dsh-office-kit 0.2.0 → 0.3.1

package/assets/skills/office-presentations/container_tools/preview_render.py

@@ -0,0 +1,382 @@
+"""Draft slide preview: rasterise a .pptx into one PNG per slide without Office.
+
+Purpose: `render_slides.py` needs pixels to look at, and a machine may have no
+office renderer at all. This module reads the package with
+python-pptx and redraws every shape onto a PyMuPDF page: rectangles, text at
+its declared size, table grids, images, and a bar/line/pie sketch of charts.
+It then rasterises each page to PNG.
+
+What it proves and what it does not:
+  * proves geometry — what sits where, what collides, what leaves the slide,
+    and which text boxes ran out of room (PyMuPDF tells us when a textbox
+    truncates);
+  * does NOT prove typography — fonts fall back to a built-in face and theme
+    colours/fills that python-pptx cannot resolve are approximated.
+
+Exit codes: 0 ok, 2 unreadable input, 3 nothing rasterised.
+"""
+
+from __future__ import annotations
+
+import json
+import sys
+from pathlib import Path
+from typing import Any
+
+sys.path.insert(0, str(Path(__file__).resolve().parent))
+
+try:  # PyMuPDF ships both names; prefer the current one
+    import pymupdf as fitz  # noqa: E402
+except ImportError:  # pragma: no cover - older wheels only expose fitz
+    import fitz  # noqa: E402
+
+from ooxml_helpers import (  # noqa: E402
+    EMU_PER_INCH,
+    box,
+    default_size_pt,
+    estimated_text_height_in,
+    has_text_frame,
+    is_chart,
+    is_picture,
+    is_table,
+    iter_shapes,
+    open_presentation,
+    placeholder_kind,
+    run_size_pt,
+    slide_size_in,
+    text_of,
+    WIDE_CHAR,
+)
+
+# Windows consoles default to a legacy code page; every payload is UTF-8 JSON.
+for _stream in (sys.stdout, sys.stderr):
+    try:
+        _stream.reconfigure(encoding="utf-8", errors="replace")
+    except (AttributeError, ValueError):
+        pass
+
+INK = (0.13, 0.15, 0.18)
+MUTED = (0.42, 0.45, 0.50)
+RULE = (0.80, 0.82, 0.85)
+ACCENT = (0.16, 0.36, 0.62)
+PANEL = (0.96, 0.97, 0.98)
+CJK_FONT = "china-s"
+
+
+def _has_cjk(text: str) -> bool:
+    return any(WIDE_CHAR.match(char) for char in text)
+
+
+def _font_for(text: str, size: float) -> tuple[str, str | None]:
+    if _has_cjk(text):
+        return CJK_FONT, None
+    return ("helv" if size < 20 else "hebo"), None
+
+
+def _solid_rgb(shape) -> tuple[float, float, float] | None:
+    try:
+        from pptx.enum.dml import MSO_FILL
+
+        if shape.fill.type != MSO_FILL.SOLID:
+            return None
+        rgb = shape.fill.fore_color.rgb
+        return (rgb[0] / 255, rgb[1] / 255, rgb[2] / 255)
+    except Exception:  # noqa: BLE001 - theme colours raise; preview just skips them
+        return None
+
+
+def _text_rgb(shape) -> tuple[float, float, float]:
+    for paragraph in getattr(shape, "text_frame", None).paragraphs if has_text_frame(shape) else []:
+        for run in paragraph.runs:
+            try:
+                rgb = run.font.color.rgb
+            except Exception:  # noqa: BLE001
