from __future__ import annotations import json import sys from pathlib import Path import matplotlib.pyplot as plt import numpy as np from matplotlib.lines import Line2D from matplotlib.patches import FancyArrowPatch, Rectangle WORK = Path(__file__).resolve().parent ASSETS = WORK / "assets" ASSETS.mkdir(parents=True, exist_ok=True) STYLE_DIR = Path(r"C:\Users\Deriv\Desktop\GreyScienx\skills\greyscienx-editorial-pdf\scripts") sys.path.insert(0, str(STYLE_DIR)) from greyscienx_style import add_figure_header, configure_matplotlib, save_figure, style_axis # noqa: E402 TOKENS = configure_matplotlib(Path(r"C:\Users\Deriv\Desktop\GreyScienx\app\globals.css")) CORAL = TOKENS["coral"] BLACK = TOKENS["black"] WHITE = TOKENS["white"] G100 = TOKENS["grey-100"] G300 = TOKENS["grey-300"] G500 = TOKENS["grey-500"] G700 = TOKENS["grey-700"] FUND = 253.54 STAGES = [ "Resource", "Primary\nrefining", "Advanced\nmaterials", "Core\ncomponents", "System\nassembly", "Services &\nrecycling", ] VALUE_WEIGHTS = np.array([0.12, 0.16, 0.20, 0.22, 0.22, 0.08]) LOCAL_SHARE = { "Current footprint": { "Manganese": [95, 45, 8, 4, 30, 8], "PGM / hydrogen": [92, 80, 45, 18, 20, 35], "Vanadium / VRFB": [70, 60, 55, 20, 35, 30], }, "Focused integration": { "Manganese": [95, 65, 65, 35, 70, 45], "PGM / hydrogen": [95, 88, 72, 52, 58, 70], "Vanadium / VRFB": [80, 75, 80, 50, 70, 65], }, "Frontier partnership": { "Manganese": [95, 80, 82, 60, 82, 70], "PGM / hydrogen": [98, 92, 85, 75, 80, 86], "Vanadium / VRFB": [90, 85, 90, 72, 85, 82], }, } CHAIN_WEIGHTS = {"Manganese": 0.40, "PGM / hydrogen": 0.35, "Vanadium / VRFB": 0.25} ITEMS = [ ("Manganese ore", "Manganese", 91, 38, 20), ("Manganese alloys", "Manganese", 63, 54, 48), ("High-purity Mn sulphate", "Manganese", 74, 91, 43), ("Precursor / cathode material", "Manganese", 58, 84, 67), ("Battery cells", "Manganese", 30, 88, 96), ("Pack, BMS and integration", "Manganese", 68, 70, 44), ("PGM refining", "PGM / hydrogen", 86, 58, 42), ("Catalyst powder and inks", "PGM / hydrogen", 72, 86, 38), ("Membrane-electrode assemblies", "PGM / hydrogen", 58, 90, 55), ("Electrolyser / fuel-cell stacks", "PGM / hydrogen", 46, 82, 72), ("Hydrogen system integration", "PGM / hydrogen", 52, 77, 78), ("PGM recycling", "PGM / hydrogen", 78, 74, 34), ("Vanadium pentoxide", "Vanadium / VRFB", 70, 59, 45), ("VRFB electrolyte", "Vanadium / VRFB", 79, 79, 32), ("Stack materials", "Vanadium / VRFB", 50, 72, 52), ("VRFB stacks", "Vanadium / VRFB", 57, 82, 61), ("VRFB system integration", "Vanadium / VRFB", 72, 75, 55), ("Electrolyte leasing / recovery", "Vanadium / VRFB", 73, 77, 36), ] CAPITAL = { "Power, water, rail and ports": 0.30, "Manganese materials hub": 0.17, "PGM and hydrogen components": 0.15, "Vanadium and VRFB chain": 0.11, "Shared chemicals and parks": 0.10, "Supplier and working capital": 0.07, "R&D, testing, skills and IP": 0.06, "Contingency and closure reserve": 0.04, } BOTTLENECKS = { "Reliable, competitively priced power": 95, "Bankable long-term offtake": 92, "Project governance and commissioning": 90, "Technology licensing and operating partner": 84, "Rail, port and specialised logistics": 80, "Water, reagents and effluent treatment": 75, "Process-engineering