excel-mcp-server 2.2.2__tar.gz → 2.2.4__tar.gz

{excel_mcp_server-2.2.2 → excel_mcp_server-2.2.4}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: excel-mcp-server
-Version: 2.2.2
+Version: 2.2.4
 Summary: MCP server for creating, reading and editing Excel workbooks without Microsoft Excel
 Keywords: mcp,mcp-server,model-context-protocol,excel,xlsx,xlsm,spreadsheet,openpyxl,charts,pivot-table,formulas,vba,llm,ai,claude,cursor
 Author: Haris Musa

{excel_mcp_server-2.2.2 → excel_mcp_server-2.2.4}/pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "excel-mcp-server"
-version = "2.2.2"
+version = "2.2.4"
 description = "MCP server for creating, reading and editing Excel workbooks without Microsoft Excel"
 readme = "README.md"
 license = "MIT"

{excel_mcp_server-2.2.2 → excel_mcp_server-2.2.4}/pyproject.toml.orig

@@ -1,6 +1,6 @@
 [project]
 name = "excel-mcp-server"
-version = "2.2.2"
+version = "2.2.4"
 description = "MCP server for creating, reading and editing Excel workbooks without Microsoft Excel"
 readme = "README.md"
 license = "MIT"

{excel_mcp_server-2.2.2 → excel_mcp_server-2.2.4}/src/excel_mcp/operations/slicer_place.py

@@ -7,6 +7,7 @@ from openpyxl.workbook import Workbook
 from openpyxl.workbook.defined_name import DefinedName
 from openpyxl.worksheet.worksheet import Worksheet
 
+from excel_mcp.errors import WorkbookError
 from excel_mcp.operations import slicer_xml as xml
 from excel_mcp.package import Link, Part, SheetPackage, state_of
 from excel_mcp.package.anchors import Geometry
@@ -119,7 +120,8 @@ def _list_in(extensions: dict[str, str], uri: str, element: str) -> None:
     assert reference is not None
     existing = extensions[uri]
     closing = re.search(r"</(?:\w+:)(?:slicerCaches|timelineCacheRefs)>", existing)
-    assert closing is not None
+    if closing is None:
+        raise WorkbookError("The workbook's list of slicer or timeline caches is malformed.")
     extensions[uri] = existing[: closing.start()] + reference[0] + existing[closing.start() :]
 
 

{excel_mcp_server-2.2.2 → excel_mcp_server-2.2.4}/src/excel_mcp/operations/sparklines.py

@@ -17,7 +17,7 @@ from excel_mcp.operations.sparkline_style import SparklineInfo, SparklineStyle,
 from excel_mcp.operations.sparkline_xml import build_group, read_groups
 from excel_mcp.operations.x14_xml import X14, XM, sheet_prefix
 from excel_mcp.package import state_of
-from excel_mcp.package.extensions import SPARKLINES, edit_sparklines
+from excel_mcp.package.extensions import SPARKLINES, edit_sparklines, sparkline_location
 from excel_mcp.refs import CellRange, cell_name, parse_range
 from excel_mcp.workspace import get_sheet
 
@@ -117,9 +117,7 @@ def _delete(extensions: dict[str, str], doomed: Callable[[str], bool]) -> int:
 
     def handle(item: str) -> list[str]:
         nonlocal removed
-        location = _LOCATION.search(item)
-        assert location is not None
-        if doomed(location[1]):
+        if doomed(sparkline_location(item)):
             removed += 1
             return []
         return [item]

{excel_mcp_server-2.2.2 → excel_mcp_server-2.2.4}/src/excel_mcp/package/consistency.py

@@ -27,7 +27,8 @@ _CACHE_LISTS = {
 }
 _SLICER_REFERENCE = re.compile(r'<(?:[\w.-]+:)?(?:slicer|timeline)\b[^>]*?\bcache="([^"]*)"')
 PIVOT_ENTRY = re.compile(r"<(?:[\w.-]+:)?pivotTable\b[^>]*/>")
-_CACHE_ENTRY = r'<(?:[\w.-]+:)?(?:slicerCache|timelineCacheRef)\b[^>]*?\b[\w.-]+:id="{id}"[^>]*/>'
+_CACHE_ENTRY = re.compile(r"<(?:[\w.-]+:)?(?:slicerCache|timelineCacheRef)\b[^>]*/>")
+_RELATIONSHIP_ID = re.compile(r'\b[\w.-]+:id="([^"]*)"')
 