+                continue
+            if rgb is not None:
+                return (rgb[0] / 255, rgb[1] / 255, rgb[2] / 255)
+    return INK
+
+
+def _draw_text(page, rect: fitz.Rect, text: str, size: float, color, align: int = 0) -> float:
+    if not text.strip():
+        return 1.0
+    fontname, _ = _font_for(text, size)
+    try:
+        return page.insert_textbox(rect, text, fontsize=size, fontname=fontname, color=color, align=align, lineheight=1.22)
+    except Exception:  # noqa: BLE001 - a bad font name must not kill the render
+        return page.insert_textbox(rect, text, fontsize=size, color=color, align=align)
+
+
+def _draw_shape(page, shape, slide_no: int, notes: list[dict[str, Any]]) -> None:
+    geometry = box(shape)
+    if geometry is None:
+        return
+    left, top, width, height = geometry
+    rect = fitz.Rect(left * 72, top * 72, (left + width) * 72, (top + height) * 72)
+    if rect.is_empty:
+        return
+
+    if is_picture(shape):
+        try:
+            page.insert_image(rect, stream=shape.image.blob, keep_proportion=False)
+        except Exception as exc:  # noqa: BLE001
+            notes.append({"slide": slide_no, "shape": shape.name, "issue": f"picture not drawn: {exc}"})
+        return
+
+    if is_table(shape):
+        _draw_table(page, shape, rect, slide_no, notes)
+        return
+
+    if is_chart(shape):
+        _draw_chart(page, shape, rect, slide_no, notes)
+        return
+
+    fill = _solid_rgb(shape)
+    if fill is not None:
+        page.draw_rect(rect, color=RULE, fill=fill, width=0.6)
+    elif getattr(shape, "shape_type", None) is not None and shape.shape_type is not None:
+        page.draw_rect(rect, color=RULE, width=0.4, dashes="[2 2] 0")
+
+    if not has_text_frame(shape):
+        return
+
+    frame = shape.text_frame
+    inset = fitz.Rect(
+        rect.x0 + (frame.margin_left or 0) / EMU_PER_INCH * 72,
+        rect.y0 + (frame.margin_top or 0) / EMU_PER_INCH * 72,
+        rect.x1 - (frame.margin_right or 0) / EMU_PER_INCH * 72,
+        rect.y1 - (frame.margin_bottom or 0) / EMU_PER_INCH * 72,
+    )
+    text = text_of(shape)
+    if not text.strip():
+        return
+    size = run_size_pt(shape) or default_size_pt(shape)
+    color = _text_rgb(shape)
+    kind = (placeholder_kind(shape) or "").upper()
+    align = 1 if kind in {"CENTER_TITLE", "TITLE"} else 0
+    # Title-like placeholders inherit middle anchoring from the master; the
+    # preview would otherwise pin every title to the top of its box.
+    if kind in {"CENTER_TITLE", "TITLE", "SUBTITLE"}:
+        needed = estimated_text_height_in(shape) or 0.0
+        slack = inset.height - needed * 72
+        if slack > 4:
+            inset = fitz.Rect(inset.x0, inset.y0 + slack / 2, inset.x1, inset.y1)
+    spare = _draw_text(page, inset, text, size, color, align)
+    if spare < 0:
+        # PyMuPDF draws nothing at all when the text cannot fit. Grow the frame
+        # downward and retry so the preview shows the overflow instead of an
+        # empty box, and still report it.
