pdfmd-cli 3.11.1__tar.gz → 3.15.0__tar.gz

{pdfmd_cli-3.11.1/pdfmd_cli.egg-info → pdfmd_cli-3.15.0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: pdfmd-cli
-Version: 3.11.1
+Version: 3.15.0
 Summary: One command from Markdown to a good-looking PDF: a Pandoc wrapper with smart defaults, project-file discovery and a multi-engine fallback chain
 Author: Ali Perdekhan
 License-Expression: MIT
@@ -20,6 +20,8 @@ Dynamic: license-file
 
 # pdfmd
 
+[![Test](https://github.com/aliperdehan/pdfmd/actions/workflows/publish.yml/badge.svg)](https://github.com/aliperdehan/pdfmd/actions/workflows/publish.yml) [![PyPI](https://img.shields.io/pypi/v/pdfmd-cli)](https://pypi.org/project/pdfmd-cli/)
+
 **One command from Markdown to a good-looking PDF.** `pdfmd` wraps
 [Pandoc](https://pandoc.org) and fills in everything you would otherwise
 have to remember: sensible fonts and margins, the right Markdown dialect,
@@ -92,6 +94,16 @@ brew install --cask mactex-no-gui           # optional: full LaTeX (large)
 sudo apt install pipx pandoc texlive-xetex  # Debian/Ubuntu
 ```
 
+```powershell
+py -m pip install --user pipx; py -m pipx ensurepath          # Windows
+winget install --id JohnMacFarlane.Pandoc; winget install --id Typst.Typst
+```
+
+Every push is tested on Windows, macOS and Linux (Pandoc + Typst: single
+files, CSV tables, a book, HTML output); see the
+[Test and publish workflow](https://github.com/aliperdehan/pdfmd/actions/workflows/publish.yml).
+LaTeX engines aren't part of that automated test on Windows yet.
+
 Optional extras: [Quarto](https://quarto.org) for `.qmd` files,
 `pandoc-crossref` for `@fig:`/`@tbl:` references, and LibreOffice for
 Office files.

{pdfmd_cli-3.11.1 → pdfmd_cli-3.15.0}/README.md

@@ -1,5 +1,7 @@
 # pdfmd
 
+[![Test](https://github.com/aliperdehan/pdfmd/actions/workflows/publish.yml/badge.svg)](https://github.com/aliperdehan/pdfmd/actions/workflows/publish.yml) [![PyPI](https://img.shields.io/pypi/v/pdfmd-cli)](https://pypi.org/project/pdfmd-cli/)
+
 **One command from Markdown to a good-looking PDF.** `pdfmd` wraps
 [Pandoc](https://pandoc.org) and fills in everything you would otherwise
 have to remember: sensible fonts and margins, the right Markdown dialect,
@@ -72,6 +74,16 @@ brew install --cask mactex-no-gui           # optional: full LaTeX (large)
 sudo apt install pipx pandoc texlive-xetex  # Debian/Ubuntu
 ```
 
+```powershell
+py -m pip install --user pipx; py -m pipx ensurepath          # Windows
+winget install --id JohnMacFarlane.Pandoc; winget install --id Typst.Typst
+```
+
+Every push is tested on Windows, macOS and Linux (Pandoc + Typst: single
+files, CSV tables, a book, HTML output); see the
+[Test and publish workflow](https://github.com/aliperdehan/pdfmd/actions/workflows/publish.yml).
+LaTeX engines aren't part of that automated test on Windows yet.
+
 Optional extras: [Quarto](https://quarto.org) for `.qmd` files,
 `pandoc-crossref` for `@fig:`/`@tbl:` references, and LibreOffice for
 Office files.

{pdfmd_cli-3.11.1 → pdfmd_cli-3.15.0}/pdfmd.py

@@ -449,7 +449,7 @@ Automatic source backups (--backup, v3.8.0; formats v3.9.0):
 # unreliable 1.x history from those gaps, versioning restarts at 2.0.0 here
 # (2026-09-16, the author's call) as an honest baseline: this is where real
 # changelog tracking begins, not a claim about how many changes preceded it.
-PDFMD_VERSION = "3.11.1"
+PDFMD_VERSION = "3.15.0"
 import argparse
 import filecmp
 from fnmatch import fnmatchcase
@@ -1301,7 +1301,8 @@ def resolve_soffice() -> str | None:
     fine for the user every day at a terminal prompt. Checked, in order:
     `soffice` on PATH, `libreoffice` on PATH (the usual Linux package
     name -- often a symlink to the same binary), then the standard macOS
-    app-bundle path directly.
+    app-bundle path directly, or on Windows the default install folders
+    under %ProgramFiles% / %ProgramFiles(x86)%.
     """
     for name in ("soffice", "libreoffice"):
         found = which(name)
@@ -1311,6 +1312,12 @@ def resolve_soffice() -> str | None:
         mac_path = Path("/Applications/LibreOffice.app/Contents/MacOS/soffice")
         if mac_path.exists():
             return str(mac_path)
+    if sys.platform == "win32":
+        # The Windows installer doesn't add LibreOffice to PATH.
+        for variable in ("ProgramFiles", "ProgramFiles(x86)"):
+            base = os.environ.get(variable)
+            if base and (Path(base) / "LibreOffice" / "program" / "soffice.exe").exists():
+                return str(Path(base) / "LibreOffice" / "program" / "soffice.exe")
     return None
 
 
@@ -1578,14 +1585,39 @@ def wrap_latex_header_includes(text: str) -> str:
 
