ls /tmp/workspace/attachments/181990fb-cda0-451c-a92e-6f773dc69cac/

python3 -c " from docx import Document doc = Document('/tmp/workspace/attachments/181990fb-cda0-451c-a92e-6f773dc69cac/A.THOUFIC AHAMED ABSTRACT (1) (1).docx') for p in doc.paragraphs: style = p.style.name if p.style else 'None' if p.text.strip(): print(f'[{style}] {p.text}') "

A professional academic e-poster in 16:9 widescreen ratio (1920x1080 landscape) on the topic "EVALUATION OF TRUE DRUG CAUSALITY USING CAUSAL AI IN REGULATORY PHARMACOVIGILANCE". The poster has a clean, modern medical/scientific design with a dark teal/navy blue header and white body background. LAYOUT (left to right, top to bottom): TOP HEADER BAR (full width, dark teal/navy): - Large bold white title: "EVALUATION OF TRUE DRUG CAUSALITY USING CAUSAL AI IN REGULATORY PHARMACOVIGILANCE" - Below title in smaller white text: "Thoufic Ahamed Abdul Aleem¹*, Mohamed Musammeel Mujifur Rahman², Vishnupriya Murali³" - Below authors: "SIR ISSAC NEWTON COLLEGE OF PHARMACY, Nagapattinam, Tamilnadu - 611 101" MAIN BODY (3-column layout on white/light grey background): LEFT COLUMN (teal accent border): Box 1 - INTRODUCTION (teal heading): "Pharmacovigilance plays a critical role in monitoring the safety of medicines by detecting, assessing and preventing adverse drug reactions (ADR). While Artificial Intelligence (AI) has significantly improved signal detection and case prioritization, conventional machine learning models mainly identify statistical associations rather than true causal relationships. This limitation reduces transparency and regulatory confidence in AI-assisted decision-making. Causal Artificial Intelligence (Causal AI) combined with Explainable AI (XAI) has emerged as a promising approach to improve causality assessment by providing interpretable and scientifically reliable evidence for regulatory pharmacovigilance." Box 2 - OBJECTIVE (teal heading): "To evaluate how explainable and causal AI can improve ADR signal detection, transparency, and regulatory decision-making." MIDDLE COLUMN (teal accent border): Box 3 - METHODOLOGY (teal heading): "A narrative review was conducted using recent literature on AI, Explainable AI (XAI), and Causal AI in pharmacovigilance. Current machine learning approaches, causal inference techniques, and regulatory challenges were critically analysed to identify their strengths, limitations, and future applications in drug safety assessment." A simple flow diagram showing: Literature Review → AI/XAI/Causal AI Analysis → Strengths & Limitations → Future Applications RIGHT COLUMN (teal accent border): Box 4 - RESULTS AND DISCUSSION (teal heading): "Causal AI integrates causal inference with explainable models to improve ADR signal detection, reduce bias, and support evidence-based regulatory decisions. However, challenges such as limited benchmark datasets, incomplete real-world data, and the need for standardized validation remain significant barriers to implementation." Box 5 - CONCLUSION (teal heading): "Causal AI represents the next generation of regulatory pharmacovigilance by moving beyond prediction toward scientifically validated causal inference. Integrating Causal AI with Explainable AI and established epidemiological methods can improve the accuracy, reliability of drug safety assessments." BOTTOM FOOTER BAR (dark teal/navy, full width): Left: KEYWORDS: "Pharmacovigilance | Causal AI | Explainable AI (XAI) | Adverse Drug Reactions | Causal Inference | Signal Detection | Regulatory Science | Drug Safety" Right: ACKNOWLEDGEMENT: "The authors gratefully acknowledge our institution and faculty members for their guidance and support." Clean academic scientific poster style, professional typography, teal and navy color scheme, white boxes with subtle shadows, all text clearly readable.