and production skills": 72, "Supplier finance and working capital": 64, } def header(fig, title, subtitle): add_figure_header(fig, title, subtitle, field="GREYSCIENCX / PAPER 4", tokens=TOKENS) def finish(fig, filename, *, left=0.10, right=0.97, top=0.77, bottom=0.14): fig.subplots_adjust(left=left, right=right, top=top, bottom=bottom) save_figure(fig, ASSETS / filename, dpi=270) plt.close(fig) def capture(chain, scenario): return float(np.dot(VALUE_WEIGHTS, np.array(LOCAL_SHARE[scenario][chain]) / 100)) def portfolio_capture(scenario): return float(sum(CHAIN_WEIGHTS[c] * capture(c, scenario) for c in CHAIN_WEIGHTS)) def make_endowment_gap(): minerals = ["Platinum", "Manganese", "Vanadium"] reserve = np.array([88, 80, 32]) production = np.array([70, 40, 7]) y = np.arange(len(minerals)) fig, ax = plt.subplots(figsize=(9.5, 5.5)) header(fig, "A mineral endowment is a head start, not an industrial complex", "South African shares cited by official sources; reserve estimates and 2024 mined production are different measures") ax.barh(y + 0.18, reserve, height=0.34, color=CORAL, label="Reserve-share estimate cited in 2025 strategy") ax.barh(y - 0.18, production, height=0.34, color=BLACK, label="Share of 2024 world mined production") for i, (r, p) in enumerate(zip(reserve, production)): ax.text(r + 1.5, i + 0.18, f"{r}%", va="center", fontsize=8, fontweight="bold") ax.text(p + 1.5, i - 0.18, f"{p}%", va="center", fontsize=8, fontweight="bold") ax.set_yticks(y, minerals) ax.set_xlim(0, 100) ax.set_xlabel("Share of reported world total (%)") ax.legend(frameon=False, fontsize=7.2, loc="lower right") style_axis(ax, TOKENS, grid_axis="x") finish(fig, "endowment-gap.png", left=0.16) def draw_node(ax, x, y, w, h, text, face, edge=BLACK, txt=BLACK, lw=0.8): rect = Rectangle((x, y), w, h, facecolor=face, edgecolor=edge, linewidth=lw) ax.add_patch(rect) ax.text(x + w / 2, y + h / 2, text, ha="center", va="center", fontsize=6.8, color=txt, fontweight="bold") def make_system_map(): fig, ax = plt.subplots(figsize=(10.3, 6.1)) header(fig, "Three mineral chains can share one industrial spine", "Coral stages are the strongest near-term localisation targets; grey stages need partnership, scale or imported technology") ax.set_xlim(0, 12) ax.set_ylim(0, 7.3) ax.axis("off") rows = { "MANGANESE": (5.5, ["Ore", "Purified\nMn", "HPMSM", "pCAM / CAM", "Cells", "Packs / BMS"]), "PGM / H2": (3.4, ["Mine", "Refine", "Catalyst\ninks", "MEA", "Stacks", "Systems"]), "VANADIUM": (1.3, ["Mine", "V2O5", "Electrolyte", "Stack\nmaterials", "VRFB stack", "Grid system"]), } target_indices = {"MANGANESE": {0, 1, 2, 5}, "PGM / H2": {0, 1, 2, 3}, "VANADIUM": {1, 2, 5}} for label, (y, nodes) in rows.items(): ax.text(0.05, y + 0.42, label, ha="left", va="center", fontsize=8.5, fontweight="bold") for i, node in enumerate(nodes): x = 1.45 + i * 1.72 face = CORAL if i in target_indices[label] else G100 txt = WHITE if face == CORAL else BLACK draw_node(ax, x, y, 1.28, 0.82, node, face, txt=txt) if i < len(nodes) - 1: ax.add_patch(FancyArrowPatch((x + 1.29, y + 0.41), (x + 1.68, y + 0.41), arrowstyle="->", mutation_scale=8, linewidth=0.8, color=G700)) ax.add_patch(Rectangle((1.45, 0.03), 9.88, 