dsh-office-kit 0.3.0 → 0.3.1

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

@@ -13,6 +13,13 @@ import json
 import sys
 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
+
 try:
     from PIL import Image, ImageDraw, ImageFont
 except ImportError:  # pragma: no cover - whitelist dependency, but fail clearly

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

@@ -42,6 +42,13 @@ from ooxml_helpers import (  # noqa: E402
     text_of,
 )
 
+# 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
+
 TITLE_OUTSIDE_MARGIN_IN = 0.0
 DEFAULT_TITLE_FLOOR_PT = 28.0
 DEFAULT_BODY_FLOOR_PT = 14.0

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

@@ -16,6 +16,13 @@ import zipfile
 from pathlib import Path
 from xml.etree import ElementTree
 
+# 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
+
 REL_NS = "{http://schemas.openxmlformats.org/package/2006/relationships}"
 CT_NS = "{http://schemas.openxmlformats.org/package/2006/content-types}"
 P_NS = "{http://schemas.openxmlformats.org/presentationml/2006/main}"

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

@@ -20,6 +20,13 @@ import zipfile
 from pathlib import Path
 from xml.etree import ElementTree
 
+# 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))
 
 C_NS = "{http://schemas.openxmlformats.org/drawingml/2006/chart}"

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

@@ -22,6 +22,13 @@ from decimal import Decimal, InvalidOperation
 from pathlib import Path
 from xml.etree import ElementTree
 
+# 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))
 
 C_NS = "{http://schemas.openxmlformats.org/drawingml/2006/chart}"

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

@@ -24,6 +24,13 @@ import sys
 from decimal import Decimal, InvalidOperation
 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))
 
 TOTAL_PATTERN = re.compile(r"\btotal\b", re.IGNORECASE)

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

@@ -23,6 +23,13 @@ import sys
 from pathlib import Path
 from typing import Any, Iterable, Iterator
 
+# 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
+
 EMU_PER_INCH = 914400
 
 # Anything that renders about a full em wide and wraps mid-token.

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

@@ -48,6 +48,13 @@ from ooxml_helpers import (  # noqa: E402
     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)

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

@@ -23,6 +23,13 @@ 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))
 
 

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

@@ -26,6 +26,13 @@ import json
 import sys
 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
+
 
 def _emit(payload: dict, code: int = 0) -> int:
     print(json.dumps(payload, ensure_ascii=False, indent=2))

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

@@ -30,6 +30,13 @@ sys.path.insert(0, str(Path(__file__).resolve().parent))
 
 from ooxml_helpers import box, is_picture, slide_size_in  # noqa: E402
 
+# 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
+
 EMU_PER_INCH = 914400
 DEFAULT_TOLERANCE_EMU = 9144  # 0.01 in
 

package/assets/skills/office-presentations/template_following_scripts/check_template_fidelity.py

@@ -20,6 +20,13 @@ sys.path.insert(0, str(Path(__file__).resolve().parent.parent / "container_tools
 
 from ooxml_helpers import box, has_visible_text, is_picture, overlap_ratio, text_of  # noqa: E402
 
+# 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
+
 PLACEHOLDER_PATTERNS = [
     re.compile(pattern, re.IGNORECASE)
     for pattern in (

package/assets/skills/office-presentations/template_following_scripts/inspect_template_deck.py

@@ -23,6 +23,13 @@ sys.path.insert(0, str(Path(__file__).resolve().parent.parent / "container_tools
 
 from ooxml_helpers import box, has_visible_text, is_chart, is_picture, is_table, placeholder_kind, slide_size_in, text_of  # noqa: E402
 
+# 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
+
 
 def _theme(presentation) -> dict:
     fonts: set[str] = set()

package/assets/skills/office-presentations/template_following_scripts/make_contact_sheet.py

@@ -15,6 +15,13 @@ import math
 import sys
 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.parent / "container_tools"))
 
 

package/assets/skills/office-presentations/template_following_scripts/prepare_template_starter_deck.py

