重构workflow

This commit is contained in:
xunbu
2025-07-30 20:48:11 +08:00
parent 8987e4ef60
commit d25f634e73
38 changed files with 351 additions and 286 deletions

View File

View File

@@ -0,0 +1,18 @@
from dataclasses import dataclass
from docutranslate.exporter.base import Exporter, ExporterConfig
from docutranslate.ir.document import Document
from docutranslate.ir.markdown_document import MarkdownDocument
@dataclass(kw_only=True)
class MDExporterConfig(ExporterConfig):
...
class MDExporter(Exporter):
def __init__(self, config: MDExporterConfig|None=None):
super().__init__(config=config)
def export(self, document: MarkdownDocument) -> Document:
...

View File

@@ -0,0 +1,73 @@
from dataclasses import dataclass
import jinja2
import markdown2
from docutranslate.exporter.md.base import MDExporter, MDExporterConfig
from docutranslate.ir.document import Document
from docutranslate.ir.markdown_document import MarkdownDocument
from docutranslate.utils.resource_utils import resource_path
@dataclass
class MD2HTMLExporterConfig(MDExporterConfig):
cdn: bool = True
class MD2HTMLExporter(MDExporter):
def __init__(self, config: MD2HTMLExporterConfig = None):
config = config or MD2HTMLExporterConfig()
super().__init__(config=config)
self.cdn=config.cdn
def export(self, document: MarkdownDocument) -> Document:
cdn = self.cdn
markdowner = markdown2.Markdown(extras=['tables', 'fenced-code-blocks', 'mermaid', "code-friendly"])
# language=html
pico = f'<style>{resource_path("static/pico.css").read_text(encoding="utf-8")}</style>' if not cdn else r'<link rel="stylesheet" href="https://s4.zstatic.net/ajax/libs/picocss/2.1.1/pico.min.css" integrity="sha512-+4kjFgVD0n6H3xt19Ox84B56MoS7srFn60tgdWFuO4hemtjhySKyW4LnftYZn46k3THUEiTTsbVjrHai+0MOFw==" crossorigin="anonymous" referrerpolicy="no-referrer" />'
html_template = resource_path("template/markdown.html").read_text(encoding="utf-8")
katex_css = f'<style>{resource_path("static/katex.css").read_text(encoding="utf-8")}</style>' if not cdn else r"""<link rel="stylesheet" href="https://s4.zstatic.net/ajax/libs/KaTeX/0.16.9/katex.min.css" integrity="sha512-fHwaWebuwA7NSF5Qg/af4UeDx9XqUpYpOGgubo3yWu+b2IQR4UeQwbb42Ti7gVAjNtVoI/I9TEoYeu9omwcC6g==" crossorigin="anonymous" referrerpolicy="no-referrer" />"""
katex_js = f'<script>{resource_path("static/katex.js").read_text(encoding="utf-8")}</script>' if not cdn else r"""<script defer src="https://s4.zstatic.net/ajax/libs/KaTeX/0.16.9/katex.min.js" integrity="sha512-LQNxIMR5rXv7o+b1l8+N1EZMfhG7iFZ9HhnbJkTp4zjNr5Wvst75AqUeFDxeRUa7l5vEDyUiAip//r+EFLLCyA==" crossorigin="anonymous" referrerpolicy="no-referrer"></script>"""
auto_render = f'<script>{resource_path("static/autoRender.js").read_text(encoding="utf-8")}</script>' if not cdn else r"""<script defer src="https://s4.zstatic.net/ajax/libs/KaTeX/0.16.9/contrib/auto-render.min.js" integrity="sha512-iWiuBS5nt6r60fCz26Nd0Zqe0nbk1ZTIQbl3Kv7kYsX+yKMUFHzjaH2+AnM6vp2Xs+gNmaBAVWJjSmuPw76Efg==" crossorigin="anonymous" referrerpolicy="no-referrer"></script>"""
# language=javascript
render_math_in_element = r"""
<script>
document.addEventListener("DOMContentLoaded", function () {
renderMathInElement(document.body, {
delimiters: [
{left: '$$', right: '$$', display: true},
{left: '\\[', right: '\\]', display: true},
{left: '$', right: '$', display: false},
{left: '\\(', right: '\\)', display: false}
],
throwOnError: false
})
});
</script>""" if cdn else r"""
<script>
document.addEventListener("DOMContentLoaded", function
() {
renderMathInElement(document.body, {
delimiters: [
{left: '$$', right: '$$', display: true},
{left: '\\[', right: '\\]', display: true},
{left: '$', right: '$', display: false},
{left: '\\(', right: '\\)', display: false}
],
fonts: false,
throwOnError: false
})
});
</script>"""
mermaid = f'<script>{resource_path("static/mermaid.js").read_text(encoding="utf-8")}</script>'
content = markdowner.convert(document.content.decode().replace("\\", "\\\\"))
# TODO:实现MathJax本地化
render = jinja2.Template(html_template).render(
title=document.stem,
pico=pico,
katexCss=katex_css,
katexJs=katex_js,
autoRender=auto_render,
markdown=content,
renderMathInElement=render_math_in_element,
mermaid=mermaid,
)
return Document.from_bytes(content=render.encode("utf-8"), suffix=".html", stem=document.stem)

View File

@@ -0,0 +1,8 @@
from docutranslate.exporter.md.base import MDExporter
from docutranslate.ir.markdown_document import MarkdownDocument, Document
class MD2MDExporter(MDExporter):
def export(self, document: MarkdownDocument) -> Document:
return Document.from_bytes(suffix=".md", content=document.content, stem=document.stem)

View File

@@ -0,0 +1,11 @@
from docutranslate.exporter.md.base import MDExporter
from docutranslate.ir.markdown_document import MarkdownDocument, Document
from docutranslate.utils.markdown_utils import unembed_base64_images_to_zip
class MD2MDZipExporter(MDExporter):
def export(self, document: MarkdownDocument) -> Document:
return Document.from_bytes(suffix=".zip", content=unembed_base64_images_to_zip(document.content.decode(),
markdown_name=document.name),
stem=document.stem)

View File

@@ -0,0 +1,3 @@
from typing import Literal
ConvertEnginType = Literal["mineru", "docling"]