0.62, facecolor=BLACK, edgecolor=BLACK)) ax.text(6.39, 0.34, "SHARED INDUSTRIAL COMMONS: POWER | WATER | ACIDS & GASES | ENGINEERING | TESTING | RAIL & PORTS | RECYCLING", ha="center", va="center", color=WHITE, fontsize=7.1, fontweight="bold") for x in [2.1, 5.54, 8.98]: ax.add_patch(FancyArrowPatch((x, 0.66), (x, 1.2), arrowstyle="->", mutation_scale=8, linewidth=0.8, color=CORAL)) finish(fig, "integrated-system-map.png", left=0.04, right=0.98, bottom=0.07, top=0.77) def make_readiness_heatmap(): labels = [i[0] for i in ITEMS] readiness = np.array([i[2] for i in ITEMS]) upside = np.array([i[3] for i in ITEMS]) matrix = np.column_stack([readiness, upside]) fig, ax = plt.subplots(figsize=(9.2, 8.3)) header(fig, "The best targets sit between the mine and the finished system", "Armchair scores out of 100: readiness measures local feasibility; upside measures strategic value if capability is built") cmap = plt.matplotlib.colors.LinearSegmentedColormap.from_list("grey_coral", [G700, G100, CORAL]) im = ax.imshow(matrix, aspect="auto", cmap=cmap, vmin=20, vmax=95) ax.set_xticks([0, 1], ["Local readiness", "Strategic upside"]) ax.set_yticks(np.arange(len(labels)), labels) for r in range(matrix.shape[0]): for c in range(matrix.shape[1]): val = int(matrix[r, c]) colour = WHITE if val >= 74 or val <= 35 else BLACK ax.text(c, r, str(val), ha="center", va="center", fontsize=7.2, fontweight="bold", color=colour) for boundary in [5.5, 11.5]: ax.axhline(boundary, color=WHITE, linewidth=3) ax.tick_params(axis="y", labelsize=7.0) ax.tick_params(axis="x", labelsize=8) cbar = fig.colorbar(im, ax=ax, fraction=0.028, pad=0.025) cbar.set_label("Score", fontsize=7.5) finish(fig, "readiness-heatmap.png", left=0.30, right=0.93, top=0.82, bottom=0.12) def make_priority_frontier(): fig, ax = plt.subplots(figsize=(9.6, 6.1)) header(fig, "Build what is ready; partner where the prize is high", "Bubble size indicates relative capital intensity; scores are transparent scenario judgements, not forecasts") chain_style = { "Manganese": (CORAL, "o"), "PGM / hydrogen": (BLACK, "s"), "Vanadium / VRFB": (G500, "D"), } selected = [2, 3, 4, 7, 8, 9, 13, 15, 16] offsets = { "High-purity Mn sulphate": (1.1, 1.0, "left"), "Precursor / cathode material": (1.1, -2.0, "left"), "Battery cells": (1.1, 1.0, "left"), "Catalyst powder and inks": (1.1, -2.0, "left"), "Membrane-electrode assemblies": (1.1, 1.0, "left"), "Electrolyser / fuel-cell stacks": (-1.1, -2.2, "right"), "VRFB electrolyte": (1.1, -2.1, "left"), "VRFB stacks": (1.1, -3.1, "left"), "VRFB system integration": (1.1, -1.0, "left"), } for idx in selected: name, chain, ready, upside, capital = ITEMS[idx] colour, marker = chain_style[chain] ax.scatter(ready, upside, s=35 + capital * 3.0, color=colour, marker=marker, alpha=0.86, edgecolor=WHITE, linewidth=0.8, zorder=3) dx, dy, align = offsets[name] ax.text(ready + dx, upside + dy, name.replace(" and ", " & "), fontsize=6.2, color=BLACK, ha=align) ax.axvline(65, color=G300, linewidth=1) ax.axhline(75, color=G300, linewidth=1) ax.text(83, 94, "BUILD / SCALE", ha="center", fontsize=8, fontweight="bold", color=CORAL) ax.text(48, 94, "PARTNER / LICENSE", ha="center", fontsize=8, fontweight="bold", color=BLACK) ax.text(83, 