+        grown = fitz.Rect(inset.x0, inset.y0, inset.x1, inset.y1 + abs(spare) + 24)
+        _draw_text(page, grown, text, size, color, align)
+        notes.append(
+            {
+                "slide": slide_no,
+                "shape": shape.name,
+                "issue": "text overflows its frame",
+                "fontPt": round(size, 1),
+                "overflowPt": round(-spare, 1),
+            }
+        )
+
+
+def _draw_table(page, shape, rect: fitz.Rect, slide_no: int, notes: list[dict[str, Any]]) -> None:
+    table = shape.table
+    columns = len(table.columns)
+    rows = len(table.rows)
+    if not columns or not rows:
+        return
+    widths = [column.width or 0 for column in table.columns]
+    total_w = sum(widths) or 1
+    heights = [row.height or int(0.4 * EMU_PER_INCH) for row in table.rows]
+    total_h = sum(heights) or 1
+    x = rect.x0
+    y = rect.y0
+    for row_index in range(rows):
+        row_h = rect.height * (heights[row_index] / total_h)
+        x = rect.x0
+        for column_index in range(columns):
+            column_w = rect.width * (widths[column_index] / total_w)
+            cell_rect = fitz.Rect(x, y, x + column_w, y + row_h)
+            page.draw_rect(cell_rect, color=RULE, width=0.5, fill=PANEL if row_index == 0 else None)
+            cell = table.cell(row_index, column_index)
+            cell_text = (cell.text or "").strip()
+            if cell_text:
+                size = 10.0 if row_index else 10.5
+                inner = fitz.Rect(cell_rect.x0 + 4, cell_rect.y0 + 2, cell_rect.x1 - 4, cell_rect.y1 - 2)
+                spare = _draw_text(page, inner, cell_text, size, INK)
+                if spare < 0:
+                    notes.append({"slide": slide_no, "shape": f"{shape.name}[{row_index},{column_index}]", "issue": "table cell clips"})
+            x += column_w
+        y += row_h
+
+
+def _draw_chart(page, shape, rect: fitz.Rect, slide_no: int, notes: list[dict[str, Any]]) -> None:
+    chart = shape.chart
+    page.draw_rect(rect, color=RULE, fill=(1, 1, 1), width=0.8)
+    title = ""
+    try:
+        if chart.has_title and chart.chart_title.text_frame.text.strip():
+            title = chart.chart_title.text_frame.text.strip()
+    except Exception:  # noqa: BLE001
+        title = ""
+    plot = fitz.Rect(rect.x0 + 18, rect.y0 + 12, rect.x1 - 14, rect.y1 - 26)
+    if title:
+        _draw_text(page, fitz.Rect(plot.x0, rect.y0 + 4, plot.x1, rect.y0 + 26), title, 11, INK)
+        plot = fitz.Rect(plot.x0, rect.y0 + 30, plot.x1, rect.y1 - 26)
+
+    try:
+        categories = [str(category) for category in chart.plots[0].categories]
+    except Exception:  # noqa: BLE001
+        categories = []
+    series: list[tuple[str, list[float]]] = []
+    try:
+        for plot_series in chart.series:
+            values = [float(value) if value is not None else 0.0 for value in plot_series.values]
+            series.append((str(plot_series.name), values))
+    except Exception:  # noqa: BLE001 - a broken chart cache must not kill the render
+        notes.append({"slide": slide_no, "shape": shape.name, "issue": "chart values unreadable"})
+
+    if not series or not categories:
+        _draw_text(page, plot, "chart data unavailable", 11, MUTED)
+        return
+
+    kind = str(getattr(chart, "chart_type", ""))
+    max_value = max([abs(value) for _, values in series for value in values] + [1.0])
+    baseline = plot.y1 - 16
+
+    if "PIE" in kind.upper() or "DOUGHNUT" in kind.upper():
+        values = series[0][1]
+        total = sum(abs(value) for value in values) or 1.0
+        center = fitz.Point((plot.x0 + plot.x1) / 2, (plot.y0 + baseline) / 2)
+        radius = min(plot.width, baseline - plot.y0) * 0.34
+        start = 0.0
+        shades = [0.25, 0.40, 0.55, 0.70, 0.85, 0.95]
+        for index, value in enumerate(values):
+            sweep = 360.0 * abs(value) / total
+            shade = shades[index % len(shades)]
+            page.draw_pie(center, radius, fitz.Point(radius * 1.0, 0), start, start + sweep, color=RULE, fill=(shade, shade, shade * 0.95 + 0.05), width=0.6)
+            start += sweep
+        return
+
+    if "LINE" in kind.upper():
+        step = plot.width / max(len(categories) - 1, 1)
+        for series_index, (_, values) in enumerate(series):
+            points = []
+            for index, value in enumerate(values):
+                x = plot.x0 + step * index
+                y = baseline - (abs(value) / max_value) * (baseline - plot.y0 - 6)