 
 @contextmanager
-def prepared_latex_inputs(paths: list[Path], latex_engine: bool) -> Iterator[list[Path]]:
-    """Yield temporary, corrected inputs for LaTeX-family renders when needed."""
+def prepared_latex_inputs(paths: list[Path], latex_engine: bool,
+                          typst_engine: bool = False, doc_count: int = 1) -> Iterator[list[Path]]:
+    """Yield temporary, corrected inputs for LaTeX-family renders when
+    needed -- or, with `typst_engine` instead, to make a margin setting
+    written for the OTHER engine family reach Typst's own template (see
+    fix_typst_margin()). The two are mutually exclusive; at most one
+    corrector ever applies to a given render, same as `latex_engine` alone
+    before this.
+
+    `paths[:doc_count]` are the main document(s) being rendered -- a
+    single `title_source` everywhere except report/book mode, where every
+    chapter file counts (``doc_count=len(files)``, passed explicitly by
+    those two call sites; every other call site keeps the default of 1).
+    Every path after that is a linked --metadata-file, which Pandoc
+    accepts EITHER fenced (``---``/``...``-delimited, same as a
+    document's own front matter) OR completely bare (no delimiters at
+    all -- see the Pandoc manual's own "metadata file" section).
+    fix_typst_margin() is told which is which (`is_metadata`) so it can
+    safely treat a bare metadata file as pure YAML, while a bare MAIN
+    DOCUMENT (no delimiters at all) is definitely markdown body text,
+    never YAML, and is always left alone regardless.
+    """
     temporary_paths: list[Path] = []
     prepared: list[Path] = []
     try:
-        for path in paths:
+        for index, path in enumerate(paths):
             text = path.read_text(encoding="utf-8-sig")
-            corrected = wrap_latex_header_includes(text) if latex_engine else text
+            if latex_engine:
+                corrected = wrap_latex_header_includes(text)
+            elif typst_engine:
+                corrected = fix_typst_margin(text, is_metadata=(index >= doc_count))
+            else:
+                corrected = text
             if corrected == text:
                 prepared.append(path)
                 continue
@@ -1942,6 +1974,189 @@ def has_geometry(md_path: Path, variables: list[str], metadata_files: list[Path]
                for preamble in preamble_files)
 