61, "SELECTIVE LOCALISATION", ha="center", fontsize=7.5, fontweight="bold", color=G700) ax.text(48, 61, "PILOT / MONITOR", ha="center", fontsize=7.5, fontweight="bold", color=G700) ax.set_xlim(25, 100) ax.set_ylim(55, 98) ax.set_xlabel("Local readiness score") ax.set_ylabel("Strategic upside score") legend = [Line2D([0], [0], marker=m, color="none", markerfacecolor=c, label=k, markersize=6) for k, (c, m) in chain_style.items()] ax.legend(handles=legend, frameon=False, fontsize=7, loc="lower right") style_axis(ax, TOKENS, grid_axis="both") finish(fig, "priority-frontier.png", left=0.10, right=0.98, top=0.78, bottom=0.12) def make_capture_ladder(): scenarios = list(LOCAL_SHARE) chains = list(CHAIN_WEIGHTS) x = np.arange(len(chains)) width = 0.23 colours = [G500, BLACK, CORAL] hatches = ["..", "//", ""] fig, ax = plt.subplots(figsize=(9.5, 5.7)) header(fig, "A focused strategy could localise two-thirds of the modelled value", "Illustrative share of value added retained locally; stage weights are held constant across three capability scenarios") for i, scenario in enumerate(scenarios): vals = [capture(c, scenario) * 100 for c in chains] bars = ax.bar(x + (i - 1) * width, vals, width=width, color=colours[i], hatch=hatches[i], label=scenario) for b, v in zip(bars, vals): ax.text(b.get_x() + b.get_width() / 2, v + 1.2, f"{v:.0f}%", ha="center", fontsize=7, fontweight="bold") ax.set_xticks(x, chains) ax.set_ylim(0, 100) ax.set_ylabel("Local share of modelled value added (%)") ax.legend(frameon=False, fontsize=7.2, ncol=3, loc="upper left") style_axis(ax, TOKENS, grid_axis="y") finish(fig, "capture-ladder.png") def make_stage_capture(): scenario = "Focused integration" weighted = np.zeros(len(STAGES)) for chain, weight in CHAIN_WEIGHTS.items(): weighted += weight * np.array(LOCAL_SHARE[scenario][chain]) retained = VALUE_WEIGHTS * weighted uncaptured = VALUE_WEIGHTS * (100 - weighted) x = np.arange(len(STAGES)) fig, ax = plt.subplots(figsize=(9.6, 5.7)) header(fig, "The deepest leakage remains in components and advanced manufacturing", "Focused-integration scenario; each bar is that stage's share of the total value pool") ax.bar(x, retained, color=CORAL, label="Retained locally") ax.bar(x, uncaptured, bottom=retained, color=G300, label="Imported or captured abroad") for i, (r, u) in enumerate(zip(retained, uncaptured)): ax.text(i, r / 2, f"{r:.1f}", ha="center", va="center", fontsize=7, fontweight="bold", color=WHITE if r > 7 else BLACK) ax.set_xticks(x, STAGES) ax.set_ylabel("Share of total chain value pool (percentage points)") ax.legend(frameon=False, fontsize=7.2, loc="upper right") style_axis(ax, TOKENS, grid_axis="y") finish(fig, "stage-capture.png", bottom=0.17) def make_capital_portfolio(): labels = list(CAPITAL) values = np.array([CAPITAL[k] * FUND for k in labels]) y = np.arange(len(labels))[::-1] fig, ax = plt.subplots(figsize=(9.6, 4.7)) header(fig, "The inherited R253.5 billion cannot be spent only on factories", "Illustrative portfolio in constant 2026 rand; 40% is reserved for shared infrastructure, industrial commons and closure risk") colours = [CORAL if i < 4 else BLACK if i < 6 else G500 for i in range(len(labels))] ax.barh(y, values, color=colours) for