+                points.append(fitz.Point(x, y))
+            if len(points) > 1:
+                shade = 0.20 + 0.18 * series_index
+                page.draw_polyline(points, color=(shade, shade, shade + 0.1), width=1.6)
+            for point in points:
+                page.draw_circle(point, 2.0, color=ACCENT, fill=ACCENT)
+        _draw_categories(page, plot, baseline, categories)
+        return
+
+    group_width = plot.width / max(len(categories), 1)
+    bar_width = group_width * 0.6 / max(len(series), 1)
+    for index, category in enumerate(categories):
+        for series_index, (_, values) in enumerate(series):
+            if index >= len(values):
+                continue
+            value = abs(values[index])
+            bar_height = (value / max_value) * (baseline - plot.y0 - 8)
+            x0 = plot.x0 + group_width * index + group_width * 0.2 + bar_width * series_index
+            bar = fitz.Rect(x0, baseline - bar_height, x0 + bar_width * 0.92, baseline)
+            shade = 0.30 + 0.22 * series_index
+            page.draw_rect(bar, color=ACCENT, fill=(shade * 0.55, shade * 0.75, shade), width=0.5)
+            _draw_text(page, fitz.Rect(bar.x0 - 6, bar.y0 - 12, bar.x1 + 6, bar.y0), _short(values[index]), 7.5, MUTED, 1)
+    page.draw_line(fitz.Point(plot.x0, baseline), fitz.Point(plot.x1, baseline), color=RULE, width=0.8)
+    _draw_categories(page, plot, baseline, categories)
+
+
+def _short(value: float) -> str:
+    return f"{value:g}"
+
+
+def _draw_categories(page, plot: fitz.Rect, baseline: float, categories: list[str]) -> None:
+    group_width = plot.width / max(max(len(categories), 1), 1)
+    for index, category in enumerate(categories):
+        cell = fitz.Rect(
+            plot.x0 + group_width * index,
+            baseline + 2,
+            plot.x0 + group_width * (index + 1) - 2,
+            baseline + 18,
+        )
+        _draw_text(page, cell, category[:18], 7.5, MUTED, 1)
+
+
+def render_preview(pptx_path: str | Path, out_dir: str | Path, width_px: int = 1600) -> dict[str, Any]:
+    """Rasterise every slide to `<out_dir>/slide-NN.png`; returns a JSON-able report."""
+    presentation = open_presentation(pptx_path)
+    destination = Path(out_dir)
+    destination.mkdir(parents=True, exist_ok=True)
+    slide_w_in, slide_h_in = slide_size_in(presentation)
+    page_w = slide_w_in * 72
+    page_h = slide_h_in * 72
+    zoom = max(1.0, width_px / page_w)
+
+    files: list[str] = []
+    notes: list[dict[str, Any]] = []
+    document = fitz.open()
+    try:
+        for slide_no, slide in enumerate(presentation.slides, start=1):
+            page = document.new_page(width=page_w, height=page_h)
+            page.draw_rect(fitz.Rect(0, 0, page_w, page_h), color=None, fill=(1, 1, 1))
+            try:
+                background = slide.background.fill
+                rgb = background.fore_color.rgb
+                page.draw_rect(
+                    fitz.Rect(0, 0, page_w, page_h),
+                    color=None,
+                    fill=(rgb[0] / 255, rgb[1] / 255, rgb[2] / 255),
+                )
+            except Exception:  # noqa: BLE001 - most decks inherit a white master
+                pass
+            for shape in iter_shapes(slide.shapes):
+                _draw_shape(page, shape, slide_no, notes)
+            page.insert_text(
+                fitz.Point(page_w - 54, page_h - 10),
+                f"{slide_no}",
+                fontsize=7,
+                fontname="helv",
+                color=MUTED,
+            )
+            png = destination / f"slide-{slide_no:02d}.png"
+            page.get_pixmap(matrix=fitz.Matrix(zoom, zoom)).save(str(png))
+            files.append(str(png.resolve()))
+    finally:
+        document.close()
+
+    report = {
+        "ok": True,
+        "engine": "draft-preview",
+        "fidelity": "geometry-only",
+        "note": "Shapes redrawn from OOXML; fonts and theme colours are approximations. Use it to judge placement, collisions and overflow, not typography.",
+        "slides": len(files),
+        "slideSize": [round(slide_w_in, 3), round(slide_h_in, 3)],
+        "files": files,
+        "overflow": notes,
+    }
+    (destination / "preview.json").write_text(json.dumps(report, ensure_ascii=False, indent=2), encoding="utf-8")
+    return report
+
+
+def main(argv: list[str] | None = None) -> int:
+    import argparse
+
+    parser = argparse.ArgumentParser(description="Rasterise a .pptx to PNGs using a draft geometry renderer (no Office required).")