mkdir -p /tmp/workspace/eposter && python3 -c " from PIL import Image, ImageDraw, ImageFont import textwrap # Canvas: 1920x1080 (16:9) W, H = 1920, 1080 img = Image.new('RGB', (W, H), '#F5F7FA') draw = ImageDraw.Draw(img) # Colors NAVY = '#0A2342' TEAL = '#1A6B72' LTEAL = '#2A9DA5' WHITE = '#FFFFFF' LGRAY = '#EAF3F4' DGRAY = '#333333' MGRAY = '#666666' ACCENT = '#F0A500' # Try to load fonts import os def get_font(size, bold=False): candidates_bold = [ '/usr/share/fonts/truetype/dejavu/DejaVuSans-Bold.ttf', '/usr/share/fonts/truetype/liberation/LiberationSans-Bold.ttf', '/usr/share/fonts/truetype/freefont/FreeSansBold.ttf', ] candidates_reg = [ '/usr/share/fonts/truetype/dejavu/DejaVuSans.ttf', '/usr/share/fonts/truetype/liberation/LiberationSans-Regular.ttf', '/usr/share/fonts/truetype/freefont/FreeSans.ttf', ] for p in (candidates_bold if bold else candidates_reg): if os.path.exists(p): return ImageFont.truetype(p, size) return ImageFont.load_default() f_title = get_font(30, bold=True) f_authors = get_font(18) f_inst = get_font(16) f_head = get_font(20, bold=True) f_body = get_font(15) f_kw = get_font(14) f_small = get_font(13) # ── HEADER ────────────────────────────────────────────────────────────────── draw.rectangle([0, 0, W, 140], fill=NAVY) draw.rectangle([0, 140, W, 148], fill=ACCENT) title = 'EVALUATION OF TRUE DRUG CAUSALITY USING CAUSAL AI IN REGULATORY PHARMACOVIGILANCE' # wrap title tw_lines = textwrap.wrap(title, width=80) y = 14 for line in tw_lines: bbox = draw.textbbox((0,0), line, font=f_title) tw = bbox[2]-bbox[0] draw.text(((W-tw)//2, y), line, fill=WHITE, font=f_title) y += 34 authors = 'Thoufic Ahamed Abdul Aleem¹*, Mohamed Musammeel Mujifur Rahman², Vishnupriya Murali³' bbox = draw.textbbox((0,0), authors, font=f_authors) draw.text(((W-bbox[2]+bbox[0])//2, y+2), authors, fill='#B8E4E8', font=f_authors) y += 24 inst = 'SIR ISSAC NEWTON COLLEGE OF PHARMACY, Nagapattinam, Tamilnadu - 611 101' bbox = draw.textbbox((0,0), inst, font=f_inst) draw.text(((W-bbox[2]+bbox[0])//2, y+4), inst, fill='#D0EEF0', font=f_inst) # ── FOOTER ─────────────────────────────────────────────────────────────────── FH = 72 draw.rectangle([0, H-FH, W, H], fill=NAVY) draw.rectangle([0, H-FH, W, H-FH+4], fill=ACCENT) kw_text = 'KEYWORDS: Pharmacovigilance | Causal AI | Explainable AI (XAI) | Adverse Drug Reactions | Causal Inference | Signal Detection | Drug Safety' draw.text((30, H-FH+12), kw_text, fill='#B8E4E8', font=f_kw) ack = 'ACKNOWLEDGEMENT: The authors gratefully acknowledge our institution and faculty members for their guidance and support.' draw.text((30, H-FH+34), ack, fill='#D0EEF0', font=f_small) # ── THREE COLUMNS ──────────────────────────────────────────────────────────── COL_TOP = 158 COL_BOT = H - FH - 10 COL_H = COL_BOT - COL_TOP PAD = 14 COL1_X = 14 COL2_X = W//3 + 7 COL3_X = 2*W//3 + 7 COL_W = W//3 - 21 def draw_section(x, y, w, title_text, body_text, wrap_w=38, max_h=None): # Section box box_h = max_h if max_h else 200 # Header bar draw.rounded_rectangle([x, y, x+w, y+box_h], radius=10, fill=WHITE, outline=TEAL, width=2) draw.rounded_rectangle([x, y, x+w, y+34], radius=10, fill=TEAL) draw.rectangle([x, y+24, x+w, y+34], fill=TEAL) # flatten bottom corners of header th_bbox = draw.textbbox((0,0), title_text, font=f_head) draw.text((x + (w - (th_bbox[2]-th_bbox[0]))//2, y+7), title_text, fill=WHITE, font=f_head) # Body text lines = textwrap.wrap(body_text, width=wrap_w) ty = y + 42 for line in lines: draw.text((x+PAD, ty), line, fill=DGRAY, font=f_body) ty += 21 return ty # bottom y used # ── Column 1: Introduction + Objective ─────────────────────────────────────── intro_text = ('Pharmacovigilance plays a critical role in monitoring the safety of medicines ' 'by detecting, assessing and preventing adverse drug reactions (ADR). While ' 'Artificial Intelligence (AI) has significantly improved signal detection and ' 'case prioritization, conventional machine learning models