 
+# Shared by frontmatter_margin_geometry_options() and fix_typst_margin()
+# below to translate a margin setting written for ONE engine family into
+# the variable the OTHER one actually reads -- geometry: (LaTeX) and
+# margin: (Typst) each mean nothing at all to the other's template, so a
+# document written/tested against one engine silently loses its margin
+# the moment it's rendered through the other (has_geometry() already
+# stops pdfmd's own DEFAULT_MARGIN from stepping on either, which is
+# correct, but does nothing to make the document's OWN setting reach the
+# engine that can't read it natively). Both directions normalize through
+# one canonical shape: a {top, bottom, left, right} dict of raw dimension
+# strings, physical sides only -- Typst's binding-aware `inside`/`outside`
+# and `rest` have no LaTeX equivalent at all, so a `margin:` mapping using
+# any of those three is left completely alone rather than guessed at.
+MARGIN_SIDE_ALIASES = {
+    "x": ("left", "right"),
+    "y": ("top", "bottom"),
+    "top": ("top",),
+    "bottom": ("bottom",),
+    "left": ("left",),
+    "right": ("right",),
+}
+
+# geometry package option names -- includes the older t/b/l/r-margin
+# spellings alongside the more common ones actually seen in practice.
+GEOMETRY_SIDE_ALIASES = {
+    "margin": ("top", "bottom", "left", "right"),
+    "hmargin": ("left", "right"),
+    "vmargin": ("top", "bottom"),
+    "top": ("top",), "tmargin": ("top",),
+    "bottom": ("bottom",), "bmargin": ("bottom",),
+    "left": ("left",), "lmargin": ("left",),
+    "right": ("right",), "rmargin": ("right",),
+}
+
+
+def parse_margin_sides(value) -> dict[str, str] | None:
+    """Normalize a parsed (via PyYAML) Typst-shaped ``margin:`` value into
+    a ``{top, bottom, left, right}`` dict of raw dimension strings -- or
+    None when `value` is a scalar-shaped None/absent, or a mapping that
+    uses ``inside``/``outside``/``rest`` (Typst's own binding-aware
+    margins, with no LaTeX equivalent) or any other key this doesn't
+    recognize, or isn't a scalar or mapping at all.
+
+    A scalar (already the case fix_typst_margin() itself exists to
+    correct on the Typst side) is uniform on all four sides.
+    """
+    if isinstance(value, (str, int, float)) and not isinstance(value, bool):
+        side_value = str(value)
+        return {"top": side_value, "bottom": side_value, "left": side_value, "right": side_value}
+    if not isinstance(value, dict):
+        return None
+    sides: dict[str, str] = {}
+    for key, side_value in value.items():
+        aliases = MARGIN_SIDE_ALIASES.get(str(key).strip().casefold())
+        if not aliases:
+            return None
+        for side in aliases:
+            sides[side] = str(side_value)
+    return sides or None
+
+
+def parse_geometry_sides(value) -> dict[str, str] | None:
+    """Normalize a parsed (via PyYAML) ``geometry:`` value -- a bare
+    "key=val" scalar, a comma-joined "key=val,key=val" scalar, or a YAML
+    list of such strings, the three shapes Pandoc's own LaTeX template
+    accepts -- into a ``{top, bottom, left, right}`` dict of raw dimension
+    strings, or None when it contains anything this can't safely
+    interpret: an option with no ``key=`` at all, or a key that isn't
+    about a margin (``includehead``, ``showframe``, ...) -- the geometry
+    package has many options that aren't page margins, and guessing wrong
+    on those would be worse than leaving Typst's own defaults in place.
+    """
+    if isinstance(value, str):
+        items = [value]
+    elif isinstance(value, list) and all(isinstance(item, str) for item in value):
+        items = value
+    else:
+        return None
+    sides: dict[str, str] = {}
+    for item in items:
+        for part in item.split(","):
+            part = part.strip()
+            if not part:
+                continue
+            if "=" not in part:
+                return None
+            key, _, side_value = part.partition("=")
+            aliases = GEOMETRY_SIDE_ALIASES.get(key.strip().casefold())
+            if not aliases:
+                return None
+            for side in aliases:
+                sides[side] = side_value.strip()
+    return sides or None
+
+
+def sides_to_geometry_options(sides: dict[str, str]) -> list[str]:
+    """Format a {top,bottom,left,right} dict as the geometry: option list
+    Pandoc's LaTeX template expects -- collapsed to one ``margin=...``
+    when all four sides given are equal, one option per side otherwise
+    (only the sides actually present -- a mapping that only set ``x:``
+    leaves top/bottom to LaTeX's own default, same as it would for Typst).
+    """
+    if len(sides) == 4 and len(set(sides.values())) == 1:
+        return [f"margin={next(iter(sides.values()))}"]
+    return [f"{side}={sides[side]}" for side in ("top", "bottom", "left", "right") if side in sides]
+
+
+def frontmatter_margin_geometry_options(md_path: Path, variables: list[str],
+                                        metadata_files: list[Path] = ()) -> list[str] | None:
+    """Return the geometry: option list a document's own (or a linked
+    --metadata-file's) ``margin:`` value translates to for LaTeX-family
+    engines -- or None when there's nothing to translate.
+
+    ``margin:`` is a real Pandoc variable for the TYPST template only (see
+    fix_typst_margin() below) -- Pandoc's LaTeX template never reads it at
+    all, only ``geometry:`` does. has_geometry() already treats a bare
+    ``margin:`` (in the document's own front matter OR a metadata file --
+    it checks both) as "a margin setting exists" (correctly, so
+    DEFAULT_MARGIN isn't injected on top of it), but nothing translated
+    that value into the one variable LaTeX's own template actually
+    consumes -- so a document with ONLY ``margin: 2.54cm`` (or a
+    ``top:``/``bottom:``/``left:``/``right:``/``x:``/``y:`` breakdown) and
+    no ``geometry:`` silently kept LaTeX's own much wider article-class
+    default margins on every LaTeX-family engine, the requested value
+    never taking effect at all -- confirmed directly (2026-09-28) alongside
+    the Typst-side bug the scalar case triggers (see fix_typst_margin()).
+
+    Same precedence pdf-engine resolution uses (see
+    frontmatter_pdfmd_options()'s callers): the document's own front
+    matter is checked first; if it sets no ``margin:`` at all, each
+    linked --metadata-file is checked next, in order, so a shared
+    metadata.yaml can set ``margin:`` once for every document that finds
+    it -- leaving it out entirely, the way this function's own first
+    version did, meant a document with margin ONLY in a shared metadata
+    file lost it silently on both engine families at once, worse than the
+    bug this function exists to fix. A real ``geometry:`` anywhere -- the
+    document's own front matter, any metadata file, or -V -- always wins,
+    untouched, checked before any ``margin:`` in any source, same as
+    has_geometry().
+
+    Returns None when: no ``margin:`` in any source; a real ``geometry:``
+    is set anywhere; PyYAML isn't installed (falls back to the
+    scalar-only regex this function used before parse_margin_sides()
+    existed, since a real parse is needed for the mapping case -- checked
+    per source, same order); or every ``margin:`` found uses Typst's
+    ``inside``/``outside``/``rest`` keys, which have no LaTeX equivalent
+    (see parse_margin_sides()).
+    """
+    if any(variable.startswith(("geometry=", "geometry:")) for variable in variables):
+        return None
+    text = md_path.read_text(encoding="utf-8-sig")
+    front_matter = re.match(r"^---\s*\n(.*?)\n---\s*(?:\n|$)", text, re.DOTALL)
+    blocks = [front_matter.group(1)] if front_matter else []
+    for metadata_file in metadata_files:
+        meta_text = metadata_file.read_text(encoding="utf-8-sig")
+        # A --metadata-file may be fenced (---...---, same as front matter)
+        # or completely bare -- Pandoc accepts either. yaml.safe_load()
+        # chokes on the fenced shape as-is (a second "---" reads as a
+        # second YAML document, "expected a single document in the
+        # stream"), so unwrap it the same way md_path's own front matter
+        # is, and only fall back to the raw text when it isn't fenced.
+        meta_front_matter = re.match(r"^---\s*\n(.*?)\n---\s*(?:\n|$)", meta_text, re.DOTALL)
+        blocks.append(meta_front_matter.group(1) if meta_front_matter else meta_text)
+    if any(re.search(r"^geometry\s*:", block, re.MULTILINE) for block in blocks):
+        return None
+    for block in blocks:
+        if yaml is None:
+            match = re.search(r"^margin\s*:[ \t]*(\S.*?)[ \t]*$", block, re.MULTILINE)
+            if not match:
+                continue
+            value = match.group(1).strip().strip("'\"")
+            return [f"margin={value}"] if value else None
+        try:
+            data = yaml.safe_load(block)
+        except yaml.YAMLError:
+            continue
+        if not isinstance(data, dict) or "margin" not in data:
+            continue
+        sides = parse_margin_sides(data["margin"])
+        return sides_to_geometry_options(sides) if sides else None
+    return None
+
+
 def pagesize_typo_value(md_path: Path, variables: list[str]) -> str | None:
     """Return a document's own ``pagesize:`` value, IF ``papersize:`` isn't
     already set anywhere Pandoc would see it -- a -V variable or the
@@ -2024,6 +2239,112 @@ def typst_papersize_translation(value: str | None) -> str | None:
     return TYPST_PAPERSIZE_ALIASES.get(value.strip().casefold())
 