 
 def cache_name(part: Part) -> str:
@@ -80,11 +81,11 @@ def without_caches(
         and link.target.content_type in _CACHES
         and cache_name(link.target) in dropped
     }
-    kept = []
-    for text in texts:
-        for identifier in gone:
-            text = re.sub(_CACHE_ENTRY.format(id=re.escape(identifier)), "", text)
-        kept.append(text)
+
+    def entry(match: re.Match[str]) -> str:
+        return "" if gone.intersection(_RELATIONSHIP_ID.findall(match[0])) else match[0]
+
+    kept = [_CACHE_ENTRY.sub(entry, text) for text in texts] if gone else texts
     return [link for link in links if link.id not in gone], kept
 
 

{excel_mcp_server-2.2.2 → excel_mcp_server-2.2.4}/src/excel_mcp/package/extensions.py

@@ -12,6 +12,7 @@ from xml.sax.saxutils import escape
 
 from openpyxl.utils.cell import column_index_from_string, get_column_letter
 
+from excel_mcp.errors import WorkbookError
 from excel_mcp.package.lines import LineEdit
 from excel_mcp.package.scan import unescape
 from excel_mcp.refs import CellRange, clip_areas, parse_range
@@ -41,21 +42,27 @@ class Rewriter(Protocol):
 
 
 _SPARKLINE = re.compile(r"<x14:sparkline>.*?</x14:sparkline>", re.S)
+# An element is its start tag, then either "/>" or its content up to its end tag. Written
+# like this, a self-closing element never makes the match run on to some later end tag, so
+# a document of many such elements is scanned once and not once per element.
 _EMPTY_GROUP = re.compile(
-    r"<x14:sparklineGroup\b[^>]*>(?:(?!<x14:sparkline>).)*?</x14:sparklineGroup>", re.S
+    r"<x14:sparklineGroup\b[^>]*?(?:/>|>(?:(?!<x14:sparkline>).)*?</x14:sparklineGroup>)", re.S
 )
-_RULE_GROUP = re.compile(r"<x14:conditionalFormatting\b[^>]*>.*?</x14:conditionalFormatting>", re.S)
-_VALIDATION = re.compile(r"<x14:dataValidation\b[^>]*>.*?</x14:dataValidation>", re.S)
+_RULE_GROUP = re.compile(
+    r"<x14:conditionalFormatting\b[^>]*?(?:/>|>.*?</x14:conditionalFormatting>)", re.S
+)
+_RULE = re.compile(r"<x14:cfRule\b[^>]*?(?:/>|>.*?</x14:cfRule>)", re.S)
+_VALIDATION = re.compile(r"<x14:dataValidation\b[^>]*?(?:/>|>.*?</x14:dataValidation>)", re.S)
 _DATED_GROUP = re.compile(
-    r"(<x14:sparklineGroup(?!s)[^>]*>(?:(?!<x14:sparklines>).)*?)<xm:f>([^<]*)</xm:f>(<x14:sparklines>)",
+    r"(<x14:sparklineGroup(?!s)[^>]*(?<!/)>(?:(?!<x14:sparklines>|</x14:sparklineGroup>).)*?)"
+    r"<xm:f>([^<]*)</xm:f>(<x14:sparklines>)",
     re.S,
 )
 _GROUP_DATES = re.compile(r"<xm:f>([^<]*)</xm:f>(?=<x14:sparklines>)")
 _RANGE = re.compile(r"<xm:sqref>([^<]*)</xm:sqref>")
 _FORMULA = re.compile(r"<xm:f>([^<]*)</xm:f>")
-_FORMULA_ITEM = re.compile(
-    r"<x14:cfRule\b.*?</x14:cfRule>|<x14:dataValidation\b.*?</x14:dataValidation>", re.S
-)
+_FORMULA_ITEM = re.compile(f"{_RULE.pattern}|{_VALIDATION.pattern}", re.S)
+_RULE_ID_ATTRIBUTE = re.compile(r'\bid="([^"]*)"')
 _GUID = re.compile(r'(<x14:id>|\b(?:xr2:uid|id)=")(\{[0-9A-Fa-f-]{36}\})')
 _CELL = re.compile(r"(\$?)\b([A-Z]{1,3})(\$?)(\d+)\b(?![!(\w])")
 
