fix: 保留上传模板的文本对齐(居中/右对齐/两端)

- docx_to_html: Mammoth转换后用BeautifulSoup注入text-align CSS
- html_to_docx: 解析style属性还原WD_ALIGN_PARAGRAPH对齐
- 支持 center/right/justify 三种对齐

backend/app/services/document_processor.py

@@ -1,14 +1,44 @@
 import mammoth
 from io import BytesIO
+from bs4 import BeautifulSoup
 from docx import Document
 from docx.shared import Pt, Inches
 from docx.enum.text import WD_ALIGN_PARAGRAPH
 from app.services.file_storage import get_file_content
 
+ALIGN_MAP = {
+    WD_ALIGN_PARAGRAPH.CENTER: "center",
+    WD_ALIGN_PARAGRAPH.RIGHT: "right",
+    WD_ALIGN_PARAGRAPH.JUSTIFY: "justify",
+}
+
 
 async def docx_to_html(file_content: bytes) -> str:
     result = mammoth.convert_to_html(BytesIO(file_content))
-    return result.value
+    html = result.value
+
+    try:
+        doc = Document(BytesIO(file_content))
+        soup = BeautifulSoup(html, "html.parser")
+        paragraphs = doc.paragraphs
+        html_paras = soup.find_all("p")
+
+        for i, para in enumerate(paragraphs):
+            if i >= len(html_paras):
+                break
+            if para.alignment and para.alignment in ALIGN_MAP:
+                css = ALIGN_MAP[para.alignment]
+                existing_style = html_paras[i].get("style", "")
+                styles = f"text-align:{css}"
+                if existing_style:
+                    styles = existing_style.rstrip(";") + ";" + styles
+                html_paras[i]["style"] = styles
+
+        html = str(soup)
+    except Exception:
+        pass
+
+    return html
 
 
 def html_to_docx_bytes(html_content: str) -> bytes:
@@ -25,7 +55,7 @@ def html_to_docx_bytes(html_content: str) -> bytes:
         def __init__(self):
             super().__init__()
             self.paragraphs: list[dict] = []
-            self.current = {"runs": []}
+            self.current = {"runs": [], "align": None}
             self.in_paragraph = False
             self.current_run = {"text": "", "bold": False, "italic": False, "underline": False}
             self.tag_stack: list[str] = []
@@ -33,9 +63,18 @@ def html_to_docx_bytes(html_content: str) -> bytes:
 
         def handle_starttag(self, tag, attrs):
             tag_lower = tag.lower()
+            attrs_dict = dict(attrs)
             if tag_lower in ("p", "div", "li"):
                 self.in_paragraph = True
+                self.current = {"runs": [], "align": None}
                 self.current_run = {"text": "", "bold": False, "italic": False, "underline": False}
+                style = attrs_dict.get("style", "")
+                if "text-align:center" in style:
+                    self.current["align"] = "center"
+                elif "text-align:right" in style:
+                    self.current["align"] = "right"
+                elif "text-align:justify" in style:
+                    self.current["align"] = "justify"
             elif tag_lower in ("h1", "h2", "h3", "h4", "h5", "h6"):
                 self.in_paragraph = True
                 self.heading_level = int(tag_lower[1])
@@ -82,6 +121,14 @@ def html_to_docx_bytes(html_content: str) -> bytes:
         else:
             p = doc.add_paragraph()
 
+        align_val = para_data.get("align")
+        if align_val == "center":
+            p.alignment = WD_ALIGN_PARAGRAPH.CENTER
+        elif align_val == "right":
+            p.alignment = WD_ALIGN_PARAGRAPH.RIGHT
+        elif align_val == "justify":
+            p.alignment = WD_ALIGN_PARAGRAPH.JUSTIFY
+
         for run_data in para_data.get("runs", []):
             run = p.add_run(run_data["text"])
             if run_data.get("bold"):