 
+TYPST_MARGIN_RE = re.compile(r"^(?P<indent>[ \t]*)margin(?P<colon>[ \t]*:[ \t]*)(?P<val>\S.*?)[ \t]*$",
+                             re.MULTILINE)
+
+
+def sides_to_typst_margin_yaml(sides: dict[str, str], indent: str, newline: str) -> str:
+    """Format a {top,bottom,left,right} dict as the YAML ``margin:``
+    mapping Pandoc's Typst template needs -- collapsed to ``x:``/``y:``
+    when all four sides given are equal, one line per side otherwise
+    (only the sides actually present, in a fixed top/bottom/left/right
+    order for a stable, readable rewrite)."""
+    if len(sides) == 4 and len(set(sides.values())) == 1:
+        value = next(iter(sides.values()))
+        return f"{indent}margin:{newline}{indent}  x: {value}{newline}{indent}  y: {value}"
+    lines = [f"{indent}margin:"]
+    lines += [f"{indent}  {side}: {sides[side]}" for side in ("top", "bottom", "left", "right")
+             if side in sides]
+    return newline.join(lines)
+
+
+def fix_typst_margin(text: str, is_metadata: bool = False) -> str:
+    """Make a document's or linked --metadata-file's own page-margin
+    setting reach Pandoc's Typst template, whichever engine family it was
+    actually written for.
+
+    Two distinct problems, both from the same root cause -- ``margin:``
+    is a real Pandoc variable for the TYPST template only, and
+    ``geometry:`` for the LATEX template only, neither reads the other's
+    key at all:
+
+    1. A bare scalar ``margin:`` (``margin: 2.54cm``) is rewritten into
+       the ``x:``/``y:`` mapping Typst's template requires -- it always
+       renders margin as ``($for(margin/pairs)$...$endfor$)``, iterating
+       key/value PAIRS a scalar has none of, so it comes out as the
+       literal, invalid ``margin: (: ,)``: a Typst syntax error
+       ("unexpected comma") instead of a page margin. Confirmed directly
+       (2026-09-28): a document with ``margin: 2.54cm`` and
+       ``pdfmd-options: {engine: typst}`` failed outright with exactly
+       that error, while rendering fine (albeit at the WRONG margin --
+       see frontmatter_margin_geometry_options()) on every LaTeX-family
+       engine.
+
+    2. A document with ONLY a ``geometry:`` (no ``margin:`` at all) --
+       written for and tested against a LaTeX-family engine -- reaches
+       Typst with no margin variable set at all: Typst's template never
+       reads ``geometry:``, so it silently falls back to its own built-in
+       default (1.25in) instead of erroring, the same "wrong margin, no
+       warning" failure frontmatter_margin_geometry_options() fixes in the
+       other direction. Parsed via parse_geometry_sides() and added as a
+       new ``margin:`` mapping -- ``geometry:`` itself is left in place;
+       it's inert for Typst, not harmful.
+
+    Neither fix touches a ``margin:`` that's already a YAML mapping --
+    Pandoc's ``/pairs`` filter already turns that straight into valid
+    Typst key/value pairs, and it may use Typst's own ``inside``/
+    ``outside``/``rest`` keys, which parse_margin_sides() can't safely
+    reinterpret but Typst's own template needs no help with anyway.
+
+    `is_metadata` marks a linked --metadata-file (passed by
+    prepared_latex_inputs() using its own `doc_count`), which Pandoc
+    accepts EITHER fenced (front-matter-shaped, handled the same as a
+    document below) OR completely bare -- no ``---``/``...`` delimiters
+    at all, just the YAML directly. A bare MAIN document (`is_metadata`
+    False) is always left alone instead: with no front matter at all,
+    that's markdown BODY text, never YAML, and must never be parsed as
+    such.
+    """
+    front_matter = re.match(r"^---\s*\n(.*?)\n---\s*(?:\n|$)", text, re.DOTALL)
+    if front_matter:
+        block = front_matter.group(1)
+        prefix = text[:front_matter.start(1)]
+        suffix = text[front_matter.end(1):]
+    elif is_metadata:
+        block = text
+        prefix = ""
+        suffix = ""
+    else:
+        return text
+    newline = "\r\n" if "\r\n" in block else "\n"
+
+    if yaml is not None:
+        try:
+            data = yaml.safe_load(block)
+        except yaml.YAMLError:
+            data = None
+        if isinstance(data, dict) and "margin" not in data and "geometry" in data:
+            sides = parse_geometry_sides(data["geometry"])
+            if not sides:
+                return text
+            insertion = sides_to_typst_margin_yaml(sides, "", newline) + newline
+            return prefix + insertion + block + suffix
+        if isinstance(data, dict) and isinstance(data.get("margin"), dict):
+            return text  # already a mapping -- Typst's template handles it as-is
+
+    def replace(match: re.Match) -> str:
+        value = match.group("val").strip().strip("'\"")
+        if not value or value.startswith(("{", "[")):
+            return match.group(0)
+        indent = match.group("indent")
+        return f"{indent}margin:{newline}{indent}  x: {value}{newline}{indent}  y: {value}"
+
+    fixed_block = TYPST_MARGIN_RE.sub(replace, block, count=1)
+    if fixed_block == block:
+        return text
+    return prefix + fixed_block + suffix
+
+
 def has_monofont(md_path: Path, variables: list[str]) -> bool:
     if any(variable.startswith("monofont=") for variable in variables):
         return True
@@ -2475,6 +2796,51 @@ def frontmatter_extra_preambles(md_path: Path) -> list[Path]:
     return resolved
 