mainly identify ' 'statistical associations rather than true causal relationships. This limitation ' 'reduces transparency and regulatory confidence in AI-assisted decision-making. ' 'Causal Artificial Intelligence (Causal AI) combined with Explainable AI (XAI) ' 'has emerged as a promising approach to improve causality assessment by providing ' 'interpretable and scientifically reliable evidence for regulatory pharmacovigilance.') intro_lines = textwrap.wrap(intro_text, width=38) intro_box_h = 46 + len(intro_lines)*21 + 10 draw_section(COL1_X, COL_TOP, COL_W, 'INTRODUCTION', intro_text, wrap_w=38, max_h=intro_box_h) obj_text = ('To evaluate how explainable and causal AI can improve ADR signal detection, ' 'transparency, and regulatory decision-making in pharmacovigilance.') obj_y = COL_TOP + intro_box_h + 12 obj_lines = textwrap.wrap(obj_text, width=38) obj_box_h = 46 + len(obj_lines)*21 + 10 draw_section(COL1_X, obj_y, COL_W, 'OBJECTIVE', obj_text, wrap_w=38, max_h=obj_box_h) # ── Column 2: Methodology ───────────────────────────────────────────────────── meth_text = ('A narrative review was conducted using recent literature on AI, Explainable AI ' '(XAI), and Causal AI in pharmacovigilance. Current machine learning approaches, ' 'causal inference techniques, and regulatory challenges were critically analysed ' 'to identify their strengths, limitations, and future applications in drug safety ' 'assessment.') meth_lines = textwrap.wrap(meth_text, width=38) meth_box_h = 46 + len(meth_lines)*21 + 10 draw_section(COL2_X, COL_TOP, COL_W, 'METHODOLOGY', meth_text, wrap_w=38, max_h=meth_box_h) # Flow diagram inside col2 below methodology box flow_y = COL_TOP + meth_box_h + 12 flow_box_h = 180 draw.rounded_rectangle([COL2_X, flow_y, COL2_X+COL_W, flow_y+flow_box_h], radius=10, fill=LGRAY, outline=LTEAL, width=2) # Title flow_title = 'STUDY WORKFLOW' ft_bbox = draw.textbbox((0,0), flow_title, font=f_head) draw.text((COL2_X + (COL_W-(ft_bbox[2]-ft_bbox[0]))//2, flow_y+8), flow_title, fill=TEAL, font=f_head) # Flow boxes steps = ['Literature Search', 'AI / XAI / Causal AI\nAnalysis', 'Strengths &\nLimitations', 'Future\nApplications'] sx = COL2_X + 18 sy = flow_y + 42 sw = (COL_W - 36 - 3*16) // 4 sh = 90 for i, step in enumerate(steps): bx = sx + i*(sw+16) draw.rounded_rectangle([bx, sy, bx+sw, sy+sh], radius=8, fill=TEAL, outline=NAVY, width=1) slines = step.split('\n') total_th = len(slines)*18 tsy = sy + (sh - total_th)//2 for sl in slines: sl_bbox = draw.textbbox((0,0), sl, font=f_small) draw.text((bx + (sw-(sl_bbox[2]-sl_bbox[0]))//2, tsy), sl, fill=WHITE, font=f_small) tsy += 18 if i < len(steps)-1: ax = bx + sw + 2 ay = sy + sh//2 draw.polygon([(ax, ay-6),(ax+12, ay),(ax, ay+6)], fill=ACCENT) # ── Column 3: Results + Conclusion ─────────────────────────────────────────── res_text = ('Causal AI integrates causal inference with explainable models to improve ADR ' 'signal detection, reduce bias, and support evidence-based regulatory decisions. ' 'However, challenges such as limited benchmark datasets, incomplete real-world ' 'data, and the need for standardized validation remain significant barriers to ' 'implementation.') res_lines = textwrap.wrap(res_text, width=38) res_box_h = 46 + len(res_lines)*21 + 10 draw_section(COL3_X, COL_TOP, COL_W, 'RESULTS & DISCUSSION', res_text, wrap_w=38, max_h=res_box_h) conc_text = ('Causal AI represents the next generation of regulatory pharmacovigilance by ' 'moving beyond prediction toward scientifically validated causal inference. ' 'Integrating Causal AI with Explainable AI and established epidemiological ' 'methods can improve the accuracy and reliability of drug safety assessments.') conc_y = COL_TOP + res_box_h + 12 conc_lines = textwrap.wrap(conc_text, width=38) conc_box_h = 46 + len(conc_lines)*21 + 10 draw_section(COL3_X, conc_y, COL_W, 'CONCLUSION', conc_text, wrap_w=38, max_h=conc_box_h) # Save out = '/tmp/workspace/eposter/Thoufic_Eposter.png' img.save(out, 'PNG', dpi=(150,150)) print('Saved to', out) print('Size:', img.size) "

