'commit'

dsRag/Test/TestReadGongShi.py

@@ -1,6 +1,4 @@
 from docx import Document
-from docx.oxml.shared import qn
-from docx.oxml import parse_xml
 
 def run_has_ole_object(run):
     """

dsRag/Test/TestReadSpire.py

@@ -0,0 +1,32 @@
+# https://www.e-iceblue.cn/doc_python_other/python-insert-or-extract-ole-objects-in-word.html
+# pip install Spire.Doc
+from spire.doc import *
+from spire.doc.common import *
+
+# 创建Document类的对象
+doc = Document()
+# 加载Word文档
+doc.LoadFromFile(r'D:\dsWork\dsProject\dsRag\static\Txt\化学方程式_CHEMISTRY_1.docx')
+
+i = 1
+# 遍历Word文档的所有节
+for k in range(doc.Sections.Count):
+    sec = doc.Sections.get_Item(k)
+    # 遍历每个节的所有子对象
+    for j in range(sec.Body.ChildObjects.Count):
+        obj = sec.Body.ChildObjects.get_Item(j)
+        # 检查子对象是否为段落
+        if isinstance(obj, Paragraph):
+            par = obj if isinstance(obj, Paragraph) else None
+            # 遍历段落中的子对象
+            for m in range(par.ChildObjects.Count):
+                o = par.ChildObjects.get_Item(m)
+                # 检查子对象是否为OLE对象
+                if o.DocumentObjectType == DocumentObjectType.OleObject:
+                    ole = o if isinstance(o, DocOleObject) else None
+                    s = ole.ObjectType
+                    if s.startswith("Equation.DSMT4"):
+                        ext = ".mathtype"
+                        print("equation")
+
+doc.Close()