@@ -196,21 +203,18 @@ def rule_sub(xml: str, rule: Callable[[str], str]) -> str:
 
 
 def without_rules(extensions: dict[str, str], identifiers: set[str]) -> None:
-    """Remove the Excel 2010 conditional format rules with these ids (``{GUID}``)."""
+    """Remove the Excel 2010 conditional format rules with these ids (``{GUID}``), and the
+    groups left without a rule."""
     xml = extensions.get(CONDITIONAL_FORMATS)
     if xml is None:
         return
-    for identifier in identifiers:
-        pattern = (
-            rf'<x14:cfRule\b(?=[^>]*\bid="{re.escape(identifier)}")'
-            r"(?:[^>]*/>|[^>]*>.*?</x14:cfRule>)"
-        )
-        xml = re.sub(pattern, "", xml, flags=re.S)
-    emptied = (
-        r"<x14:conditionalFormatting\b[^>]*>(?:(?!<x14:cfRule\b).)*?"
-        r"</x14:conditionalFormatting>"
-    )
-    xml = re.sub(emptied, "", xml, flags=re.S)
+
+    def dropped(rule: re.Match[str]) -> str:
+        start_tag = rule[0][: rule[0].index(">") + 1]
+        return "" if identifiers.intersection(_RULE_ID_ATTRIBUTE.findall(start_tag)) else rule[0]
+
+    xml = _RULE.sub(dropped, xml)
+    xml = _RULE_GROUP.sub(lambda group: group[0] if "<x14:cfRule" in group[0] else "", xml)
     if "<x14:cfRule" in xml:
         extensions[CONDITIONAL_FORMATS] = xml
     else:
@@ -229,11 +233,21 @@ def _sparklines(xml: str, handle: Callable[[str], list[str]]) -> str | None:
     return xml if re.search(r"<x14:sparklineGroup\b", xml) else None
 
 
-def _moved_sparkline(item: str, rewriter: Rewriter) -> list[str]:
+def sparkline_location(item: str) -> str:
+    """The cell of a ``<x14:sparkline>``. Excel refuses a file with a sparkline that has none."""
     location = _RANGE.search(item)
+    if location is None:
+        raise WorkbookError(
+            "The workbook has a sparkline without a location (<xm:sqref>), which Excel "
+            "cannot open. Remove it from the file."
+        )
+    return location[1]
+
+
+def _moved_sparkline(item: str, rewriter: Rewriter) -> list[str]:
+    location = sparkline_location(item)
     source = _FORMULA.search(item)
-    assert location is not None
-    moved = rewriter.ranges(location[1], grow=False)
+    moved = rewriter.ranges(location, grow=False)
     if moved is None:
         return []
     item = _RANGE.sub(lambda _: f"<xm:sqref>{moved}</xm:sqref>", item)
@@ -243,7 +257,7 @@ def _moved_sparkline(item: str, rewriter: Rewriter) -> list[str]:
             item = item.replace(source[0], "")
         else:
             item = item.replace(source[0], f"<xm:f>{escape(read)}</xm:f>")
-    return [item, *_inserted_copies(item, location[1], rewriter)]
+    return [item, *_inserted_copies(item, location, rewriter)]
 
 
 def _moved_formula(match: re.Match[str], rewriter: Rewriter) -> str:

{excel_mcp_server-2.2.2 → excel_mcp_server-2.2.4}/src/excel_mcp/package/opc.py

@@ -4,7 +4,9 @@ import io
 import posixpath
 import re
 import zipfile
+from collections.abc import Callable
 from dataclasses import dataclass
+from itertools import count
 from xml.sax.saxutils import quoteattr
 
 from excel_mcp.errors import WorkbookError
@@ -75,15 +77,25 @@ def write_rels(rels: list[Rel]) -> bytes:
     return (XML_DECLARATION + "".join(lines)).encode()
 
