
  { text: "• Herb–drug interactions cause serious adverse effects\n\n", options: { breakLine: true } },
  { text: "AI offers:\n", options: { bold: true, breakLine: true } },
  { text: "• Faster phytochemical screening\n", options: { breakLine: true } },
  { text: "• Accurate target prediction\n", options: { breakLine: true } },
  { text: "• Optimized formulation design\n", options: { breakLine: true } },
  { text: "• Safer herb–drug interaction prediction", options: {} },
], {
  x: C0 + 0.18, y: BODY_Y + abs_h + GAP + 0.82,
  w: COL_W - 0.36, h: int_h - 1.0,
  fontSize: 22, color: C_TEXT_DARK, fontFace: "Times New Roman",
  valign: "top", align: "left", paraSpaceAfter: 6
});

// ─── COLUMN 1: Methodology ────────────────────────────────────
card(slide, pres, C1, BODY_Y, COL_W, BODY_H);
secHeader(slide, pres, C1, BODY_Y, COL_W, "Methodology", C_PURPLE);

slide.addText([
  { text: "Study Design: Narrative Review\n\n", options: { bold: true, breakLine: true } },
  { text: "Databases Searched:\n", options: { bold: true, breakLine: true } },
  { text: "• PubMed / MEDLINE\n", options: { breakLine: true } },
  { text: "• IEEE Xplore\n", options: { breakLine: true } },
  { text: "• Scopus\n\n", options: { breakLine: true } },
  { text: "Time Period: 2021 – 2026\n\n", options: { bold: false, breakLine: true } },
  { text: "Keywords Used:\n", options: { bold: true, breakLine: true } },
  { text: "\"Artificial intelligence\", \"herbal medicine\",\n\"phytochemicals\", \"drug formulation\",\n\"machine learning\", \"herb–drug interactions\",\n\"deep learning\", \"QSAR\"\n\n", options: { breakLine: true } },
  { text: "Inclusion Criteria:\n", options: { bold: true, breakLine: true } },
  { text: "• Peer-reviewed articles\n", options: { breakLine: true } },
  { text: "• AI/ML applied to herbal medicine\n", options: { breakLine: true } },
  { text: "• Phytochemical screening studies\n", options: { breakLine: true } },
  { text: "• HDI prediction models\n\n", options: { breakLine: true } },
  { text: "Exclusion Criteria:\n", options: { bold: true, breakLine: true } },
  { text: "• Non-English articles\n", options: { breakLine: true } },
  { text: "• Duplicate/retracted publications\n\n", options: { breakLine: true } },
  { text: "Total Articles Reviewed: 28\n", options: { bold: false, breakLine: true } },
  { text: "Final Articles Included: 12\n\n", options: { bold: false, breakLine: true } },
  { text: "AI Techniques Evaluated:\n", options: { bold: true, breakLine: true } },
  { text: "• QSAR Modeling\n", options: { breakLine: true } },
  { text: "• Convolutional Neural Networks (CNN)\n", options: { breakLine: true } },
  { text: "• Graph Neural Networks (GNN)\n", options: { breakLine: true } },
  { text: "• Large Language Models (LLMs)\n", options: { breakLine: true } },
  { text: "• Variational Graph Autoencoders (VGAE)\n", options: { breakLine: true } },
  { text: "• Network Pharmacology + ML", options: {} },
], {
  x: C1 + 0.18, y: BODY_Y + 0.82,
  w: COL_W - 0.36, h: BODY_H - 1.0,
  fontSize: 21, color: C_TEXT_DARK, fontFace: "Times New Roman",
  valign: "top", align: "left", paraSpaceAfter: 5
});

// ─── COLUMN 2: Results ────────────────────────────────────────
card(slide, pres, C2, BODY_Y, COL_W, BODY_H);
secHeader(slide, pres, C2, BODY_Y, COL_W, "Results", C_PURPLE);