+    parser.add_argument("presentation", help="path to the .pptx")
+    parser.add_argument("--out-dir", required=True, help="directory for slide-NN.png and preview.json")
+    parser.add_argument("--width", type=int, default=1600, help="output pixel width (default 1600)")
+    args = parser.parse_args(argv)
+
+    report = render_preview(args.presentation, args.out_dir, args.width)
+    json.dump(report, sys.stdout, ensure_ascii=False, indent=2)
+    sys.stdout.write("\n")
+    return 0 if report["slides"] else 3
+
+
+if __name__ == "__main__":
+    raise SystemExit(main())

package/assets/skills/office-presentations/container_tools/render_slides.py

@@ -0,0 +1,146 @@
+"""Render a deck to PNGs, one per slide, using whatever engine is available.
+
+Purpose: the verification loop needs pixels. Ask the plugin CLI first: it drives
+Microsoft Office, WPS Office, or LibreOffice, whichever is installed. When none
+is, fall back to the draft geometry renderer in `preview_render.py` and say so
+loudly in the JSON — a draft preview proves placement, not typography.
+
+Inputs may be a .pptx or a .pdf. Output is `<out-dir>/slide-NN.png` for every
+page, plus `<out-dir>/render.json`.
+
+Exit codes:
+  0  at least one page rendered
+  2  input missing or unreadable
+  3  no engine produced output
+"""
+
+from __future__ import annotations
+
+import json
+import shutil
+import subprocess
+import sys
+import tempfile
+from pathlib import Path
+
+# Windows consoles default to a legacy code page; every payload is UTF-8 JSON.
+for _stream in (sys.stdout, sys.stderr):
+    try:
+        _stream.reconfigure(encoding="utf-8", errors="replace")
+    except (AttributeError, ValueError):
+        pass
+
+sys.path.insert(0, str(Path(__file__).resolve().parent))
+
+
+def office_cli() -> Path:
+    """The plugin's helper CLI, which owns renderer discovery."""
+    return Path(__file__).resolve().parents[4] / "python" / "office_cli.py"
+
+
+def rasterise_pdf(pdf_path: Path, out_dir: Path, dpi: int) -> list[str]:
+    try:
+        import pymupdf as fitz
+    except ImportError:  # pragma: no cover
+        import fitz
+
+    files: list[str] = []
+    document = fitz.open(str(pdf_path))
+    try:
+        zoom = dpi / 72.0
+        for index, page in enumerate(document, start=1):
+            png = out_dir / f"slide-{index:02d}.png"
+            page.get_pixmap(matrix=fitz.Matrix(zoom, zoom)).save(str(png))
+            files.append(str(png))
+    finally:
+        document.close()
+    return files
+
+
+def render_with_office(source: Path, out_dir: Path, dpi: int) -> dict:
+    """Export through the CLI, which tries every installed office suite."""
