Python代码实现轻松提取Word文档的正文与表格

pip install spire.doc.free

from spire.doc import *
from spire.doc.common import *

def extractText(docxPath):
    doc = Document()
    doc.LoadFromFile(docxPath)
    lines = []
    for i in range(doc.Sections.Count):
        section = doc.Sections.get_Item(i)
        for j in range(section.Paragraphs.Count):
            para = section.Paragraphs.get_Item(j)
            lines.append(para.Text)
    doc.Close()
    return lines

def cellText(cell):
    parts = []
    for k in range(cell.Paragraphs.Count):
        parts.append(cell.Paragraphs.get_Item(k).Text)
    return " ".join(parts)
def extractTables(docxPath):
    doc = Document()
    doc.LoadFromFile(docxPath)
    tables = []
    for i in range(doc.Sections.Count):
        section = doc.Sections.get_Item(i)
        for t in range(section.Tables.Count):
            table = section.Tables[t]
            rows = []
            for r in range(table.Rows.Count):
                row = table.Rows[r]
                rows.append([cellText(row.Cells[c]) for c in range(row.Cells.Count)])
            tables.append(rows)
    doc.Close()
    return tables
# 写入CSV
outDir   = "output_csv"
os.makedirs(outDir, exist_ok=True)
for i, table in enumerate(extractTables(docxPath), 1):
    path = os.path.join(outDir, f"{os.path.splitext(os.path.basename(docxPath))[0]}_{i:02d}.csv")
    with open(path, "w", newline="", encoding="utf-8-sig") as f:
        csv.writer(f).writerows(table)
    print("已写入:", path)

def extractByStyle(docxPath, styleName):
    doc = Document()
    doc.LoadFromFile(docxPath)
    items = []
    for i in range(doc.Sections.Count):
        section = doc.Sections.get_Item(i)
        for j in range(section.Paragraphs.Count):
            para = section.Paragraphs.get_Item(j)
            if para.StyleName is not None and para.StyleName == styleName:
                items.append(para.Text)
    doc.Close()
    return items

def extractBookmark(docxPath, bookmarkName):
    doc = Document()
    doc.LoadFromFile(docxPath)
    navigator = BookmarksNavigator(doc)
    navigator.MoveToBookmark(bookmarkName)
    textBodyPart = navigator.GetBookmarkContent()
    parts = []
    for i in range(textBodyPart.BodyItems.Count):
        item = textBodyPart.BodyItems.get_Item(i)
        if isinstance(item, Paragraph):
            para = item
            buf = []
            for j in range(para.ChildObjects.Count):
                child = para.ChildObjects.get_Item(j)
                if isinstance(child, TextRange):
                    buf.append(child.Text)
            parts.append("".join(buf))
        elif isinstance(item, Table):
            tbl = item
            lines = []
            for r in range(tbl.Rows.Count):
                row = tbl.Rows[r]
                lines.append("\t".join(cellText(row.Cells[c]) for c in range(row.Cells.Count)))
            parts.append("\n".join(lines))
    doc.Close()
    return "\n".join(parts)

FILES = [
    "技术服务合同_华东数据.docx",
    "技术服务合同_华北运维.docx",
    "技术服务合同_华南集成.docx",
]

summaryRows = []
totalAmount = 0
for name in FILES:
    tables = extractTables(os.path.join(ART_DIR, name))
    payTable = tables[0]
    for row in payTable[1:]:
        stage, ratio, amount = row[0], row[1], row[2]
        summaryRows.append([name, stage, ratio, amount])
        totalAmount += int(amount.replace(",", ""))