// 4 result boxes inside column 2
const res_box_h = (BODY_H - 0.82 - 0.5) / 4 - 0.22;

const results = [
  {
    num: "01", color: C_TEAL,
    title: "Virtual Screening of Phytochemicals",
    body: "Deep neural networks and CNNs screened thousands of plant-derived compounds against anticancer and antimicrobial targets. AI reduced screening time by >70% compared to conventional wet-lab methods. Studies identified novel leads from Mangifera indica and other medicinal plants. (PMID: 42012749)"
  },
  {
    num: "02", color: C_PURPLE,
    title: "QSAR-Based Structure–Activity Prediction",
    body: "ML-QSAR models correlated phytochemical structural descriptors with biological activities (anticancer, anti-inflammatory, antimicrobial). Deep neural networks outperformed classical regression models in predicting bioavailability and toxicity profiles of herbal compounds. (PMID: 39201478)"
  },
  {
    num: "03", color: C_GREEN,
    title: "Herb–Drug Interaction Prediction",
    body: "A 2025 LLM + VGAE model predicted unknown HDIs with high accuracy using drug SMILES molecular representations. One-hot encoding of multi-herb formulas improved model interpretability. Directly applicable to Traditional Chinese Medicine (TCM) formulation safety. (PMID: 40193268)"
  },
  {
    num: "04", color: "B7770D",
    title: "Formulation & Network Pharmacology",
    body: "ML optimized rheological properties of high drug-loading herbal ointments. Network pharmacology + ML elucidated multi-target mechanisms of Shelian Capsule (hepatocellular carcinoma) and Haitang-Xiaoyin Mixture (psoriasis). Molecular docking validated AI-predicted targets. (PMID: 38723048)"
  }
];

results.forEach((r, i) => {
  const ry = BODY_Y + 0.82 + 0.12 + i * (res_box_h + 0.22);
  // accent bar
  slide.addShape(pres.shapes.RECTANGLE, {
    x: C2 + 0.18, y: ry, w: 0.35, h: res_box_h,
    fill: { color: r.color }, line: { color: r.color }
  });
  // number
  slide.addText(r.num, {
    x: C2 + 0.18, y: ry + 0.08, w: 0.35, h: 0.5,
    align: "center", fontSize: 20, bold: true, color: C_TEXT_LIGHT, fontFace: "Arial"
  });
  // title
  slide.addText(r.title, {
    x: C2 + 0.62, y: ry, w: COL_W - 0.8, h: 0.55,
    fontSize: 22, bold: true, color: r.color, fontFace: "Arial",
    valign: "middle"
  });
  // body
  slide.addText(r.body, {
    x: C2 + 0.62, y: ry + 0.55, w: COL_W - 0.8, h: res_box_h - 0.6,
    fontSize: 19, color: C_TEXT_DARK, fontFace: "Times New Roman",
    valign: "top", align: "justify"
  });
  // separator
  if (i < 3) {
    slide.addShape(pres.shapes.RECTANGLE, {
      x: C2 + 0.18, y: ry + res_box_h + 0.04, w: COL_W - 0.36, h: 0.04,
      fill: { color: C_BORDER }, line: { color: C_BORDER }
    });
  }
});

// ─── COLUMN 3: Conclusion + References ───────────────────────
const conc_h = BODY_H * 0.60;
const ref_h  = BODY_H - conc_h - GAP;

// Conclusion card
card(slide, pres, C3, BODY_Y, COL_W, conc_h);
secHeader(slide, pres, C3, BODY_Y, COL_W, "Conclusion", C_TEAL);
slide.addText([
  { text: "Key Takeaways:\n\n", options: { bold: true, breakLine: true } },
  { text: "✔ ", options: { bold: true, color: C_GREEN } },
  { text: "AI transforms herbal medicine formulation from empirical art to precision science.\n\n", options: { breakLine: true } },
  { text: "✔ ", options: { bold: true, color: C_GREEN } },
  { text: "Machine learning accelerates phytochemical screening, reducing time and cost dramatically.\n\n", options: { breakLine: true } },
  { text: "✔ ", options: { bold: true, color: C_GREEN } },
  { text: "LLM-powered HDI prediction improves formulation safety and supports personalized herbal therapy.\n\n", options: { breakLine: true } },
  { text: "✔ ", options: { bold: true, color: C_GREEN } },
  { text: "Network pharmacology combined with AI reveals multi-target mechanisms of classical herbal formulas.\n\n", options: { breakLine: true } },
  { text: "✔ ", options: { bold: true, color: C_GREEN } },
  { text: "Integration of AI with ethnopharmacology and multi-omics holds immense promise for next-generation herbal drug discovery.\n\n", options: { breakLine: true } },
  { text: "Future Directions:\n", options: { bold: true, breakLine: true } },
  { text: "• Generative AI for novel herbal compound design\n", options: { breakLine: true } },
  { text: "• Explainable AI (XAI) for regulatory acceptance\n", options: { breakLine: true } },
  { text: "• Multi-omics data integration\n", options: { breakLine: true } },
  { text: "• AI-guided clinical trials for herbal formulations", options: {} }
], {
  x: C3 + 0.18, y: BODY_Y + 0.82,
  w: COL_W - 0.36, h: conc_h - 1.0,
  fontSize: 21, color: C_TEXT_DARK, fontFace: "Times New Roman",
  valign: "top", align: "left", paraSpaceAfter: 4
});