+    scratch = Path(tempfile.mkdtemp(prefix="dsh-render-"))
+    try:
+        completed = subprocess.run(
+            [sys.executable, str(office_cli()), "pptx", "render",
+             "--input", str(source), "--out-dir", str(scratch), "--dpi", str(dpi)],
+            capture_output=True, text=True, timeout=300, check=False,
+        )
+        if completed.returncode != 0:
+            raise RuntimeError((completed.stderr or completed.stdout or "render failed").strip()[:400])
+        report = json.loads(completed.stdout)
+        files: list[str] = []
+        for index, page in enumerate(report.get("pages") or [], start=1):
+            target = out_dir / f"slide-{index:02d}.png"
+            shutil.move(page, target)
+            files.append(str(target))
+        return {
+            "ok": bool(files),
+            "engine": report.get("renderer") or "office",
+            "fidelity": "faithful",
+            "slides": len(files),
+            "dpi": dpi,
+            "files": files,
+        }
+    finally:
+        shutil.rmtree(scratch, ignore_errors=True)
+
+
+def render(input_path: str | Path, out_dir: str | Path, dpi: int = 110, engine: str = "auto", width: int = 1600) -> dict:
+    source = Path(input_path)
+    destination = Path(out_dir)
+    if not source.is_file():
+        return {"ok": False, "error": {"code": "input_missing", "message": f"no such file: {source}"}}
+    destination.mkdir(parents=True, exist_ok=True)
+
+    if source.suffix.lower() == ".pdf":
+        files = rasterise_pdf(source, destination, dpi)
+        return {"ok": bool(files), "engine": "pdf-raster", "fidelity": "faithful", "slides": len(files), "dpi": dpi, "files": files}
+
+    notes: list[str] = []
+    if engine in {"auto", "office"}:
+        try:
+            return render_with_office(source, destination, dpi)
+        except Exception as exc:  # noqa: BLE001 - fall through to the draft engine
+            notes.append(f"office render failed: {exc}")
+
+    if engine == "office":
+        return {"ok": False, "engine": "office", "notes": notes, "error": {"code": "no_engine", "message": "no Microsoft Office, WPS Office, or LibreOffice renderer is available"}}
+
+    from preview_render import render_preview
+
+    report = render_preview(source, destination, width)
+    report["notes"] = notes + [
+        "Draft preview: geometry is exact, type is approximate. Treat colours and fonts as indicative only.",
+    ]
+    report["files"] = [str(Path(item).resolve()) for item in report["files"]]
+    return report
+
+
+def main(argv: list[str] | None = None) -> int:
+    import argparse
+
+    parser = argparse.ArgumentParser(description="Render a .pptx or .pdf into one PNG per page.")
+    parser.add_argument("--input", required=True, help="deck or PDF to render")
+    parser.add_argument("--out-dir", required=True, help="directory for slide-NN.png and render.json")
+    parser.add_argument("--dpi", type=int, default=110, help="raster dpi when a real renderer is used (default 110)")
+    parser.add_argument("--width", type=int, default=1600, help="pixel width for the draft renderer (default 1600)")
+    parser.add_argument("--engine", choices=("auto", "office", "draft"), default="auto", help="force an engine (default auto)")
+    parser.add_argument("--json-out", help="also write the report here")
+    args = parser.parse_args(argv)
+
+    report = render(args.input, args.out_dir, args.dpi, args.engine, args.width)
+    out_dir = Path(args.out_dir)
+    out_dir.mkdir(parents=True, exist_ok=True)
+    (out_dir / "render.json").write_text(json.dumps(report, ensure_ascii=False, indent=2), encoding="utf-8")
+    if args.json_out:
+        Path(args.json_out).write_text(json.dumps(report, ensure_ascii=False, indent=2), encoding="utf-8")
+    json.dump(report, sys.stdout, ensure_ascii=False, indent=2)
+    sys.stdout.write("\n")
+    if report.get("ok"):
+        return 0
+    return 2 if report.get("error", {}).get("code") == "input_missing" else 3
+
+
+if __name__ == "__main__":
+    raise SystemExit(main())