python3 -c " from PIL import Image img = Image.open('/tmp/workspace/eposter/Thoufic_Eposter.png') print('Size:', img.size, '| Mode:', img.mode, '| Format:', img.format) import os print('File size:', round(os.path.getsize('/tmp/workspace/eposter/Thoufic_Eposter.png')/1024/1024, 2), 'MB') "

python3 -c " from PIL import Image img = Image.open('/tmp/workspace/eposter/Thoufic_Eposter.png') # crop a sample to see if content renders (check col1 area) crop = img.crop((14, 158, 640, 500)) crop.save('/tmp/workspace/eposter/preview_col1.png') print('Preview saved') "

mkdir -p /tmp/workspace/eposter && cat > /tmp/workspace/eposter/build_pptx.py << 'PYEOF' from pptx import Presentation from pptx.util import Inches, Pt, Emu from pptx.dml.color import RGBColor from pptx.enum.text import PP_ALIGN, MSO_ANCHOR from pptx.util import Inches, Pt import copy # ── Colours ────────────────────────────────────────────────────────────────── NAVY = RGBColor(0x0A, 0x23, 0x42) TEAL = RGBColor(0x1A, 0x6B, 0x72) LTEAL = RGBColor(0x2A, 0x9D, 0xA5) WHITE = RGBColor(0xFF, 0xFF, 0xFF) LGRAY = RGBColor(0xEA, 0xF3, 0xF4) DGRAY = RGBColor(0x33, 0x33, 0x33) MGRAY = RGBColor(0x55, 0x55, 0x55) ACCENT = RGBColor(0xF0, 0xA5, 0x00) LTBLUE = RGBColor(0xD0, 0xEE, 0xF0) prs = Presentation() prs.slide_width = Inches(13.333) prs.slide_height = Inches(7.5) blank = prs.slide_layouts[6] # fully blank # ── Helper: solid fill a shape ──────────────────────────────────────────────── def solid(shape, color): fill = shape.fill fill.solid() fill.fore_color.rgb = color # ── Helper: add text box ────────────────────────────────────────────────────── def add_tb(slide, x, y, w, h, text, size=12, bold=False, color=DGRAY, align=PP_ALIGN.LEFT, wrap=True, italic=False, anchor=MSO_ANCHOR.TOP): tb = slide.shapes.add_textbox(Inches(x), Inches(y), Inches(w), Inches(h)) tf = tb.text_frame tf.word_wrap = wrap tf.vertical_anchor = anchor tf.margin_left = Pt(4) tf.margin_right = Pt(4) tf.margin_top = Pt(3) tf.margin_bottom= Pt(3) p = tf.paragraphs[0] p.alignment = align run = p.add_run() run.text = text run.font.name = "Calibri" run.font.size = Pt(size) run.font.bold = bold run.font.italic= italic run.font.color.rgb = color return tb # ── Helper: add rectangle ───────────────────────────────────────────────────── def add_rect(slide, x, y, w, h, fill_color, line_color=None, line_width=Pt(0)): from pptx.util import Pt as PPt shape = slide.shapes.add_shape( 1, # MSO_SHAPE_TYPE.RECTANGLE Inches(x), Inches(y), Inches(w), Inches(h) ) solid(shape, fill_color) if line_color: shape.line.color.rgb = line_color shape.line.width = line_width else: shape.line.fill.background() return shape # ── Helper: multi-bullet textbox ───────────────────────────────────────────── def add_bullets(slide, x, y, w, h, bullets, size=11, color=DGRAY, head_color=TEAL, head=None, head_size=13): tb = slide.shapes.add_textbox(Inches(x), Inches(y), Inches(w), Inches(h)) tf = tb.text_frame tf.word_wrap = True tf.margin_left = Pt(6) tf.margin_right = Pt(4) tf.margin_top = Pt(4) tf.margin_bottom= Pt(4) first = True if head: p = tf.paragraphs[0] p.alignment = PP_ALIGN.LEFT run = p.add_run() run.text = head run.font.name = "Calibri" run.font.size = Pt(head_size) run.font.bold = True run.font.color.rgb = head_color first = False for b in bullets: p = tf.add_paragraph() if not first else tf.paragraphs[0] first = False p.alignment = PP_ALIGN.LEFT run = p.add_run() run.text = b run.font.name = "Calibri" run.font.size = Pt(size) run.font.color.rgb = color return tb # ════════════════════════════════════════════════════════════════════════════════ # SLIDE 1 – Title Slide # ════════════════════════════════════════════════════════════════════════════════ s1 = prs.slides.add_slide(blank) # Full background add_rect(s1, 0, 0, 13.333, 7.5, NAVY) # Teal accent band add_rect(s1, 0, 5.3, 13.333, 0.06, ACCENT) # Bottom strip add_rect(s1, 