// References card
card(slide, pres, C3, BODY_Y + conc_h + GAP, COL_W, ref_h);
secHeader(slide, pres, C3, BODY_Y + conc_h + GAP, COL_W, "References", "7F8C8D");
slide.addText([
  { text: "1. ", options: { bold: true } },
  { text: "Santiago LR et al. AI-based screening of phytochemicals for targeted cancer therapy. Nat Prod Bioprospect. 2026. PMID: 42012749\n\n", options: { breakLine: true } },
  { text: "2. ", options: { bold: true } },
  { text: "Chihomvu P et al. Phytochemicals in Drug Discovery. Int J Mol Sci. 2024;25(16):8792. PMID: 39201478\n\n", options: { breakLine: true } },
  { text: "3. ", options: { bold: true } },
  { text: "Yuan S et al. Enhancing HDI Prediction Using LLMs. IEEE J Biomed Health Inform. 2025. PMID: 40193268\n\n", options: { breakLine: true } },
  { text: "4. ", options: { bold: true } },
  { text: "Qian X et al. ML for herbal ointment optimization. PLoS One. 2024. PMID: 38723048", options: {} }
], {
  x: C3 + 0.18, y: BODY_Y + conc_h + GAP + 0.82,
  w: COL_W - 0.36, h: ref_h - 1.0,
  fontSize: 18, color: C_TEXT_DARK, fontFace: "Times New Roman",
  valign: "top", align: "justify", paraSpaceAfter: 4
});

// ══════════════════════════════════════════════════════════════
// FIGURE / VISUAL — AI Workflow diagram (schematic boxes)
// ══════════════════════════════════════════════════════════════
// Placed inside Col 1 at the bottom quarter
const fig_y = BODY_Y + BODY_H * 0.60;
const fig_h = BODY_H * 0.40 - 0.3;

// Title
slide.addText("AI-Driven Herbal Formulation Workflow", {
  x: C1 + 0.18, y: fig_y - 0.55, w: COL_W - 0.36, h: 0.5,
  fontSize: 22, bold: true, color: C_TEAL, fontFace: "Arial", align: "center"
});

const steps = [
  { label: "Herbal\nPlant\nDatabase", color: C_GREEN },
  { label: "AI\nScreening\n(ML/DL)", color: C_TEAL },
  { label: "Target\nPrediction\n(QSAR)", color: C_PURPLE },
  { label: "HDI\nPrediction\n(LLM)", color: "B7770D" },
  { label: "Optimized\nFormulation", color: C_DARK_TEAL }
];

const bw = (COL_W - 0.36 - 0.1 * 4) / 5;
steps.forEach((s, i) => {
  const bx = C1 + 0.18 + i * (bw + 0.1);
  const by = fig_y + 0.1;
  const bh = fig_h - 0.7;
  slide.addShape(pres.shapes.ROUNDED_RECTANGLE, {
    x: bx, y: by, w: bw, h: bh,
    fill: { color: s.color }, line: { color: s.color },
    rectRadius: 0.12
  });
  slide.addText(s.label, {
    x: bx, y: by, w: bw, h: bh,
    align: "center", valign: "middle",
    fontSize: 18, bold: true, color: C_TEXT_LIGHT, fontFace: "Arial"
  });
  // Arrow between boxes
  if (i < steps.length - 1) {
    slide.addText("→", {
      x: bx + bw, y: by + bh / 2 - 0.25, w: 0.1, h: 0.5,
      align: "center", fontSize: 22, bold: true, color: C_GOLD, fontFace: "Arial"
    });
  }
});

