RAGflow/deepdoc/parser/docx_parser.py

228 lines
7.9 KiB
Python
Raw Normal View History

2025-03-24 11:19:28 +08:00
#
# Copyright 2025 The InfiniFlow Authors. All Rights Reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
#
from docx import Document
import re
import pandas as pd
from collections import Counter
from rag.nlp import rag_tokenizer
from io import BytesIO
class RAGFlowDocxParser:
"""
Word文档(.docx)解析器用于提取文档中的文本内容和表格
该解析器能够
1. 按页面范围提取文档中的段落文本及其样式
2. 识别文档中的表格并将其转换为结构化文本
3. 智能处理表格头部和内容生成语义化的文本描述
"""
2025-03-24 11:19:28 +08:00
def __extract_table_content(self, tb):
"""
从Word表格对象中提取内容并转换为DataFrame
参数:
tb: docx库的Table对象
返回:
处理后的表格内容文本列表
"""
2025-03-24 11:19:28 +08:00
df = []
for row in tb.rows:
df.append([c.text for c in row.cells])
return self.__compose_table_content(pd.DataFrame(df))
def __compose_table_content(self, df):
"""
将表格DataFrame转换为语义化的文本描述
通过识别表格的结构特征(如表头数据类型等)将表格转换为更易于理解的文本形式
参数:
df: 包含表格内容的DataFrame
返回:
表格内容的文本表示列表
"""
2025-03-24 11:19:28 +08:00
def blockType(b):
"""
识别单元格内容的类型
通过正则表达式和文本特征分析将单元格内容分类为不同类型
- Dt: 日期类型
- Nu: 数字类型
- Ca: 代码/ID类型
- En: 英文文本
- NE: 数字和文本混合
- Sg: 单字符
- Tx: 短文本
- Lx: 长文本
- Nr: 人名
- Ot: 其他类型
参数:
b: 单元格文本内容
返回:
内容类型的字符串标识
"""
2025-03-24 11:19:28 +08:00
patt = [
("^(20|19)[0-9]{2}[年/-][0-9]{1,2}[月/-][0-9]{1,2}日*$", "Dt"),
(r"^(20|19)[0-9]{2}年$", "Dt"),
(r"^(20|19)[0-9]{2}[年/-][0-9]{1,2}月*$", "Dt"),
("^[0-9]{1,2}[月/-][0-9]{1,2}日*$", "Dt"),
(r"^第*[一二三四1-4]季度$", "Dt"),
(r"^(20|19)[0-9]{2}年*[一二三四1-4]季度$", "Dt"),
(r"^(20|19)[0-9]{2}[ABCDE]$", "DT"),
("^[0-9.,+%/ -]+$", "Nu"),
(r"^[0-9A-Z/\._~-]+$", "Ca"),
(r"^[A-Z]*[a-z' -]+$", "En"),
(r"^[0-9.,+-]+[0-9A-Za-z/$¥%<>()' -]+$", "NE"),
(r"^.{1}$", "Sg")
]
for p, n in patt:
if re.search(p, b):
return n
tks = [t for t in rag_tokenizer.tokenize(b).split() if len(t) > 1]
if len(tks) > 3:
if len(tks) < 12:
return "Tx"
else:
return "Lx"
if len(tks) == 1 and rag_tokenizer.tag(tks[0]) == "nr":
return "Nr"
return "Ot"
# 表格至少需要两行才能处理
2025-03-24 11:19:28 +08:00
if len(df) < 2:
return []
# 统计表格中最常见的内容类型
2025-03-24 11:19:28 +08:00
max_type = Counter([blockType(str(df.iloc[i, j])) for i in range(
1, len(df)) for j in range(len(df.iloc[i, :]))])
max_type = max(max_type.items(), key=lambda x: x[1])[0]
# 获取表格列数
2025-03-24 11:19:28 +08:00
colnm = len(df.iloc[0, :])
# 默认第一行为表头
hdrows = [0] # 表头不一定出现在第一行
# 如果表格主要是数字类型,则识别非数字行作为表头
2025-03-24 11:19:28 +08:00
if max_type == "Nu":
for r in range(1, len(df)):
tys = Counter([blockType(str(df.iloc[r, j]))
for j in range(len(df.iloc[r, :]))])
tys = max(tys.items(), key=lambda x: x[1])[0]
if tys != max_type:
hdrows.append(r)
# 处理表格内容,将每行转换为文本
2025-03-24 11:19:28 +08:00
lines = []
for i in range(1, len(df)):
# 跳过表头行
2025-03-24 11:19:28 +08:00
if i in hdrows:
continue
# 计算当前行之前的表头行
2025-03-24 11:19:28 +08:00
hr = [r - i for r in hdrows]
hr = [r for r in hr if r < 0]
# 找到最近的连续表头行
2025-03-24 11:19:28 +08:00
t = len(hr) - 1
while t > 0:
if hr[t] - hr[t - 1] > 1:
hr = hr[t:]
break
t -= 1
# 为每列构建表头描述
2025-03-24 11:19:28 +08:00
headers = []
for j in range(len(df.iloc[i, :])):
t = []
for h in hr:
x = str(df.iloc[i + h, j]).strip()
if x in t:
continue
t.append(x)
t = ",".join(t)
if t:
t += ": "
headers.append(t)
# 构建每行的文本表示
2025-03-24 11:19:28 +08:00
cells = []
for j in range(len(df.iloc[i, :])):
if not str(df.iloc[i, j]):
continue
cells.append(headers[j] + str(df.iloc[i, j]))
lines.append(";".join(cells))
# 根据列数决定返回格式
2025-03-24 11:19:28 +08:00
if colnm > 3:
return lines
return ["\n".join(lines)]
def __call__(self, fnm, from_page=0, to_page=100000000):
"""
解析Word文档提取指定页面范围内的文本和表格
参数:
fnm: 文件名或二进制内容
from_page: 起始页码(从0开始)
to_page: 结束页码
返回:
元组(secs, tbls)其中:
- secs: 段落内容列表每项为(文本内容, 样式名称)的元组
- tbls: 表格内容列表
"""
# 根据输入类型创建Document对象
2025-03-24 11:19:28 +08:00
self.doc = Document(fnm) if isinstance(
fnm, str) else Document(BytesIO(fnm))
pn = 0 # 当前解析页码
secs = [] # 存储解析的段落内容
# 遍历文档中的所有段落
2025-03-24 11:19:28 +08:00
for p in self.doc.paragraphs:
# 如果超出指定页码范围,停止解析
2025-03-24 11:19:28 +08:00
if pn > to_page:
break
runs_within_single_paragraph = [] # 保存在页面范围内的文本片段
# 遍历段落中的所有文本片段(run)
2025-03-24 11:19:28 +08:00
for run in p.runs:
if pn > to_page:
break
# 如果当前页码在指定范围内且段落有内容,则添加文本
2025-03-24 11:19:28 +08:00
if from_page <= pn < to_page and p.text.strip():
runs_within_single_paragraph.append(run.text) # 先添加文本片段
2025-03-24 11:19:28 +08:00
# 检查页面分隔符
2025-03-24 11:19:28 +08:00
if 'lastRenderedPageBreak' in run._element.xml:
pn += 1
# 将段落文本和样式添加到结果列表
secs.append(("".join(runs_within_single_paragraph), p.style.name if hasattr(p.style, 'name') else '')) # 然后将文本片段连接为段落的一部分
2025-03-24 11:19:28 +08:00
# 提取所有表格内容
2025-03-24 11:19:28 +08:00
tbls = [self.__extract_table_content(tb) for tb in self.doc.tables]
return secs, tbls