0, 5.36, 13.333, 2.14, TEAL) # Decorative circle top-left from pptx.util import Emu from pptx.enum.shapes import MSO_SHAPE_TYPE from pptx.dml.color import RGBColor as RGB cir = s1.shapes.add_shape(9, Inches(-1), Inches(-1), Inches(4), Inches(4)) # oval solid(cir, RGBColor(0x1A, 0x6B, 0x72)) cir.line.fill.background() cir2 = s1.shapes.add_shape(9, Inches(10.5), Inches(4.5), Inches(3.5), Inches(3.5)) solid(cir2, RGBColor(0x0D, 0x3A, 0x5C)) cir2.line.fill.background() # Tag line top add_tb(s1, 1, 0.6, 11, 0.4, "ACADEMIC RESEARCH POSTER PRESENTATION", size=11, bold=False, color=LTBLUE, align=PP_ALIGN.CENTER) # Main Title add_tb(s1, 0.6, 1.1, 12.1, 2.4, "EVALUATION OF TRUE DRUG CAUSALITY USING\nCAUSAL AI IN REGULATORY PHARMACOVIGILANCE", size=32, bold=True, color=WHITE, align=PP_ALIGN.CENTER) # Accent line add_rect(s1, 4.5, 3.55, 4.333, 0.07, ACCENT) # Authors add_tb(s1, 0.5, 3.75, 12.333, 0.5, "Thoufic Ahamed Abdul Aleem¹* | Mohamed Musammeel Mujifur Rahman² | Vishnupriya Murali³", size=14, bold=False, color=LTBLUE, align=PP_ALIGN.CENTER) # Institution box inst_box = add_rect(s1, 3.2, 4.35, 6.933, 0.65, LTEAL) inst_box.line.fill.background() add_tb(s1, 3.2, 4.35, 6.933, 0.65, "SIR ISSAC NEWTON COLLEGE OF PHARMACY\nNagapattinam, Tamilnadu - 611 101", size=12, bold=True, color=WHITE, align=PP_ALIGN.CENTER, anchor=MSO_ANCHOR.MIDDLE) # Bottom labels add_tb(s1, 0.5, 5.5, 6, 0.4, "Regulatory Pharmacovigilance | Drug Safety | Causal AI", size=11, color=LTBLUE, align=PP_ALIGN.LEFT) add_tb(s1, 7, 5.5, 6, 0.4, "Sir Issac Newton College of Pharmacy | 2026", size=11, color=LTBLUE, align=PP_ALIGN.RIGHT) # ════════════════════════════════════════════════════════════════════════════════ # SLIDE 2 – Introduction # ════════════════════════════════════════════════════════════════════════════════ s2 = prs.slides.add_slide(blank) add_rect(s2, 0, 0, 13.333, 7.5, RGBColor(0xF5, 0xF7, 0xFA)) add_rect(s2, 0, 0, 13.333, 1.0, NAVY) add_rect(s2, 0, 1.0, 13.333, 0.07, ACCENT) add_tb(s2, 0.3, 0.12, 12.7, 0.75, "INTRODUCTION", size=28, bold=True, color=WHITE, align=PP_ALIGN.LEFT) # Slide number add_tb(s2, 12.5, 0.15, 0.8, 0.4, "02", size=14, bold=True, color=LTBLUE, align=PP_ALIGN.RIGHT) # Left card add_rect(s2, 0.3, 1.25, 6.1, 5.8, WHITE) s2.shapes[-1].line.color.rgb = TEAL s2.shapes[-1].line.width = Pt(1.5) add_rect(s2, 0.3, 1.25, 6.1, 0.45, TEAL) s2.shapes[-1].line.fill.background() add_tb(s2, 0.35, 1.27, 6.0, 0.4, "Background", size=14, bold=True, color=WHITE) add_tb(s2, 0.45, 1.82, 5.85, 5.1, "Pharmacovigilance plays a critical role in monitoring the safety of medicines by " "detecting, assessing and preventing adverse drug reactions (ADR).\n\n" "While Artificial Intelligence (AI) has significantly improved signal detection and " "case prioritization, conventional machine learning models mainly identify statistical " "associations rather than true causal relationships.\n\n" "This limitation reduces transparency and regulatory confidence in AI-assisted " "decision-making.", size=12, color=DGRAY, wrap=True) # Right card add_rect(s2, 6.9, 1.25, 6.1, 5.8, WHITE) s2.shapes[-1].line.color.rgb = TEAL s2.shapes[-1].line.width = Pt(1.5) add_rect(s2, 6.9, 1.25, 6.1, 0.45, TEAL) s2.shapes[-1].line.fill.background() add_tb(s2, 6.95, 1.27, 6.0, 0.4, "The Challenge", size=14, bold=True, color=WHITE) add_tb(s2, 7.05, 1.82, 5.85, 5.1, "Causal Artificial Intelligence (Causal AI) combined with Explainable AI (XAI) has " "emerged as a promising approach to improve causality assessment.\n\n" "It provides interpretable and scientifically reliable evidence for regulatory " "pharmacovigilance decision-making.