mirror of
https://github.com/haris-musa/excel-mcp-server.git
synced 2025-12-08 17:12:41 +08:00
Merge pull request #45 from haris-musa/fix/issue-40-read-range
Fix: Correctly read data when not starting at A1. Fixes #40
This commit is contained in:
Haris
@ -51,21 +51,25 @@ def read_excel_range(
|
||||
except ValueError as e:
|
||||
raise DataError(f"Invalid end cell format: {str(e)}")
|
||||
else:
|
||||
# Dynamically expand range until all values are empty
|
||||
end_row, end_col = start_row, start_col
|
||||
while end_row <= ws.max_row and any(ws.cell(row=end_row, column=c).value is not None for c in range(start_col, ws.max_column + 1)):
|
||||
end_row += 1
|
||||
while end_col <= ws.max_column and any(ws.cell(row=r, column=end_col).value is not None for r in range(start_row, ws.max_row + 1)):
|
||||
end_col += 1
|
||||
end_row -= 1 # Adjust back to last non-empty row
|
||||
end_col -= 1 # Adjust back to last non-empty column
|
||||
# If no end_cell, use the full data range of the sheet
|
||||
if ws.max_row == 1 and ws.max_column == 1 and ws.cell(1, 1).value is None:
|
||||
# Handle empty sheet
|
||||
end_row, end_col = start_row, start_col
|
||||
else:
|
||||
# Use the sheet's own boundaries
|
||||
start_row, start_col = ws.min_row, ws.min_column
|
||||
end_row, end_col = ws.max_row, ws.max_column
|
||||
|
||||
# Validate range bounds
|
||||
if start_row > ws.max_row or start_col > ws.max_column:
|
||||
raise DataError(
|
||||
f"Start cell out of bounds. Sheet dimensions are "
|
||||
f"A1:{get_column_letter(ws.max_column)}{ws.max_row}"
|
||||
# This case can happen if start_cell is outside the used area on a sheet with data
|
||||
# or on a completely empty sheet.
|
||||
logger.warning(
|
||||
f"Start cell {start_cell} is outside the sheet's data boundary "
|
||||
f"({get_column_letter(ws.min_column)}{ws.min_row}:{get_column_letter(ws.max_column)}{ws.max_row}). "
|
||||
f"No data will be read."
|
||||
)
|
||||
return []
|
||||
|
||||
data = []
|
||||
for row in range(start_row, end_row + 1):
|
||||
@ -131,91 +135,6 @@ def write_data(
|
||||
logger.error(f"Failed to write data: {e}")
|
||||
raise DataError(str(e))
|
||||
|
||||
def _looks_like_headers(row_dict):
|
||||
"""Check if a data row appears to be headers (keys match values)."""
|
||||
return all(
|
||||
isinstance(value, str) and str(value).strip() == str(key).strip()
|
||||
for key, value in row_dict.items()
|
||||
)
|
||||
|
||||
def _check_for_headers_above(worksheet, start_row, start_col, headers):
|
||||
"""Check if cells above start position contain headers."""
|
||||
if start_row <= 1:
|
||||
return False # Nothing above row 1
|
||||
|
||||
# Look for header-like content above
|
||||
for check_row in range(max(1, start_row - 5), start_row):
|
||||
# Count matches for this row
|
||||
header_count = 0
|
||||
cell_count = 0
|
||||
|
||||
for i, header in enumerate(headers):
|
||||
if i >= 10: # Limit check to first 10 columns for performance
|
||||
break
|
||||
|
||||
cell = worksheet.cell(row=check_row, column=start_col + i)
|
||||
cell_count += 1
|
||||
|
||||
# Check if cell is formatted like a header (bold)
|
||||
is_formatted = cell.font.bold if hasattr(cell.font, 'bold') else False
|
||||
|
||||
# Check for any content that could be a header
|
||||
if cell.value is not None:
|
||||
# Case 1: Direct match with expected header
|
||||
if str(cell.value).strip().lower() == str(header).strip().lower():
|
||||
header_count += 2 # Give higher weight to exact matches
|
||||
# Case 2: Any formatted cell with content
|
||||
elif is_formatted and cell.value:
|
||||
header_count += 1
|
||||
# Case 3: Any cell with content in the first row we check
|
||||
elif check_row == max(1, start_row - 5):
|
||||
header_count += 0.5
|
||||
|
||||
# If we have a significant number of matching cells, consider it a header row
|
||||
if cell_count > 0 and header_count >= cell_count * 0.5:
|
||||
return True
|
||||
|
||||
# No headers found above
|
||||
return False
|
||||
|
||||
def _determine_header_behavior(worksheet, start_row, start_col, data):
|
||||
"""Determine if headers should be written based on context."""