 
-def free_id(rels: list[Rel], wanted: str) -> str:
-    """``wanted`` if it is a free ``rIdN``, else the first free one."""
+def free_ids(rels: list[Rel]) -> Callable[[str], str]:
+    """A function that gives, for each id wanted, that id if it is a free ``rIdN`` and else the
+    first free one, and then counts it as taken. Many calls cost no more than one scan."""
     taken = {rel.id for rel in rels}
-    if re.fullmatch(r"rId\d+", wanted) and wanted not in taken:
-        return wanted
-    number = 1
-    while f"rId{number}" in taken:
-        number += 1
-    return f"rId{number}"
+    numbers = count(1)
+
+    def allocate(wanted: str) -> str:
+        if re.fullmatch(r"rId\d+", wanted) and wanted not in taken:
+            chosen = wanted
+        else:
+            chosen = next(f"rId{n}" for n in numbers if f"rId{n}" not in taken)
+        taken.add(chosen)
+        return chosen
+
+    return allocate
+
+
+def free_id(rels: list[Rel], wanted: str) -> str:
+    return free_ids(rels)(wanted)
 
 
 class ContentTypes:
@@ -135,15 +147,16 @@ class Package:
 
     def __init__(self, data: bytes) -> None:
         self._source = zipfile.ZipFile(io.BytesIO(data))
+        self._source_names = set(self._source.namelist())
         self._changes: dict[str, bytes] = {}
         self._rels: dict[str, list[Rel]] = {}
         self.content_types = ContentTypes(self.read(CONTENT_TYPES))
 
     def names(self) -> set[str]:
-        return set(self._source.namelist()) | set(self._changes)
+        return self._source_names | set(self._changes)
 
     def exists(self, name: str) -> bool:
-        return name in self._changes or name in self._source.namelist()
+        return name in self._changes or name in self._source_names
 
     def read(self, name: str) -> bytes:
         if name in self._changes:
@@ -162,13 +175,12 @@ class Package:
 
     def unique_name(self, wanted: str) -> str:
         """``wanted``, or the same name with the next free number in it."""
-        taken = self.names()
-        if wanted not in taken:
+        if not self.exists(wanted):
             return wanted
         match = re.fullmatch(r"(.*?)(\d*)(\.[^./]*)?", wanted)
         stem, number, extension = match.groups() if match else (wanted, "", "")
         counter = int(number or 0) + 1
-        while f"{stem}{counter}{extension or ''}" in taken:
+        while self.exists(f"{stem}{counter}{extension or ''}"):
             counter += 1
         return f"{stem}{counter}{extension or ''}"
 
@@ -189,6 +201,6 @@ class Package:
                 elif content:
                     target.writestr(name, content)
             for name, content in self._changes.items():
-                if content and name not in self._source.namelist():
+                if content and name not in self._source_names:
                     target.writestr(name, content)
         return output.getvalue()

{excel_mcp_server-2.2.2 → excel_mcp_server-2.2.4}/src/excel_mcp/package/patch.py

@@ -2,6 +2,7 @@
 
 import re
 from collections.abc import Callable
+from itertools import pairwise
 
 from excel_mcp.errors import WorkbookError
 from excel_mcp.package.scan import Scan, scan
@@ -146,17 +147,44 @@ def _with_extension(rule: bytes, qname: str, extension: str, offset: int) -> Edi
     return (offset, offset + len(rule), text)
 
 
+# Looking each cell up by its address reads the sheet once per cell, which is fastest for a
+# few cells; for more, one pass over all the cells does not grow with the number to patch.
+_FEW_CELLS = 16
+_CELL = re.compile(rb'<c r="([^"]*)"([^>]*?)(?:/>|>(.*?)</c>)', re.S)
+
+
 def patch_cells(data: bytes, edits: dict[str, CellEdit]) -> bytes:
     """Rewrite chosen cells of ``sheetData``; an edit gets a cell's attributes and its content."""
     if not edits:
         return data
-    names = b"|".join(re.escape(name.encode()) for name in edits)
-    cell = re.compile(rb'<c r="(' + names + rb')"([^>]*?)(?:/>|>(.*?)</c>)', re.S)
+    if len(edits) <= _FEW_CELLS:
+        return splice(data, _located_edits(data, edits))
 