\n\n" "Moving beyond statistical correlation toward validated causal inference is the " "next frontier in drug safety science.", size=12, color=DGRAY, wrap=True) # ════════════════════════════════════════════════════════════════════════════════ # SLIDE 3 – Objective # ════════════════════════════════════════════════════════════════════════════════ s3 = prs.slides.add_slide(blank) add_rect(s3, 0, 0, 13.333, 7.5, RGBColor(0xF5, 0xF7, 0xFA)) add_rect(s3, 0, 0, 13.333, 1.0, NAVY) add_rect(s3, 0, 1.0, 13.333, 0.07, ACCENT) add_tb(s3, 0.3, 0.12, 12.7, 0.75, "OBJECTIVE", size=28, bold=True, color=WHITE) add_tb(s3, 12.5, 0.15, 0.8, 0.4, "03", size=14, bold=True, color=LTBLUE, align=PP_ALIGN.RIGHT) # Central objective statement add_rect(s3, 1.2, 1.4, 10.933, 1.8, TEAL) s3.shapes[-1].line.fill.background() add_tb(s3, 1.3, 1.45, 10.733, 1.7, "To evaluate how explainable and causal AI can improve ADR signal detection, " "transparency, and regulatory decision-making in pharmacovigilance.", size=18, bold=True, color=WHITE, align=PP_ALIGN.CENTER, anchor=MSO_ANCHOR.MIDDLE) # 3 sub-goal boxes goals = [ ("Signal Detection", "Enhance sensitivity and specificity of ADR signal detection beyond statistical association."), ("Transparency", "Improve interpretability of AI models for regulatory reviewers and pharmacovigilance teams."), ("Decision-Making", "Support evidence-based regulatory decisions with causally validated AI outputs."), ] for i, (title, desc) in enumerate(goals): bx = 0.4 + i * 4.3 add_rect(s3, bx, 3.5, 4.0, 3.5, WHITE) s3.shapes[-1].line.color.rgb = LTEAL s3.shapes[-1].line.width = Pt(1.5) add_rect(s3, bx, 3.5, 4.0, 0.55, LTEAL) s3.shapes[-1].line.fill.background() # number badge badge = s3.shapes.add_shape(9, Inches(bx+1.6), Inches(4.15), Inches(0.8), Inches(0.8)) solid(badge, ACCENT) badge.line.fill.background() add_tb(s3, bx+1.6, 4.15, 0.8, 0.8, str(i+1), size=18, bold=True, color=WHITE, align=PP_ALIGN.CENTER, anchor=MSO_ANCHOR.MIDDLE) add_tb(s3, bx+0.1, 3.52, 3.8, 0.5, title, size=14, bold=True, color=WHITE) add_tb(s3, bx+0.15, 5.1, 3.7, 1.8, desc, size=11, color=DGRAY, wrap=True) # ════════════════════════════════════════════════════════════════════════════════ # SLIDE 4 – Methodology # ════════════════════════════════════════════════════════════════════════════════ s4 = prs.slides.add_slide(blank) add_rect(s4, 0, 0, 13.333, 7.5, RGBColor(0xF5, 0xF7, 0xFA)) add_rect(s4, 0, 0, 13.333, 1.0, NAVY) add_rect(s4, 0, 1.0, 13.333, 0.07, ACCENT) add_tb(s4, 0.3, 0.12, 12.7, 0.75, "METHODOLOGY", size=28, bold=True, color=WHITE) add_tb(s4, 12.5, 0.15, 0.8, 0.4, "04", size=14, bold=True, color=LTBLUE, align=PP_ALIGN.RIGHT) # Design description box add_rect(s4, 0.3, 1.25, 12.733, 1.3, WHITE) s4.shapes[-1].line.color.rgb = TEAL s4.shapes[-1].line.width = Pt(1.5) add_rect(s4, 0.3, 1.25, 12.733, 0.45, TEAL) s4.shapes[-1].line.fill.background() add_tb(s4, 0.4, 1.27, 5, 0.4, "Study Design", size=14, bold=True, color=WHITE) add_tb(s4, 0.45, 1.8, 12.433, 0.65, "A narrative review was conducted using recent literature on Artificial Intelligence (AI), " "Explainable AI (XAI), and Causal AI in pharmacovigilance.", size=12, color=DGRAY, wrap=True) # Workflow arrow steps steps = [ ("1", "Literature\nSearch", "AI, XAI &\nCausal AI"), ("2", "Critical\nAnalysis", "ML approaches\n& techniques"), ("3", "Identify\nStrengths", "& Limitations"), ("4", "Future\nApplications", "Drug safety\nassessment"), ] arrow_y = 2.85 for i, (num, line1, line2) in enumerate(steps): bx = 0.4 + i * 3.2 # Box add_rect(s4, bx, arrow_y, 2.7, 2.4, TEAL) s4.shapes[-1].line.fill.background() # Number badge = s4.shapes.add_shape(9, Inches(bx+1.05), Inches(arrow_y+0.1), Inches(0.6), Inches(0.6)) solid(badge, ACCENT) badge.line.fill.background() add_tb(s4, bx+1.05, arrow_y+0.1, 0.6, 0.6, num, size=14, bold=True, color=WHITE, align=PP_ALIGN.CENTER, anchor=MSO_ANCHOR.MIDDLE) add_tb(s4, bx+0.1, arrow_y+0.8, 2.5, 0.5, line1, size=13, bold=True, color=WHITE, align=PP_ALIGN.CENTER) add_tb(s4, bx+0.1, arrow_y+1.35, 2.5, 0.9, line2, size=11, color=LTBLUE, align=PP_ALIGN.CENTER) # Arrow if i < len(steps)-1: arr = s4.shapes.add_shape(13, Inches(bx+2.7), Inches(arrow_y+0.9), Inches(0.5), Inches(0.5)) solid(arr, ACCENT) arr.line.fill.background() # Key areas analysed add_rect(s4, 0.3, 5.45, 