@@ -23,6 +23,13 @@ sys.path.insert(0, str(Path(__file__).resolve().parent.parent / "container_tools
 
 from ooxml_helpers import box, placeholder_kind, slide_size_in  # noqa: E402
 
+# 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
+
 PLACEHOLDER_ALIASES = {
     "title": ("TITLE", "CENTER_TITLE"),
     "subtitle": ("SUBTITLE",),

package/assets/skills/office-presentations/template_following_scripts/validate_template_plan.py

@@ -26,6 +26,13 @@ import json
 import sys
 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.parent / "container_tools"))
 
 ALLOWED_ACTIONS = {"rewrite", "rewrite-and-reposition", "replace", "delete", "keep", "fill-placeholder"}

package/assets/skills/office-spreadsheets/container_tools/add_chart.py

@@ -35,6 +35,13 @@ import json
 import sys
 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
+
 if hasattr(sys.stdout, "reconfigure"):
     sys.stdout.reconfigure(errors="replace")
 

package/assets/skills/office-spreadsheets/container_tools/add_image.py

@@ -23,6 +23,13 @@ import json
 import sys
 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
+
 if hasattr(sys.stdout, "reconfigure"):
     sys.stdout.reconfigure(errors="replace")
 

package/assets/skills/office-spreadsheets/container_tools/append_rows.py

@@ -29,6 +29,13 @@ import json
 import sys
 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
+
 if hasattr(sys.stdout, "reconfigure"):
     sys.stdout.reconfigure(errors="replace")
 

package/assets/skills/office-spreadsheets/container_tools/diff_workbooks.py

@@ -21,6 +21,13 @@ import json
 import sys
 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
+
 if hasattr(sys.stdout, "reconfigure"):
     # Legacy console codepages (cp936, cp1252) must not abort a report.
     sys.stdout.reconfigure(errors="replace")

package/assets/skills/office-spreadsheets/container_tools/inspect_workbook.py

@@ -22,6 +22,13 @@ import sys
 import zipfile
 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
+
 if hasattr(sys.stdout, "reconfigure"):
     # Legacy console codepages (cp936, cp1252) must not abort a report.
     sys.stdout.reconfigure(errors="replace")

package/assets/skills/office-spreadsheets/container_tools/recalc_gate.py

@@ -28,6 +28,13 @@ import subprocess
 import sys
 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
+
 if hasattr(sys.stdout, "reconfigure"):
     # Legacy console codepages (cp936, cp1252) must not abort a report.
     sys.stdout.reconfigure(errors="replace")
@@ -116,23 +123,28 @@ def _referenced_sheets(formula: str) -> list[str]:
     return found
 
 
-def _recalculate(python: str, cli: Path, source: Path, out: Path) -> tuple[bool, str | None]:
+def _recalculate(python: str, cli: Path, source: Path, out: Path) -> tuple[bool, str | None, str | None]:
     command = [python, str(cli), "xlsx", "recalc", "--input", str(source), "--out", str(out)]
     try:
         completed = subprocess.run(command, capture_output=True, text=True, timeout=300)
     except FileNotFoundError as error:
-        return False, f"cannot start the helper CLI: {error}"
+        return False, f"cannot start the helper CLI: {error}", None
     except subprocess.TimeoutExpired:
-        return False, "recalculation timed out after 300 seconds"
+        return False, "recalculation timed out after 300 seconds", None
     if completed.returncode == 0 and out.is_file():
-        return True, None
+        renderer = None
+        try:
+            renderer = json.loads(completed.stdout).get("renderer")
+        except (json.JSONDecodeError, AttributeError):
+            renderer = None
+        return True, None, renderer
     message = (completed.stderr or completed.stdout or "").strip()
     try:
         parsed = json.loads(message)
         message = str(parsed.get("error", message))
     except json.JSONDecodeError:
         pass
-    return False, message or f"helper CLI exited with code {completed.returncode}"
+    return False, message or f"helper CLI exited with code {completed.returncode}", None
 
 
 def _ignored(entries: str | None) -> dict[str, set[str]]:
@@ -176,10 +188,10 @@ def main() -> int:
         if not cli.is_file():
             recalc_error = f"helper CLI not found: {cli}; pass --cli or use --no-recalc"
         else:
-            recalculated, recalc_error = _recalculate(args.python, cli, source, out)
+            recalculated, recalc_error, renderer = _recalculate(args.python, cli, source, out)
             if recalculated:
                 checked = out
-                engine = "microsoft excel"
+                engine = renderer or "recalculation engine"
 