     def rewrite(match: re.Match[bytes]) -> bytes:
-        coordinate = match[1].decode()
-        attributes, body = edits[coordinate](match[2].decode(), (match[3] or b"").decode())
-        start = f'<c r="{coordinate}"{attributes}'
-        return f"{start}>{body}</c>".encode() if body else f"{start}/>".encode()
-
-    return cell.sub(rewrite, data)
+        edit = edits.get(match[1].decode())
+        return _rewritten(match, edit) if edit else match[0]
+
+    return _CELL.sub(rewrite, data)
+
+
+def _located_edits(data: bytes, edits: dict[str, CellEdit]) -> list[Edit]:
+    found: list[Edit] = []
+    for coordinate, edit in edits.items():
+        opening = b'<c r="%b"' % coordinate.encode()
+        start = data.find(opening)
+        while start >= 0:
+            match = _CELL.match(data, start)
+            if match is None:
+                raise WorkbookError(f"Cell {coordinate} of a worksheet is malformed.")
+            found.append((match.start(), match.end(), _rewritten(match, edit)))
+            start = data.find(opening, match.end())
+    found.sort()
+    if any(before[1] > after[0] for before, after in pairwise(found)):
+        raise WorkbookError("A worksheet has cells inside other cells.")
+    return found
+
+
+def _rewritten(match: re.Match[bytes], edit: CellEdit) -> bytes:
+    attributes, body = edit(match[2].decode(), (match[3] or b"").decode())
+    start = f'<c r="{match[1].decode()}"{attributes}'
+    return f"{start}>{body}</c>".encode() if body else f"{start}/>".encode()

{excel_mcp_server-2.2.2 → excel_mcp_server-2.2.4}/src/excel_mcp/package/references.py

@@ -26,6 +26,7 @@ from xml.sax.saxutils import escape
 from openpyxl import Workbook
 from openpyxl.worksheet.worksheet import Worksheet
 
+from excel_mcp.errors import WorkbookError
 from excel_mcp.package import extensions
 from excel_mcp.package.anchors import Geometry, Moved, move_anchors, move_vml_shape
 from excel_mcp.package.lines import LineEdit
@@ -120,7 +121,8 @@ def _ranged_items(xml: str, item: str, attribute: str, rewriter: _Rewriter) -> s
     def move(match: re.Match[str]) -> str:
         text = match[0]
         value = re.search(rf'\b{attribute}="([^"]*)"', text)
-        assert value is not None
+        if value is None:
+            raise WorkbookError(f"A <{item}> of the workbook has no {attribute} attribute.")
         moved = rewriter.ranges(value[1], grow=attribute == "sqref")
         return "" if moved is None else text.replace(value[0], f'{attribute}="{moved}"', 1)
 
@@ -134,7 +136,8 @@ def _sort_state(xml: str, rewriter: _Rewriter) -> str | None:
 
     def move(match: re.Match[str]) -> str:
         found = ref.search(match[0])
-        assert found is not None
+        if found is None:
+            raise WorkbookError("A <sortCondition> of the workbook has no ref attribute.")
         moved = rewriter.ranges(found[1])
         return "" if moved is None else match[0].replace(found[0], f'ref="{moved}"', 1)
 

{excel_mcp_server-2.2.2 → excel_mcp_server-2.2.4}/src/excel_mcp/package/restore.py

@@ -14,7 +14,7 @@ from excel_mcp.package.model import (
     state_of,
     workbook_sheets,
 )
-from excel_mcp.package.opc import Package, Rel, free_id
+from excel_mcp.package.opc import Package, Rel, free_ids
 from excel_mcp.package.patch import WORKBOOK_ORDER, append_to_root, insertions, splice
 from excel_mcp.package.restore_sheet import restore_sheet
 from excel_mcp.package.scan import attributes, rel_id, remap_relationship_ids, scan, tags
@@ -103,13 +103,18 @@ class Restorer:
     ) -> list[str]:
         """Add relationships to a part that is being merged into, and renumber the XML to match."""
         rels = self.package.rels(owner)
+        free_id = free_ids(rels)
+        known: dict[tuple[str, str], Rel] = {}
+        for rel in rels:
+            known.setdefault((rel.type, rel.target), rel)
         mapping: dict[str, str] = {}
         for link in links:
             rel = Rel(link.id, link.type, self.target(link), isinstance(link.target, str))
-            existing = next((r for r in rels if (r.type, r.target) == (rel.type, rel.target)), None)
+            existing = known.get((rel.type, rel.target))
             if existing is None:
-                rel.id = free_id(rels, link.id)
+                rel.id = free_id(link.id)
                 rels.append(rel)
+                known[rel.type, rel.target] = rel
             mapping[link.id] = (existing or rel).id
         return [remap_relationship_ids(text, mapping, scope) for text in texts]