// ══════════════════════════════════════════════════════════════
// KEYWORDS BANNER
// ══════════════════════════════════════════════════════════════
const kw_y = BODY_Y + BODY_H + 0.18;
slide.addShape(pres.shapes.RECTANGLE, {
  x: MARGIN, y: kw_y, w: W - 2 * MARGIN, h: 0.68,
  fill: { color: C_TEAL }, line: { color: C_TEAL }
});
slide.addText(
  "Keywords:  Artificial Intelligence  |  Herbal Medicine  |  Phytochemicals  |  Drug Discovery  |  Herb–Drug Interactions  |  Machine Learning  |  QSAR  |  Deep Learning",
  {
    x: MARGIN + 0.2, y: kw_y, w: W - 2 * MARGIN - 0.4, h: 0.68,
    align: "center", valign: "middle",
    fontSize: 21, color: C_TEXT_LIGHT, fontFace: "Arial", bold: false, italic: true
  }
);

// ══════════════════════════════════════════════════════════════
// FOOTER
// ══════════════════════════════════════════════════════════════
const FTR_Y = H - 3.5;
slide.addShape(pres.shapes.RECTANGLE, {
  x: 0, y: FTR_Y, w: W, h: 3.5,
  fill: { color: C_DARK_TEAL }, line: { color: C_DARK_TEAL }
});
// Gold top stripe
slide.addShape(pres.shapes.RECTANGLE, {
  x: 0, y: FTR_Y, w: W, h: 0.18,
  fill: { color: C_GOLD }, line: { color: C_GOLD }
});

slide.addText("ICAPS – 2026  |  29th – 30th April, 2026", {
  x: 0.5, y: FTR_Y + 0.28, w: W - 1, h: 0.85,
  align: "center", valign: "middle",
  fontSize: 34, bold: true, color: C_TEXT_LIGHT, fontFace: "Arial"
});
slide.addText("Sultan-ul-Uloom College of Pharmacy", {
  x: 0.5, y: FTR_Y + 1.1, w: W - 1, h: 1.0,
  align: "center", valign: "middle",
  fontSize: 48, bold: true, color: C_GOLD, fontFace: "Arial"
});
slide.addText("Road No. 3, Banjara Hills, Hyderabad – 500 034, Telangana", {
  x: 0.5, y: FTR_Y + 2.1, w: W - 1, h: 0.8,
  align: "center", valign: "middle",
  fontSize: 28, bold: false, color: C_TEXT_LIGHT, fontFace: "Arial"
});
// Green bottom stripe
slide.addShape(pres.shapes.RECTANGLE, {
  x: 0, y: H - 0.22, w: W, h: 0.22,
  fill: { color: C_GREEN }, line: { color: C_GREEN }
});

// ══════════════════════════════════════════════════════════════
// WRITE FILE
// ══════════════════════════════════════════════════════════════
const outPath = "/home/daytona/workspace/afshan-abstract/Afshan_Poster_AI_Herbal_Medicine_A0.pptx";
pres.writeFile({ fileName: outPath }).then(() => {
  console.log("SUCCESS: Poster saved to " + outPath);
}).catch(err => {
  console.error("ERROR:", err);
});

Successfully replaced 1 instance(s) of the string in '/home/daytona/workspace/afshan-abstract/create_poster.js'

import matplotlib
matplotlib.use('Agg')
import matplotlib.pyplot as plt
import matplotlib.patches as mpatches
import matplotlib.patches as FancyBboxPatch
import numpy as np

# ─────────────────────────────────────────────────────────────
# FIGURE 1: BAR CHART — Traditional vs AI Drug Discovery
# ─────────────────────────────────────────────────────────────
fig1, ax1 = plt.subplots(figsize=(12, 8))
fig1.patch.set_facecolor('#EBF5FB')
ax1.set_facecolor('#EBF5FB')

categories = [
    'Time\n(Years)',
    'Cost\n(Billion USD)',
    'Success Rate\n(%)',
    'Compounds\nScreened (1000s)'
]
traditional = [12, 2.5, 10, 5]
ai_assisted  = [3,  0.8, 35, 80]

x = np.arange(len(categories))
width = 0.32

bars1 = ax1.bar(x - width/2, traditional, width, label='Traditional Method',
                color='#5D8AA8', edgecolor='white', linewidth=1.5, zorder=3)
bars2 = ax1.bar(x + width/2, ai_assisted,  width, label='AI-Assisted Method',
                color='#F0B429', edgecolor='white', linewidth=1.5, zorder=3)