12.733, 1.7, WHITE) s4.shapes[-1].line.color.rgb = LTEAL s4.shapes[-1].line.width = Pt(1) add_rect(s4, 0.3, 5.45, 12.733, 0.45, LTEAL) s4.shapes[-1].line.fill.background() add_tb(s4, 0.4, 5.47, 8, 0.4, "Key Areas Critically Analysed", size=13, bold=True, color=WHITE) areas = ["✔ Current machine learning approaches in pharmacovigilance", "✔ Causal inference techniques and their application", "✔ Regulatory challenges in AI-assisted safety decisions", "✔ Strengths, limitations and future applications in drug safety assessment"] add_tb(s4, 0.45, 5.98, 12.5, 1.1, " ".join(areas[:2]) + "\n" + " ".join(areas[2:]), size=11, color=DGRAY, wrap=True) # ════════════════════════════════════════════════════════════════════════════════ # SLIDE 5 – Results & Discussion # ════════════════════════════════════════════════════════════════════════════════ s5 = prs.slides.add_slide(blank) add_rect(s5, 0, 0, 13.333, 7.5, RGBColor(0xF5, 0xF7, 0xFA)) add_rect(s5, 0, 0, 13.333, 1.0, NAVY) add_rect(s5, 0, 1.0, 13.333, 0.07, ACCENT) add_tb(s5, 0.3, 0.12, 12.7, 0.75, "RESULTS & DISCUSSION", size=28, bold=True, color=WHITE) add_tb(s5, 12.5, 0.15, 0.8, 0.4, "05", size=14, bold=True, color=LTBLUE, align=PP_ALIGN.RIGHT) # Key finding banner add_rect(s5, 0.3, 1.25, 12.733, 0.9, TEAL) s5.shapes[-1].line.fill.background() add_tb(s5, 0.4, 1.28, 12.533, 0.85, "Causal AI integrates causal inference with explainable models to improve ADR signal " "detection, reduce bias, and support evidence-based regulatory decisions.", size=14, bold=True, color=WHITE, align=PP_ALIGN.CENTER, anchor=MSO_ANCHOR.MIDDLE) # Two column result + challenge res_items = [ ("Key Findings", [ "• Causal AI improves ADR signal detection beyond statistical methods", "• Explainable AI (XAI) increases model transparency and interpretability", "• Integration reduces bias in regulatory decision-making", "• Supports evidence-based pharmacovigilance assessment", "• Addresses limitations of conventional ML correlation-based models", ]), ] chal_items = [ ("Challenges Identified", [ "• Limited benchmark datasets for causal model validation", "• Incomplete real-world data affecting model reliability", "• Need for standardized validation frameworks", "• Significant barriers remain for regulatory implementation", "• Gap between AI research and regulatory practice", ]), ] for col_i, (cards, bx) in enumerate([( res_items, 0.3), (chal_items, 6.9)]): for title, bullets in cards: add_rect(s5, bx, 2.35, 6.1, 4.85, WHITE) s5.shapes[-1].line.color.rgb = TEAL if col_i == 0 else RGBColor(0xC0, 0x50, 0x20) s5.shapes[-1].line.width = Pt(1.5) hdr_color = TEAL if col_i == 0 else RGBColor(0xC0, 0x50, 0x20) add_rect(s5, bx, 2.35, 6.1, 0.48, hdr_color) s5.shapes[-1].line.fill.background() add_tb(s5, bx+0.1, 2.37, 5.9, 0.44, title, size=14, bold=True, color=WHITE) ty = 2.95 for b in bullets: add_tb(s5, bx+0.15, ty, 5.85, 0.52, b, size=11.5, color=DGRAY) ty += 0.7 # ════════════════════════════════════════════════════════════════════════════════ # SLIDE 6 – Conclusion # ════════════════════════════════════════════════════════════════════════════════ s6 = prs.slides.add_slide(blank) add_rect(s6, 0, 0, 13.333, 7.5, RGBColor(0xF5, 0xF7, 0xFA)) add_rect(s6, 0, 0, 13.333, 1.0, NAVY) add_rect(s6, 0, 1.0, 13.333, 0.07, ACCENT) add_tb(s6, 0.3, 0.12, 12.7, 0.75, "CONCLUSION", size=28, bold=True, color=WHITE) add_tb(s6, 12.5, 0.15, 0.8, 0.4, "06", size=14, bold=True, color=LTBLUE, align=PP_ALIGN.RIGHT) add_rect(s6, 0.5, 1.25, 12.333, 2.4, NAVY) s6.shapes[-1].line.fill.background() add_tb(s6, 0.65, 1.3, 12.033, 2.3, "Causal AI represents the next generation of regulatory pharmacovigilance by moving " "beyond prediction toward scientifically validated causal inference.