|
||||
if not data:
|
||||
return False # No data means no headers
|
||||
|
||||
# Check if we're in the title area (rows 1-4)
|
||||
if start_row <= 4:
|
||||
return False # Don't add headers in title area
|
||||
|
||||
# If we already have data in the sheet, be cautious about adding headers
|
||||
if worksheet.max_row > 1:
|
||||
# Check if the target row already has content
|
||||
has_content = any(
|
||||
worksheet.cell(row=start_row, column=start_col + i).value is not None
|
||||
for i in range(min(5, len(data[0].keys())))
|
||||
)
|
||||
|
||||
if has_content:
|
||||
return False # Don't overwrite existing content with headers
|
||||
|
||||
# Check if first row appears to be headers
|
||||
first_row_is_headers = _looks_like_headers(data[0])
|
||||
|
||||
# Check extensively for headers above (up to 5 rows)
|
||||
has_headers_above = _check_for_headers_above(worksheet, start_row, start_col, list(data[0].keys()))
|
||||
|
||||
# Be conservative - don't add headers if we detect headers above or the data has headers
|
||||
if has_headers_above or first_row_is_headers:
|
||||
return False
|
||||
|
||||
# If we're appending data immediately after existing data, don't add headers
|
||||
if any(worksheet.cell(row=start_row-1, column=start_col + i).value is not None
|
||||
for i in range(min(5, len(data[0].keys())))):
|
||||
return False
|
||||
|
||||
# For completely new sheets or empty areas far from content, add headers
|
||||
return True
|
||||
|
||||
def _write_data_to_worksheet(
|
||||
worksheet: Worksheet,
|
||||
data: list[list],
|
||||
@ -295,25 +214,32 @@ def read_excel_range_with_metadata(
|
||||
except ValueError as e:
|
||||
raise DataError(f"Invalid end cell format: {str(e)}")
|
||||
else:
|
||||
# Dynamically expand range until all values are empty
|
||||
end_row, end_col = start_row, start_col
|
||||
while end_row <= ws.max_row and any(ws.cell(row=end_row, column=c).value is not None for c in range(start_col, ws.max_column + 1)):
|
||||
end_row += 1
|
||||
while end_col <= ws.max_column and any(ws.cell(row=r, column=end_col).value is not None for r in range(start_row, ws.max_row + 1)):
|
||||
end_col += 1
|
||||
end_row -= 1 # Adjust back to last non-empty row
|
||||
end_col -= 1 # Adjust back to last non-empty column
|
||||
# If no end_cell, use the full data range of the sheet
|
||||
if ws.max_row == 1 and ws.max_column == 1 and ws.cell(1, 1).value is None:
|
||||
# Handle empty sheet
|
||||
end_row, end_col = start_row, start_col
|
||||
else:
|
||||
# Use the sheet's own boundaries, but respect the provided start_cell
|
||||
end_row, end_col = ws.max_row, ws.max_column
|
||||
# If start_cell is 'A1' (default), we should find the true start
|
||||
if start_cell == 'A1':
|
||||
start_row, start_col = ws.min_row, ws.min_column
|
||||
|
||||
# Validate range bounds
|
||||
if start_row > ws.max_row or start_col > ws.max_column:
|
||||
raise DataError(
|
||||
f"Start cell out of bounds. Sheet dimensions are "
|
||||
f"A1:{get_column_letter(ws.max_column)}{ws.max_row}"
|
||||
# This case can happen if start_cell is outside the used area on a sheet with data
|
||||
# or on a completely empty sheet.
|
||||
logger.warning(
|
||||
f"Start cell {start_cell} is outside the sheet's data boundary "
|
||||
f"({get_column_letter(ws.min_column)}{ws.min_row}:{get_column_letter(ws.max_column)}{ws.max_row}). "
|
||||
f"No data will be read."
|
||||
)
|
||||
return {"range": f"{start_cell}:", "sheet_name": sheet_name, "cells": []}
|
||||
|
||||
# Build structured cell data
|
||||
range_str = f"{get_column_letter(start_col)}{start_row}:{get_column_letter(end_col)}{end_row}"
|
||||
range_data = {
|
||||
"range": f"{start_cell}:{get_column_letter(end_col)}{end_row}" if end_cell else start_cell,
|
||||
"range": range_str,
|
||||
"sheet_name": sheet_name,
|
||||
"cells": []
|
||||
}
|
||||
|
||||
Reference in New Issue
Block a user