 
+def frontmatter_extra_lua_filters(md_path: Path) -> list[Path]:
+    """Read ``pdfmd-options: lua-filter: ...`` -- a document naming its own
+    Pandoc Lua filter(s) explicitly, instead of relying on
+    find_lua_filters()'s fixed auto-discovery names (``<doc-stem>.lua`` or
+    ``nulabreport.lua``).
+
+    Accepts a bare string or a YAML list of strings; each path is resolved
+    relative to `md_path`'s own directory (not cwd -- a document should be
+    runnable from any directory), same convention as
+    frontmatter_extra_preambles. A path that doesn't exist is a hard
+    SystemExit, not a silent skip: naming a filter explicitly means it's
+    required, unlike auto-discovery's own best-effort search.
+
+    Merged with (not a replacement for) whatever find_lua_filters() already
+    auto-discovered -- see this function's one call site in convert_one,
+    which appends these AFTER the auto-discovered ones and dedupes by
+    resolved path. Both land on the command line after --citeproc (see
+    run()'s own comment on that ordering) -- required for a filter, like
+    fullcite.lua, that consumes citeproc's resolved bibliography div.
+
+    Reuses the existing "lua" --no-auto KIND, same as
+    frontmatter_extra_preambles reuses "preamble": a pdfmd-options.lua-filter
+    key is written IN the document, so --no-auto lua (or a bare --no-auto)
+    suppresses it the same as auto-discovery.
+
+    Added 2026-09-28 for a real case: an annotated-bibliography.md using a
+    citeproc-dependent filter (fullcite.lua) that isn't named after the
+    document's own stem, so find_lua_filters() alone never picked it up --
+    the document had to name it explicitly instead.
+    """
+    value = frontmatter_pdfmd_options(md_path).get("lua-filter")
+    if value is None:
+        return []
+    names = [value] if isinstance(value, str) else (
+        [str(item) for item in value] if isinstance(value, list) else [])
+    resolved = []
+    for name in names:
+        path = (md_path.parent / name).resolve()
+        if not path.is_file():
+            raise SystemExit(f"{md_path}: pdfmd-options.lua-filter names {name!r}, "
+                             f"which doesn't exist at {path}")
+        resolved.append(path)
+    return resolved
+
+
 # --stamp / pdfmd-options.stamp: appends (or updates) a "BUILD NOTES" HTML
 # comment at/near the end of a document recording what compiled it and when
 # -- a lab-notebook-style provenance note, invisible in the rendered output
@@ -3446,6 +3812,12 @@ def default_output_path(path: Path, target_format: str) -> Path:
 
 def open_file(path: Path) -> None:
     """Open a finished file with the platform's default viewer, for --open."""
+    if sys.platform == "win32":
+        try:
+            os.startfile(path)  # Windows' own "open with default app"
+        except OSError as error:
+            print(f"WARN  --open: could not open {path}: {error}", file=sys.stderr)
+        return
     opener = "open" if sys.platform == "darwin" else "xdg-open" if sys.platform.startswith("linux") else None
     if opener is None or not which(opener):
         print(f"WARN  --open: no '{opener or 'file opener'}' found for this platform; "
@@ -3697,7 +4069,7 @@ def run_tex_engine(engine: str, tex_path: Path, output_dir: Path,
         # points bibtex back at the source directory to find the .bib
         # file, since it has no --input-directory flag of its own.
         bib_env = os.environ.copy()
-        bib_env["BIBINPUTS"] = f"{tex_path.parent}:{bib_env.get('BIBINPUTS', '')}"
+        bib_env["BIBINPUTS"] = f"{tex_path.parent}{os.pathsep}{bib_env.get('BIBINPUTS', '')}"
         bib_cmd = (["biber", f"--input-directory={tex_path.parent}", tex_path.stem] if bib_tool == "biber"
                    else ["bibtex", tex_path.stem])
         bib_result = run(bib_cmd, output_dir, env=bib_env)
@@ -3872,11 +4244,15 @@ def convert_via_soffice_bridge(md_path: Path, output: Path, effective_from: str
         cmd += resource_path_option(md_path.parent, pandoc_cwd, metadata_files)
         for variable in variables:
             cmd += ["-V", variable]
-        cmd += pandoc_options
         cmd += crossref_filter_args(md_path, pandoc_options, no_auto, str(md_path))
         cmd += csv_table_filter_args(md_path, no_auto, csv_filter)
         if contains_citations(md_path) and "--citeproc" not in pandoc_options and not CITEPROC_DISABLED:
             cmd.append("--citeproc")
+        # pandoc_options after --citeproc: any --lua-filter/--filter a caller
+        # passes through needs resolved citations already in the AST, same
+        # invariant as the auto-discovered lua_filters below (see run()'s
+        # matching comment in convert_one).
+        cmd += pandoc_options
         for lua_filter in lua_filters:
             cmd += ["--lua-filter", str(lua_filter)]
         log_cmd(cmd, pandoc_cwd, verbose)
@@ -4102,6 +4478,10 @@ def _convert_one(md_path: Path, out_dir: Path | None, presentation: bool, font:
     if reader_reason:
         note("READER", reader_reason)
     lua_filters = [] if auto_disabled(no_auto, "lua") else find_lua_filters(md_path, metadata_files)
+    if not auto_disabled(no_auto, "lua"):
+        for extra_filter in frontmatter_extra_lua_filters(md_path):
+            if extra_filter not in lua_filters:
+                lua_filters.append(extra_filter)
     for lua_filter in lua_filters:
         note("LUA", str(lua_filter))
 