\n\n" "Integrating Causal AI with Explainable AI and established epidemiological methods can " "improve the accuracy and reliability of drug safety assessments.", size=15, bold=False, color=WHITE, align=PP_ALIGN.CENTER, anchor=MSO_ANCHOR.MIDDLE) # Take-away points points = [ ("Next Generation", "Causal AI moves pharmacovigilance beyond statistical correlation to validated causal inference."), ("Integration", "Combining Causal AI + XAI + epidemiological methods yields superior drug safety evaluation."), ("Accuracy", "Improved accuracy and reliability of ADR assessments for regulatory confidence."), ("Future", "Standardized validation frameworks needed to unlock full regulatory implementation."), ] for i, (head, body) in enumerate(points): bx = 0.4 + (i % 2) * 6.5 by = 3.9 + (i // 2) * 1.7 add_rect(s6, bx, by, 6.0, 1.45, WHITE) s6.shapes[-1].line.color.rgb = LTEAL s6.shapes[-1].line.width = Pt(1.2) add_rect(s6, bx, by, 0.15, 1.45, TEAL) s6.shapes[-1].line.fill.background() add_tb(s6, bx+0.25, by+0.08, 5.65, 0.38, head, size=13, bold=True, color=TEAL) add_tb(s6, bx+0.25, by+0.5, 5.65, 0.85, body, size=11, color=DGRAY, wrap=True) # ════════════════════════════════════════════════════════════════════════════════ # SLIDE 7 – Keywords & Acknowledgement # ════════════════════════════════════════════════════════════════════════════════ s7 = prs.slides.add_slide(blank) add_rect(s7, 0, 0, 13.333, 7.5, NAVY) add_rect(s7, 0, 0, 13.333, 0.07, ACCENT) # Decorative elements cir3 = s7.shapes.add_shape(9, Inches(10), Inches(-1), Inches(5), Inches(5)) solid(cir3, RGBColor(0x1A, 0x6B, 0x72)) cir3.line.fill.background() cir4 = s7.shapes.add_shape(9, Inches(-0.5), Inches(5.5), Inches(3), Inches(3)) solid(cir4, RGBColor(0x0D, 0x3A, 0x5C)) cir4.line.fill.background() add_tb(s7, 0.5, 0.5, 12.333, 0.7, "KEYWORDS", size=22, bold=True, color=ACCENT, align=PP_ALIGN.CENTER) keywords = [ "Pharmacovigilance", "Causal Artificial Intelligence (Causal AI)", "Explainable Artificial Intelligence (XAI)", "Adverse Drug Reactions (ADR)", "Causal Inference", "Signal Detection", "Regulatory Science", "Drug Safety" ] kw_row1 = " • ".join(keywords[:4]) kw_row2 = " • ".join(keywords[4:]) add_tb(s7, 0.5, 1.3, 12.333, 0.45, kw_row1, size=12, color=LTBLUE, align=PP_ALIGN.CENTER) add_tb(s7, 0.5, 1.8, 12.333, 0.45, kw_row2, size=12, color=LTBLUE, align=PP_ALIGN.CENTER) add_rect(s7, 2.5, 2.45, 8.333, 0.07, ACCENT) add_tb(s7, 0.5, 2.7, 12.333, 0.6, "ACKNOWLEDGEMENT", size=22, bold=True, color=ACCENT, align=PP_ALIGN.CENTER) add_tb(s7, 1.0, 3.4, 11.333, 1.1, "The authors gratefully acknowledge our institution and faculty members for their " "guidance and support in preparing this presentation.", size=13, color=LTBLUE, align=PP_ALIGN.CENTER, wrap=True) add_rect(s7, 2.5, 4.6, 8.333, 0.07, TEAL) # Authors & affil add_tb(s7, 0.5, 4.85, 12.333, 0.45, "Thoufic Ahamed Abdul Aleem¹* | Mohamed Musammeel Mujifur Rahman² | Vishnupriya Murali³", size=13, bold=True, color=WHITE, align=PP_ALIGN.CENTER) add_tb(s7, 0.5, 5.35, 12.333, 0.45, "SIR ISSAC NEWTON COLLEGE OF PHARMACY, Nagapattinam, Tamilnadu - 611 101", size=12, color=LTBLUE, align=PP_ALIGN.CENTER) add_rect(s7, 0, 6.1, 13.333, 0.07, ACCENT) add_tb(s7, 0.5, 6.25, 12.333, 0.4, "Evaluation of True Drug Causality Using Causal AI in Regulatory Pharmacovigilance | 2026", size=10, color=RGBColor(0x80, 0x9A, 0xAA), align=PP_ALIGN.CENTER) # Save out = "/tmp/workspace/eposter/Thoufic_CausalAI_Presentation.pptx" prs.save(out) print("Saved:", out) PYEOF python3 /tmp/workspace/eposter/build_pptx.py

python3 -c " from pptx import Presentation import os prs = Presentation('/tmp/workspace/eposter/Thoufic_CausalAI_Presentation.pptx') print('Slides:', len(prs.slides)) print('Dimensions:', round(prs.slide_width.inches,2), 'x', round(prs.slide_height.inches,2), 'inches (16:9)') print('File size:', round(os.path.getsize('/tmp/workspace/eposter/Thoufic_CausalAI_Presentation.pptx')/1024, 1), 'KB') for i, slide in enumerate(prs.slides, 1): texts = [] for shape in slide.shapes: if shape.has_text_frame and shape.text_frame.text.strip(): texts.append(shape.text_frame.text[:50].replace(chr(10),' ')) print(f' Slide {i}: {texts[0] if texts else \"(empty)\"}') "