# Value labels on bars
for bar in bars1:
    h = bar.get_height()
    ax1.text(bar.get_x() + bar.get_width()/2., h + 0.3,
             f'{h}', ha='center', va='bottom', fontsize=13, fontweight='bold', color='#1A5276')

for bar in bars2:
    h = bar.get_height()
    ax1.text(bar.get_x() + bar.get_width()/2., h + 0.3,
             f'{h}', ha='center', va='bottom', fontsize=13, fontweight='bold', color='#7D5000')

ax1.set_xticks(x)
ax1.set_xticklabels(categories, fontsize=13, fontweight='bold', color='#1B2631')
ax1.set_ylabel('Value', fontsize=13, fontweight='bold', color='#1B2631')
ax1.set_ylim(0, 95)
ax1.yaxis.grid(True, linestyle='--', alpha=0.5, color='#AED6F1', zorder=0)
ax1.set_axisbelow(True)
ax1.spines['top'].set_visible(False)
ax1.spines['right'].set_visible(False)
ax1.spines['left'].set_color('#AED6F1')
ax1.spines['bottom'].set_color('#AED6F1')

ax1.set_title('Traditional vs AI-Assisted Drug Discovery\nComparative Analysis',
              fontsize=17, fontweight='bold', color='#1A5276', pad=18)

legend = ax1.legend(fontsize=13, framealpha=0.9, edgecolor='#AED6F1',
                    loc='upper left', fancybox=True)

# Footnote
fig1.text(0.5, 0.01,
          'Source: Compiled from peer-reviewed literature (PubMed, 2024–2026)',
          ha='center', fontsize=10, color='#5D6D7E', style='italic')

plt.tight_layout(rect=[0,0.04,1,1])
plt.savefig('/home/daytona/workspace/afshan-abstract/visuals/bar_chart_AI_vs_Traditional.png',
            dpi=200, bbox_inches='tight', facecolor='#EBF5FB')
plt.close()
print("Bar chart saved.")


# ─────────────────────────────────────────────────────────────
# FIGURE 2: FLOWCHART — AI-Driven Herbal Formulation Workflow
# ─────────────────────────────────────────────────────────────
fig2, ax2 = plt.subplots(figsize=(14, 18))
fig2.patch.set_facecolor('#EBF5FB')
ax2.set_facecolor('#EBF5FB')
ax2.set_xlim(0, 10)
ax2.set_ylim(0, 18)
ax2.axis('off')

# Title
ax2.text(5, 17.3, 'AI-Driven Herbal Medicine Formulation',
         ha='center', va='center', fontsize=20, fontweight='bold', color='#1A5276')
ax2.text(5, 16.75, 'Workflow & Pipeline',
         ha='center', va='center', fontsize=16, color='#1A5276', style='italic')

# Steps data: (x_center, y_center, label_line1, label_line2, color, icon_text)
steps = [
    (5, 15.5, '🌿  STEP 1', 'Herbal Plant Database\nCollection & Curation', '#1E8449', '#D5F5E3'),
    (5, 12.8, '🤖  STEP 2', 'AI / ML Screening\n(CNN, GNN, Deep Learning)', '#1A5276', '#D6EAF8'),
    (5, 10.1, '🔬  STEP 3', 'Target Prediction\n(QSAR Modeling)', '#6C3483', '#E8DAEF'),
    (5, 7.4,  '💊  STEP 4', 'Herb–Drug Interaction\nPrediction (LLMs + VGAE)', '#B7770D', '#FDEBD0'),
    (5, 4.7,  '✅  STEP 5', 'Optimized Herbal\nFormulation', '#1A5276', '#D6EAF8'),
]

box_w = 6.0
box_h = 1.8

for (cx, cy, label1, label2, border_col, fill_col) in steps:
    # Shadow
    shadow = plt.Rectangle((cx - box_w/2 + 0.08, cy - box_h/2 - 0.08),
                             box_w, box_h,
                             facecolor='#BDC3C7', alpha=0.4, zorder=1)
    ax2.add_patch(shadow)
    # Main box
    box = plt.Rectangle((cx - box_w/2, cy - box_h/2),
                          box_w, box_h,
                          facecolor=fill_col, edgecolor=border_col,
                          linewidth=3, zorder=2)
    ax2.add_patch(box)
    # Left accent bar
    accent = plt.Rectangle((cx - box_w/2, cy - box_h/2),
                             0.22, box_h,
                             facecolor=border_col, zorder=3)
    ax2.add_patch(accent)
    # Step label (top)
    ax2.text(cx - box_w/2 + 0.45, cy + 0.35, label1,
             ha='left', va='center', fontsize=13, fontweight='bold',
             color=border_col, zorder=4)
    # Description (bottom)
    ax2.text(cx - box_w/2 + 0.45, cy - 0.25, label2,
             ha='left', va='center', fontsize=12,
             color='#1B2631', zorder=4)