@@ -4153,6 +4533,11 @@ def _convert_one(md_path: Path, out_dir: Path | None, presentation: bool, font:
             if is_tex_target and geometry_needed:
                 note("MARGIN", f"{md_path}: no geometry/margin set; "
                               f"using geometry:margin={DEFAULT_MARGIN} on LaTeX-family targets")
+            margin_options = (None if auto_disabled(no_auto, "margin")
+                              else frontmatter_margin_geometry_options(md_path, variables, metadata_files))
+            if is_tex_target and margin_options:
+                note("MARGIN", f"{md_path}: margin: isn't a Pandoc variable LaTeX-family targets "
+                              f"read (geometry: is) -- using geometry:{','.join(margin_options)}")
             if is_tex_target and monofont_needed:
                 note("MONOFONT", f"{md_path}: has code but no monofont set; "
                                 f"using {default_monofont()} on LaTeX-family targets")
@@ -4163,7 +4548,8 @@ def _convert_one(md_path: Path, out_dir: Path | None, presentation: bool, font:
                                   f"(papersize: is) -- using papersize={pagesize_typo} on LaTeX-family "
                                   "targets. Set papersize: yourself, or --no-auto papersize, to silence "
                                   "this and keep the Letter default")
-            with prepared_latex_inputs([title_source, *metadata_files], is_tex_target) as prepared, \
+            with prepared_latex_inputs([title_source, *metadata_files], is_tex_target,
+                                       typst_engine=(target_format == "typst")) as prepared, \
                     document_header_file(md_path, (bool(preamble_files) or bool(pdf_meta_snippet_text))
                                          and is_tex_target) as header_file, \
                     pdf_metadata_header_file(pdf_meta_snippet_text if is_tex_target else None) as pdf_meta_file:
@@ -4199,6 +4585,9 @@ def _convert_one(md_path: Path, out_dir: Path | None, presentation: bool, font:
                         cmd += ["-V", f"mainfont={first_font}"]
                     if geometry_needed:
                         cmd += ["-V", f"geometry:margin={DEFAULT_MARGIN}"]
+                    elif margin_options:
+                        for margin_option in margin_options:
+                            cmd += ["-V", f"geometry:{margin_option}"]
                     if monofont_needed:
                         cmd += ["-V", f"monofont={default_monofont()}"]
                     if pagesize_typo:
@@ -4217,7 +4606,6 @@ def _convert_one(md_path: Path, out_dir: Path | None, presentation: bool, font:
                         cmd += ["-V", f"papersize={typst_size}"]
                 if shift_heading:
                     cmd += ["--shift-heading-level-by=-1"]
-                cmd += pandoc_options
                 if preamble_files and is_tex_target:
                     for preamble_file in preamble_files:
                         cmd += ["--include-in-header", str(preamble_file)]
@@ -4243,6 +4631,11 @@ def _convert_one(md_path: Path, out_dir: Path | None, presentation: bool, font:
                         cmd.append(f"--{citation_engine}")
                     else:
                         cmd.append("--citeproc")
+                # pandoc_options after the citation-engine flag above: a
+                # caller-supplied --lua-filter/--filter needs resolved
+                # citations already in the AST (same invariant as the
+                # auto-discovered lua_filters below).
+                cmd += pandoc_options
                 if is_tex_target and tablewidth_auto:
                     cmd += ["--lua-filter", str(width_filter)]
                 for lua_filter in lua_filters:
@@ -4263,6 +4656,11 @@ def _convert_one(md_path: Path, out_dir: Path | None, presentation: bool, font:
         if geometry_needed:
             note("MARGIN", f"{md_path}: no geometry/margin set; "
                           f"using geometry:margin={DEFAULT_MARGIN} on LaTeX-family engines")
+        margin_options = (None if auto_disabled(no_auto, "margin")
+                         else frontmatter_margin_geometry_options(md_path, variables, metadata_files))
+        if margin_options:
+            note("MARGIN", f"{md_path}: margin: isn't a Pandoc variable LaTeX-family engines "
+                          f"read (geometry: is) -- using geometry:{','.join(margin_options)}")
         monofont_needed = (not auto_disabled(no_auto, "monofont")
                           and has_code_spans(md_path.read_text(encoding="utf-8-sig"))
                           and not has_monofont(md_path, variables))
@@ -4330,7 +4728,8 @@ def _convert_one(md_path: Path, out_dir: Path | None, presentation: bool, font:
                 remaining = [e for e in engines[engine_index + 1:] if ENGINE_FAMILY.get(e, e) not in failed_families]
                 report_engine_failure(str(md_path), engine, result, remaining, debug)
                 continue
-            with prepared_latex_inputs([title_source, *metadata_files], engine in LATEX_ENGINES) as prepared, \
+            with prepared_latex_inputs([title_source, *metadata_files], engine in LATEX_ENGINES,
+                                       typst_engine=(engine == "typst")) as prepared, \
                     document_header_file(md_path, (bool(preamble_files) or bool(pdf_meta_snippet_text))
                                          and engine in LATEX_ENGINES) as header_file, \
                     pdf_metadata_header_file(pdf_meta_snippet_text if engine in LATEX_ENGINES else None) as pdf_meta_file:
@@ -4354,6 +4753,9 @@ def _convert_one(md_path: Path, out_dir: Path | None, presentation: bool, font:
                         cmd += ["-V", f"mainfontfallback={fallback_font()}"]
                     if geometry_needed and engine in LATEX_ENGINES:
                         cmd += ["-V", f"geometry:margin={DEFAULT_MARGIN}"]
+                    elif margin_options and engine in LATEX_ENGINES:
+                        for margin_option in margin_options:
+                            cmd += ["-V", f"geometry:{margin_option}"]
                     if monofont_needed and engine in LATEX_ENGINES:
                         cmd += ["-V", f"monofont={default_monofont()}"]
                     if pagesize_typo and engine in LATEX_ENGINES:
@@ -4372,7 +4774,6 @@ def _convert_one(md_path: Path, out_dir: Path | None, presentation: bool, font:
                             cmd += ["-V", f"papersize={typst_size}"]
                     if shift_heading:
                         cmd += ["--shift-heading-level-by=-1"]
-                    cmd += pandoc_options
                     if preamble_files and engine in LATEX_ENGINES:
                         for preamble_file in preamble_files:
                             cmd += ["--include-in-header", str(preamble_file)]
@@ -4389,6 +4790,15 @@ def _convert_one(md_path: Path, out_dir: Path | None, presentation: bool, font:
                     cmd += crossref_filter_args(md_path, pandoc_options, no_auto, str(md_path))
                     if contains_citations(md_path) and "--citeproc" not in pandoc_options and not CITEPROC_DISABLED:
                         cmd.append("--citeproc")
+                    # pandoc_options after --citeproc, same reason as below: a
+                    # caller-supplied --lua-filter/--filter needs resolved
+                    # citations already in the AST. (This used to sit before
+                    # the --citeproc append above, which silently broke any
+                    # citeproc-dependent filter passed via extra CLI args --
+                    # e.g. `--lua-filter=some.lua` landed ahead of --citeproc
+                    # on the actual pandoc command line. Confirmed via
+                    # --verbose CMD output before this fix.)
+                    cmd += pandoc_options
                     # Lua filters go last: pandoc applies --citeproc and filters in
                     # command-line order, and a filter that renders cell contents to