     try:
         cached = _cached_values(checked, cells) if cells else {}

package/assets/skills/office-spreadsheets/container_tools/set_cells.py

@@ -26,6 +26,13 @@ import json
 import sys
 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
+
 if hasattr(sys.stdout, "reconfigure"):
     sys.stdout.reconfigure(errors="replace")
 

package/assets/skills/office-spreadsheets/container_tools/sheet_ops.py

@@ -24,6 +24,13 @@ import json
 import sys
 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
+
 if hasattr(sys.stdout, "reconfigure"):
     sys.stdout.reconfigure(errors="replace")
 

package/assets/skills/office-spreadsheets/container_tools/style_cells.py

@@ -30,6 +30,13 @@ import json
 import sys
 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
+
 if hasattr(sys.stdout, "reconfigure"):
     sys.stdout.reconfigure(errors="replace")
 

package/assets/skills/shared/check_runtime.py

@@ -24,6 +24,13 @@ import subprocess
 import sys
 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
+
 DEFAULT_ROOT = Path(__file__).resolve().parents[3]
 RENDERER_HINT = (
     "docx/pptx/xlsx rendering needs LibreOffice, Microsoft Office, or WPS Office; "

package/dist/index.js

@@ -324,20 +324,27 @@ function runPostinstall() {
 }
 let cache;
 let healing;
+/** A cached `ok` is only worth reusing while its interpreter still exists on disk. */
+function runtimeAlive(status) {
+	return status.state !== "ok" || status.python != null && isFile(status.python.path);
+}
 /**
 * Resolve a runtime that is proven to work, self-healing via postinstall.mjs when it
 * is missing or incomplete. Never returns a bare interpreter without dependencies.
+* Pass `force` for an explicit status check that must not trust the process cache.
 */
-function ensureRuntime(configFile = officePaths().configFile) {
-	if (cache != null && (cache.status.state === "ok" || Date.now() - cache.at < FAILURE_RETRY_DELAY)) return Promise.resolve(cache.status);
+function ensureRuntime(configFile = officePaths().configFile, force = false) {
+	if (!force && cache != null) {
+		if (cache.status.state === "ok" ? runtimeAlive(cache.status) : Date.now() - cache.at < FAILURE_RETRY_DELAY) return Promise.resolve(cache.status);
+	}
 	healing ??= (async () => {
-		const fresh = await inspectRuntime(configFile);
-		if (fresh.state === "ok") {
+		const probed = await inspectRuntime(configFile);
+		if (probed.state === "ok") {
 			cache = {
-				status: fresh,
+				status: probed,
 				at: Date.now()
 			};
-			return fresh;
+			return probed;
 		}
 		const healError = await runPostinstall();
 		const status = {
@@ -519,7 +526,7 @@ function apply(ctx) {
 		},
 		timeoutMs: 6e5,
 		async execute() {
-			const runtime = await ensureRuntime(paths.configFile);
+			const runtime = await ensureRuntime(paths.configFile, true);
 			const report = {
 				plugin: {
 					name,

package/package.json

@@ -1,7 +1,7 @@
 {
   "name": "dsh-office-kit",
   "type": "module",
-  "version": "0.3.0",
+  "version": "0.3.1",
   "description": "Office documents, PDF, presentations, and spreadsheets for DeepSeek Harness",
   "author": "Hairy",
   "license": "MIT",

package/python/office_cli.py

@@ -21,7 +21,18 @@ import tempfile
 from pathlib import Path
 from typing import Any, Iterable
 
-VERSION = "0.1.0"
+PACKAGE_ROOT = Path(__file__).resolve().parents[1]
+
+
+def _package_version() -> str:
+    try:
+        data = json.loads((PACKAGE_ROOT / "package.json").read_text(encoding="utf-8"))
+        return str(data["version"])
+    except (OSError, ValueError, KeyError, TypeError):
+        return "0.0.0"
+
+
+VERSION = _package_version()
 
