تجزیه جداول OneNote در Python — استخراج داده‌های ساختاریافته از فایل‌های .one

pip install aspose-note

from aspose.note import Document, Table

doc = Document("MyNotes.one")
tables = doc.GetChildNodes(Table)
print(f"Found {len(tables)} table(s)")

from aspose.note import Document, Table, TableRow, TableCell, RichText

doc = Document("MyNotes.one")

for t_num, table in enumerate(doc.GetChildNodes(Table), start=1):
    print(f"\nTable {t_num}:")
    for r_num, row in enumerate(table.GetChildNodes(TableRow), start=1):
        cells = row.GetChildNodes(TableCell)
        row_values = [
            " ".join(rt.Text for rt in cell.GetChildNodes(RichText)).strip()
            for cell in cells
        ]
        print(f"  Row {r_num}: {row_values}")

from aspose.note import Document, Table

doc = Document("MyNotes.one")
for i, table in enumerate(doc.GetChildNodes(Table), start=1):
    print(f"Table {i}: {len(table.ColumnWidths)} column(s)")
    print(f"  Widths (pts): {table.ColumnWidths}")
    print(f"  Borders visible: {table.BordersVisible}")

import csv, io
from aspose.note import Document, Table, TableRow, TableCell, RichText

doc = Document("MyNotes.one")
output = io.StringIO()
writer = csv.writer(output)

for table in doc.GetChildNodes(Table):
    for row in table.GetChildNodes(TableRow):
        values = [
            " ".join(rt.Text for rt in cell.GetChildNodes(RichText)).strip()
            for cell in row.GetChildNodes(TableCell)
        ]
        writer.writerow(values)
    writer.writerow([])   # blank row between tables

with open("tables.csv", "w", encoding="utf-8", newline="") as f:
    f.write(output.getvalue())

print("Saved tables.csv")

import json
from aspose.note import Document, Table, TableRow, TableCell, RichText

doc = Document("MyNotes.one")
result = []

for table in doc.GetChildNodes(Table):
    rows = []
    for row in table.GetChildNodes(TableRow):
        cells = [
            " ".join(rt.Text for rt in cell.GetChildNodes(RichText)).strip()
            for cell in row.GetChildNodes(TableCell)
        ]
        rows.append(cells)
    result.append({"rows": rows, "column_widths": table.ColumnWidths})

print(json.dumps(result, indent=2))

from aspose.note import Document, Table, TableRow, TableCell, RichText

doc = Document("MyNotes.one")

for table in doc.GetChildNodes(Table):
    rows = table.GetChildNodes(TableRow)
    if not rows:
        continue

    def row_text(row):
        return [
            " ".join(rt.Text for rt in cell.GetChildNodes(RichText)).strip()
            for cell in row.GetChildNodes(TableCell)
        ]

    headers = row_text(rows[0])
    print("Headers:", headers)
    for row in rows[1:]:
        record = dict(zip(headers, row_text(row)))
        print("  Record:", record)