                     # LaTeX needs the citations already resolved. csv-table has to
@@ -4618,7 +5028,7 @@ def build_parser() -> argparse.ArgumentParser:
                              "with --engine when a document's rendered output doesn't match what the "
                              "Markdown predicts and the short reason alone isn't enough to tell why")
     parser.add_argument("--open", action="store_true",
-                        help="open the finished file (via macOS 'open'/Linux 'xdg-open') once "
+                        help="open the finished file (via macOS 'open', Linux 'xdg-open' or Windows' default app) once "
                              "conversion succeeds. Applies to single-file and report/book mode, "
                              "each of which produce exactly one output; ignored in batch mode "
                              "(-b), which would otherwise pop open every file in the directory")
@@ -4935,13 +5345,22 @@ def main() -> None:
                 if report_geometry_needed:
                     report_note("MARGIN", "REPORT: no geometry/margin set; "
                                          f"using geometry:margin={DEFAULT_MARGIN} on LaTeX-family targets")
+                report_margin_options = (None if auto_disabled(report_no_auto, "margin")
+                                         else frontmatter_margin_geometry_options(files[0], variables,
+                                                                                  metadata_files))
+                if is_tex_target and report_margin_options:
+                    report_note("MARGIN", "REPORT: margin: isn't a Pandoc variable LaTeX-family "
+                                         "targets read (geometry: is) -- using "
+                                         f"geometry:{','.join(report_margin_options)}")
                 report_monofont_needed = (is_tex_target and not auto_disabled(report_no_auto, "monofont")
                                           and any(has_code_spans(file.read_text(encoding="utf-8-sig")) for file in files)
                                           and not has_monofont(files[0], variables))
                 if report_monofont_needed:
                     report_note("MONOFONT", "REPORT: has code but no monofont set; "
                                            f"using {default_monofont()} on LaTeX-family targets")
-                with prepared_latex_inputs([*files, *metadata_files], is_tex_target) as prepared, \
+                with prepared_latex_inputs([*files, *metadata_files], is_tex_target,
+                                           typst_engine=(target_format == "typst"),
+                                           doc_count=len(files)) as prepared, \
                         document_header_file(files[0], (bool(report_preambles) or bool(report_pdf_meta_snippet_text))
                                              and is_tex_target) as report_header_file, \
                         pdf_metadata_header_file(report_pdf_meta_snippet_text) as report_pdf_meta_file:
@@ -4969,11 +5388,13 @@ def main() -> None:
                             cmd += ["-V", f"mainfont={report_first_font}"]
                         if report_geometry_needed:
                             cmd += ["-V", f"geometry:margin={DEFAULT_MARGIN}"]
+                        elif report_margin_options:
+                            for margin_option in report_margin_options:
+                                cmd += ["-V", f"geometry:{margin_option}"]
                         if report_monofont_needed:
                             cmd += ["-V", f"monofont={default_monofont()}"]
                     for variable in variables:
                         cmd += ["-V", variable]
-                    cmd += pandoc_options
                     if report_preambles and is_tex_target:
                         for preamble in report_preambles:
                             cmd += ["--include-in-header", str(preamble)]
@@ -4991,6 +5412,9 @@ def main() -> None:
                             and "--citeproc" not in pandoc_options
                             and not CITEPROC_DISABLED):
                         cmd.append("--citeproc")
+                    # pandoc_options after --citeproc: see convert_one's
+                    # matching comment.
+                    cmd += pandoc_options
                     if is_tex_target and not auto_disabled(report_no_auto, "tablewidth"):
                         cmd += ["--lua-filter", str(width_filter)]
                     log_cmd(cmd, pandoc_cwd, args.verbose)
@@ -5014,6 +5438,12 @@ def main() -> None:
                 if geometry_needed:
                     report_note("MARGIN", "REPORT: no geometry/margin set; "
                                          f"using geometry:margin={DEFAULT_MARGIN} on LaTeX-family engines")
+                margin_options = (None if auto_disabled(report_no_auto, "margin")
+                                 else frontmatter_margin_geometry_options(files[0], variables, metadata_files))
+                if margin_options:
+                    report_note("MARGIN", "REPORT: margin: isn't a Pandoc variable LaTeX-family "
+                                         "engines read (geometry: is) -- using "
+                                         f"geometry:{','.join(margin_options)}")
                 monofont_needed = (not auto_disabled(report_no_auto, "monofont")
                                   and any(has_code_spans(file.read_text(encoding="utf-8-sig")) for file in files)
                                   and not has_monofont(files[0], variables))
@@ -5044,7 +5474,9 @@ def main() -> None:
                         print(f"SKIP  REPORT: {engine} shares the {family} engine with an earlier "
                               f"failure; skipping", file=sys.stderr)
                         continue
-                    with prepared_latex_inputs([*files, *metadata_files], engine in LATEX_ENGINES) as prepared, \
+                    with prepared_latex_inputs([*files, *metadata_files], engine in LATEX_ENGINES,
+                                               typst_engine=(engine == "typst"),
+                                               doc_count=len(files)) as prepared, \
                             pdf_metadata_header_file(
                                 report_pdf_meta_snippet_text if engine in LATEX_ENGINES else None
                             ) as report_pdf_meta_file:
@@ -5059,11 +5491,13 @@ def main() -> None:
                         cmd += report_resource_path
                         if geometry_needed and engine in LATEX_ENGINES:
                             cmd += ["-V", f"geometry:margin={DEFAULT_MARGIN}"]
+                        elif margin_options and engine in LATEX_ENGINES:
+                            for margin_option in margin_options:
+                                cmd += ["-V", f"geometry:{margin_option}"]
                         if monofont_needed and engine in LATEX_ENGINES:
                             cmd += ["-V", f"monofont={default_monofont()}"]
                         for variable in variables:
                             cmd += ["-V", variable]
-                        cmd += pandoc_options
                         if preambles and engine in LATEX_ENGINES:
                             for preamble in preambles:
                                 cmd += ["--include-in-header", str(preamble)]
@@ -5075,6 +5509,9 @@ def main() -> None:
                                 and "--citeproc" not in pandoc_options
                                 and not CITEPROC_DISABLED):
                             cmd.append("--citeproc")
+                        # pandoc_options after --citeproc: see convert_one's
+                        # matching comment.
+                        cmd += pandoc_options
                         if report_tablewidth_auto and engine in LATEX_ENGINES:
                             cmd += ["--lua-filter", str(width_filter)]
                         log_cmd(cmd, pandoc_cwd, args.verbose)