# Arrows between steps
arrow_props = dict(arrowstyle='->', color='#1A5276',
                   lw=2.5, mutation_scale=22)
arrow_xs = [5, 5, 5, 5]
arrow_ys_start = [15.5 - box_h/2, 12.8 - box_h/2, 10.1 - box_h/2, 7.4 - box_h/2]
arrow_ys_end   = [12.8 + box_h/2, 10.1 + box_h/2, 7.4  + box_h/2, 4.7 + box_h/2]

for xs, ys, ye in zip(arrow_xs, arrow_ys_start, arrow_ys_end):
    ax2.annotate('', xy=(xs, ye + 0.05), xytext=(xs, ys - 0.05),
                 arrowprops=arrow_props, zorder=5)

# Side labels for arrows
side_labels = [
    'Data Mining\n& Filtering',
    'Virtual\nScreening',
    'ADMET &\nToxicity Check',
    'Safety\nValidation'
]
side_ys = [14.15, 11.45, 8.75, 6.05]
for lbl, sy in zip(side_labels, side_ys):
    ax2.text(7.2, sy, lbl, ha='left', va='center', fontsize=10.5,
             color='#5D6D7E', style='italic',
             bbox=dict(boxstyle='round,pad=0.3', facecolor='white',
                       edgecolor='#AED6F1', alpha=0.85))

# Bottom note
ax2.text(5, 3.1,
         'Output: Evidence-based, AI-optimized herbal formulation\nready for pre-clinical validation',
         ha='center', va='center', fontsize=12, color='#1A5276',
         fontweight='bold',
         bbox=dict(boxstyle='round,pad=0.5', facecolor='#D6EAF8',
                   edgecolor='#1A5276', linewidth=2))

# Footnote
ax2.text(5, 0.35,
         'Reference: Compiled from PubMed literature review (2021–2026)',
         ha='center', va='center', fontsize=10, color='#5D6D7E', style='italic')

plt.tight_layout()
plt.savefig('/home/daytona/workspace/afshan-abstract/visuals/flowchart_AI_Herbal_Workflow.png',
            dpi=200, bbox_inches='tight', facecolor='#EBF5FB')
plt.close()
print("Flowchart saved.")


# ─────────────────────────────────────────────────────────────
# FIGURE 3: PIE CHART — AI Techniques Used in Herbal Medicine
# ─────────────────────────────────────────────────────────────
fig3, ax3 = plt.subplots(figsize=(10, 8))
fig3.patch.set_facecolor('#EBF5FB')
ax3.set_facecolor('#EBF5FB')

labels = ['QSAR Modeling', 'Deep Learning\n(CNN/GNN)', 'LLMs &\nVGAE',
          'Network\nPharmacology+ML', 'Other AI\nMethods']
sizes  = [28, 25, 20, 18, 9]
colors = ['#1A5276', '#1E8449', '#6C3483', '#F0B429', '#5D8AA8']
explode = (0.04, 0.04, 0.04, 0.04, 0.04)

wedges, texts, autotexts = ax3.pie(
    sizes, explode=explode, labels=labels, colors=colors,
    autopct='%1.0f%%', startangle=140,
    pctdistance=0.78,
    wedgeprops=dict(edgecolor='white', linewidth=2.5),
    textprops=dict(fontsize=12, fontweight='bold', color='#1B2631')
)
for at in autotexts:
    at.set_fontsize(12)
    at.set_fontweight('bold')
    at.set_color('white')

ax3.set_title('AI Techniques Used in Herbal Medicine Research\n(Based on Literature Review 2021–2026)',
              fontsize=15, fontweight='bold', color='#1A5276', pad=20)

fig3.text(0.5, 0.01,
          'Source: Narrative review of 12 peer-reviewed articles (PubMed, IEEE, Scopus)',
          ha='center', fontsize=10, color='#5D6D7E', style='italic')

plt.tight_layout(rect=[0,0.04,1,1])
plt.savefig('/home/daytona/workspace/afshan-abstract/visuals/pie_chart_AI_Techniques.png',
            dpi=200, bbox_inches='tight', facecolor='#EBF5FB')
plt.close()
print("Pie chart saved.")
print("ALL DONE!")

Successfully replaced 1 instance(s) of the string in '/home/daytona/workspace/afshan-abstract/visuals/generate_visuals.py'