yy, val in zip(y, values): ax.text(val + 1.2, yy, f"R{val:.1f}bn", va="center", fontsize=7.5, fontweight="bold") ax.set_yticks(y, labels) ax.set_xlim(0, 86) ax.set_xlabel("Capital allocation, R billion (2026 rand)") style_axis(ax, TOKENS, grid_axis="x") finish(fig, "capital-portfolio.png", left=0.31, right=0.95) def make_bottleneck_bar(): labels = list(BOTTLENECKS) values = np.array(list(BOTTLENECKS.values())) y = np.arange(len(labels))[::-1] fig, ax = plt.subplots(figsize=(9.6, 5.8)) header(fig, "The binding constraints are institutional before they are geological", "Indicative criticality score: how strongly failure of one condition can strand otherwise viable downstream investment") bars = ax.barh(y, values, color=[CORAL if v >= 90 else BLACK if v >= 80 else G500 for v in values]) for bar, val in zip(bars, values): ax.text(val + 1, bar.get_y() + bar.get_height() / 2, str(val), va="center", fontsize=7.5, fontweight="bold") ax.set_yticks(y, labels) ax.set_xlim(0, 103) ax.set_xlabel("Constraint criticality score (0-100)") style_axis(ax, TOKENS, grid_axis="x") finish(fig, "bottlenecks.png", left=0.34, right=0.95) def make_policy_choice(): labels = ["Current footprint", "Focused integration", "Frontier partnership", "Forced autarky"] nominal = [portfolio_capture("Current footprint") * 100, portfolio_capture("Focused integration") * 100, portfolio_capture("Frontier partnership") * 100, 78] viable = [nominal[0] * 0.88, nominal[1] * 0.93, nominal[2] * 0.90, nominal[3] * 0.67] x = np.arange(len(labels)) fig, ax = plt.subplots(figsize=(9.5, 5.7)) header(fig, "Local content is not the same as viable local value", "The autarky case loses utilisation and market access; partnership retains less nominal control but more realised value") ax.bar(x, nominal, color=G300, label="Nominal local-content ambition") ax.bar(x, viable, color=[G500, BLACK, CORAL, G700], width=0.58, label="Realised local value after viability factor") for i, v in enumerate(viable): ax.text(i, v + 1.5, f"{v:.0f}%", ha="center", fontsize=8, fontweight="bold") ax.set_xticks(x, ["Current\nfootprint", "Focused\nintegration", "Frontier\npartnership", "Forced\nautarky"]) ax.set_ylim(0, 100) ax.set_ylabel("Share of modelled value pool (%)") ax.legend(frameon=False, fontsize=7.2, loc="upper left") style_axis(ax, TOKENS, grid_axis="y") finish(fig, "policy-choice.png", bottom=0.17) def make_sequence(): lanes = [ ("Utilities, corridors and offtake", 2026, 2032, CORAL), ("HPMSM, electrolyte and catalysts", 2027, 2033, CORAL), ("Packs, balance-of-plant and systems", 2028, 2036, BLACK), ("Cathodes, MEAs and stack materials", 2030, 2039, BLACK), ("Selective cells and complete stacks", 2034, 2044, G500), ("Recycling and second-life networks", 2030, 2045, G500), ] fig, ax = plt.subplots(figsize=(9.6, 5.7)) header(fig, "Sequencing beats simultaneous localisation", "The plan earns the right to move downstream by first proving inputs, demand, quality and plant utilisation") y = np.arange(len(lanes))[::-1] for yy, (name, start, end, colour) in zip(y, lanes): ax.barh(yy, end - start, left=start, height=0.52, color=colour) ax.text(start + 0.25, yy, name, va="center", ha="left", fontsize=7.1, color=WHITE, fontweight="bold") ax.text(end + 0.18, yy, str(end), va="center", fontsize=7, color=BLACK) for year in [2030, 2035, 2040]: ax.axvline(year, color=G300, linewidth=0.8, zorder=-1) ax.set_yticks([]) ax.set_xlim(2025.5, 2046.5) ax.set_xlabel("Illustrative capability-building window") style_axis(ax, TOKENS, grid_axis="x") finish(fig, "sequence.png", left=0.08, right=0.96) def make_value_chain_table_figure(): rows = [ ("Manganese", "HPMSM", "pCAM / CAM with partner", "Cells", "Packs, BMS, recycling"), ("PGM / hydrogen", "Refining and catalyst inks", "MEA with licence", "Commodity electrolyser factories", "Systems, recycling, industrial use"), ("Vanadium / VRFB", "V2O5 and electrolyte", "Stack materials and stacks", "Membranes at small scale", "Integration, leasing, recovery"), ] fig, ax = plt.subplots(figsize=(10.2, 4.8)) header(fig, "The practical strategy is make, partner, import and learn", "A mineral-by-mineral decision rule avoids treating every downstream stage as equally localisable") ax.axis("off") columns = ["CHAIN", "MAKE / SCALE", "PARTNER / LICENSE", "IMPORT FOR NOW", "CLOSE THE LOOP"] table = ax.table(cellText=rows, colLabels=columns, loc="center", cellLoc="left", colLoc="left", colWidths=[0.15, 0.22, 0.23, 0.20, 0.23]) table.auto_set_font_size(False) table.set_fontsize(7.0) table.scale(1, 2.0) for (r, c), cell in table.get_celld().items(): cell.set_linewidth(0.4) cell.set_edgecolor(WHITE) if r == 0: cell.set_facecolor(BLACK) cell.get_text().set_color(WHITE) cell.get_text().set_fontweight("bold") else: cell.set_facecolor(G100 if r % 2 == 0 else WHITE) if c == 1: cell.get_text().set_color(CORAL) cell.get_text().set_fontweight("bold") finish(fig, "make-partner-import.png", left=0.03, right=0.97, top=0.70, bottom=0.10) def write_results(): results = { "resource_envelope_r_bn_2026": FUND, "stage_value_weights": dict(zip(STAGES, [round(float(v), 2) for v in VALUE_WEIGHTS])), "portfolio_local_value_share": {s: round(portfolio_capture(s), 4) for s in LOCAL_SHARE}, "chain_local_value_share": { s: {c: round(capture(c, s), 4) for c in CHAIN_WEIGHTS} for s in LOCAL_SHARE }, "capital_allocation_r_bn": {k: round(v * FUND, 2) for k, v in CAPITAL.items()}, "bottleneck_scores": BOTTLENECKS, "readiness_items": [ {"item": n, "chain": c, "readiness": r, "strategic_upside": u, "capital_intensity": cap} for n, c, r, u, cap in ITEMS ], "interpretation": { "focused_capture_percent": round(portfolio_capture("Focused integration") * 100, 1), "frontier_capture_percent": round(portfolio_capture("Frontier partnership") * 100, 1), "current_capture_percent": round(portfolio_capture("Current footprint") * 100, 1), "note": "All localisation shares, stage weights, readiness scores and capital allocations are transparent scenarios, not forecasts or investment advice.", }, } (WORK / "model_results.json").write_text(json.dumps(results, indent=2), encoding="utf-8") def main(): make_endowment_gap() make_system_map() make_readiness_heatmap() make_priority_frontier() make_capture_ladder() make_stage_capture() make_capital_portfolio() make_bottleneck_bar() make_policy_choice() make_sequence() make_value_chain_table_figure() write_results() print(WORK / "model_results.json") if __name__ == "__main__": main()