{pdfmd_cli-3.11.1 → pdfmd_cli-3.15.0/pdfmd_cli.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: pdfmd-cli
-Version: 3.11.1
+Version: 3.15.0
 Summary: One command from Markdown to a good-looking PDF: a Pandoc wrapper with smart defaults, project-file discovery and a multi-engine fallback chain
 Author: Ali Perdekhan
 License-Expression: MIT
@@ -20,6 +20,8 @@ Dynamic: license-file
 
 # pdfmd
 
+[![Test](https://github.com/aliperdehan/pdfmd/actions/workflows/publish.yml/badge.svg)](https://github.com/aliperdehan/pdfmd/actions/workflows/publish.yml) [![PyPI](https://img.shields.io/pypi/v/pdfmd-cli)](https://pypi.org/project/pdfmd-cli/)
+
 **One command from Markdown to a good-looking PDF.** `pdfmd` wraps
 [Pandoc](https://pandoc.org) and fills in everything you would otherwise
 have to remember: sensible fonts and margins, the right Markdown dialect,
@@ -92,6 +94,16 @@ brew install --cask mactex-no-gui           # optional: full LaTeX (large)
 sudo apt install pipx pandoc texlive-xetex  # Debian/Ubuntu
 ```
 
+```powershell
+py -m pip install --user pipx; py -m pipx ensurepath          # Windows
+winget install --id JohnMacFarlane.Pandoc; winget install --id Typst.Typst
+```
+
+Every push is tested on Windows, macOS and Linux (Pandoc + Typst: single
+files, CSV tables, a book, HTML output); see the
+[Test and publish workflow](https://github.com/aliperdehan/pdfmd/actions/workflows/publish.yml).
+LaTeX engines aren't part of that automated test on Windows yet.
+
 Optional extras: [Quarto](https://quarto.org) for `.qmd` files,
 `pandoc-crossref` for `@fig:`/`@tbl:` references, and LibreOffice for
 Office files.