 # Windows consoles default to a legacy code page; every payload is UTF-8 JSON.
 for _stream in (sys.stdout, sys.stderr):
@@ -339,6 +350,16 @@ def _docx_add_equation(paragraph, latex: str) -> bool:
         return False
 
 
+def _docx_mark_header_row(row) -> None:
+    """Repeat this row across pages and mark it as a header for assistive tech."""
+    from docx.oxml.ns import qn
+
+    properties = row._tr.get_or_add_trPr()
+    if properties.find(qn("w:tblHeader")) is None:
+        header = properties.makeelement(qn("w:tblHeader"), {})
+        properties.append(header)
+
+
 def _docx_add_block(document, block: dict[str, Any], counters: dict[str, int]) -> None:
     from docx.shared import Inches, Pt
 
@@ -391,6 +412,7 @@ def _docx_add_block(document, block: dict[str, Any], counters: dict[str, int]) -
                 for paragraph in cells[index].paragraphs:
                     for run in paragraph.runs:
                         run.bold = True
+            _docx_mark_header_row(table.rows[0])
         for row in rows:
             cells = table.add_row().cells
             for index, value in enumerate(row):
@@ -459,6 +481,10 @@ def docx_create(spec: dict[str, Any], out: Path) -> dict[str, Any]:
     if spec.get("author") or spec.get("title"):
         document.core_properties.title = str(spec.get("title", ""))
         document.core_properties.author = str(spec.get("author", ""))
+        # A fresh python-docx document carries the library name in comments; a
+        # delivered file should not advertise the tool that wrote it.
+        document.core_properties.comments = str(spec.get("comments", ""))
+        document.core_properties.last_modified_by = str(spec.get("author", ""))
     if spec.get("header"):
         document.sections[0].header.paragraphs[0].text = str(spec["header"])
     if spec.get("footer"):
@@ -659,17 +685,21 @@ def pdf_create(spec: dict[str, Any], out: Path) -> dict[str, Any]:
         author=str(spec.get("author", "")),
     )
     counters = {key: 0 for key in ("headings", "paragraphs", "list_items", "tables", "images")}
+    following: list[dict[str, Any]] = []
+    for page in spec.get("pages", []):
+        if isinstance(page, dict):
+            following.extend(page.get("blocks", []))
+            if page.get("break", True):
+                following.append({"type": "page_break"})
+    following.extend(spec.get("blocks", []))
+    kinds = {str(block.get("type", "")).lower() for block in following if isinstance(block, dict)}
     blocks: list[dict[str, Any]] = []
-    if spec.get("title"):
+    # `title` is metadata; only draw it when the blocks do not already carry one.
+    if spec.get("title") and "title" not in kinds:
         blocks.append({"type": "title", "text": spec["title"]})
     if spec.get("subtitle"):
         blocks.append({"type": "paragraph", "text": spec["subtitle"]})
-    for page in spec.get("pages", []):
-        if isinstance(page, dict):
-            blocks.extend(page.get("blocks", []))
-            if page.get("break", True):
-                blocks.append({"type": "page_break"})
-    blocks.extend(spec.get("blocks", []))
+    blocks.extend(following)
     document.build(_pdf_flowables(blocks, font, counters))
     return {"ok": True, "kind": "pdf", "command": "create", "output": str(out), "font": font, "counts": counters}
 
@@ -849,6 +879,10 @@ def pptx_create(spec: dict[str, Any], out: Path) -> dict[str, Any]:
     prs.slide_height = int(7.5 * 914400)
     prs.core_properties.title = str(spec.get("title", ""))
     prs.core_properties.author = str(spec.get("author", "DSH Office Suite"))
+    # python-pptx leaves its own name in the properties; a delivered deck should
+    # carry the author's, not the library's.
+    prs.core_properties.comments = str(spec.get("comments", ""))
+    prs.core_properties.last_modified_by = str(spec.get("author", "DSH Office Suite"))
     counters = {key: 0 for key in ("slides", "bullets", "tables", "charts", "images")}
     for slide_spec in spec.get("slides", []):
         if isinstance(slide_spec, dict):