Leon/docutranslate
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

修复空白分块造成翻译错误的问题

Browse Source
This commit is contained in:
xunbu
2025-05-21 11:02:08 +08:00
parent c29ea337e0
commit 21c12d3b45
4 changed files with 37 additions and 37 deletions
Download Patch File Download Diff File Expand all files Collapse all files

17
.idea/workspace.xml generated
View File

@@ -5,19 +5,10 @@
</component>
<component name="ChangeListManager">
<list default="true" id="6b18b44a-df57-4212-a857-9e291ebe5dd2" name="更改" comment="">
<change afterPath="$PROJECT_DIR$/docutranslate/global_values/__init__.py" afterDir="false" />
<change afterPath="$PROJECT_DIR$/docutranslate/global_values/conditional_import.py" afterDir="false" />
<change afterPath="$PROJECT_DIR$/lite.spec" afterDir="false" />
<change beforePath="$PROJECT_DIR$/.idea/workspace.xml" beforeDir="false" afterPath="$PROJECT_DIR$/.idea/workspace.xml" afterDir="false" />
<change beforePath="$PROJECT_DIR$/README.md" beforeDir="false" afterPath="$PROJECT_DIR$/README.md" afterDir="false" />
<change beforePath="$PROJECT_DIR$/app.spec" beforeDir="false" afterPath="$PROJECT_DIR$/full.spec" afterDir="false" />
<change beforePath="$PROJECT_DIR$/docutranslate/__init__.py" beforeDir="false" afterPath="$PROJECT_DIR$/docutranslate/__init__.py" afterDir="false" />
<change beforePath="$PROJECT_DIR$/docutranslate/app.py" beforeDir="false" afterPath="$PROJECT_DIR$/docutranslate/app.py" afterDir="false" />
<change beforePath="$PROJECT_DIR$/docutranslate/converter/__init__.py" beforeDir="false" afterPath="$PROJECT_DIR$/docutranslate/converter/__init__.py" afterDir="false" />
<change beforePath="$PROJECT_DIR$/docutranslate/agents/agent.py" beforeDir="false" afterPath="$PROJECT_DIR$/docutranslate/agents/agent.py" afterDir="false" />
<change beforePath="$PROJECT_DIR$/docutranslate/static/index.html" beforeDir="false" afterPath="$PROJECT_DIR$/docutranslate/static/index.html" afterDir="false" />
<change beforePath="$PROJECT_DIR$/docutranslate/translater.py" beforeDir="false" afterPath="$PROJECT_DIR$/docutranslate/translater.py" afterDir="false" />
<change beforePath="$PROJECT_DIR$/pyproject.toml" beforeDir="false" afterPath="$PROJECT_DIR$/pyproject.toml" afterDir="false" />
<change beforePath="$PROJECT_DIR$/uv.lock" beforeDir="false" afterPath="$PROJECT_DIR$/uv.lock" afterDir="false" />
<change beforePath="$PROJECT_DIR$/docutranslate/utils/markdown_splitter.py" beforeDir="false" afterPath="$PROJECT_DIR$/docutranslate/utils/markdown_splitter.py" afterDir="false" />
</list>
<option name="SHOW_DIALOG" value="false" />
<option name="HIGHLIGHT_CONFLICTS" value="true" />
@@ -612,7 +603,7 @@
<workItem from="1747739438735" duration="826000" />
<workItem from="1747740341909" duration="145000" />
<workItem from="1747752718385" duration="81000" />
<workItem from="1747754618316" duration="9919000" />
<workItem from="1747754618316" duration="12015000" />
</task>
<servers />
</component>
@@ -631,7 +622,7 @@
</option>
</component>
<component name="com.intellij.coverage.CoverageDataManagerImpl">
<SUITE FILE_PATH="coverage/filetranslate$app_test__1_.coverage" NAME="app_test (1) 覆盖结果" MODIFIED="1747794139415" SOURCE_PROVIDER="com.intellij.coverage.DefaultCoverageFileProvider" RUNNER="coverage.py" COVERAGE_BY_TEST_ENABLED="false" COVERAGE_TRACING_ENABLED="false" WORKING_DIRECTORY="$PROJECT_DIR$/tests" />
<SUITE FILE_PATH="coverage/filetranslate$app_test__1_.coverage" NAME="app_test (1) 覆盖结果" MODIFIED="1747796284047" SOURCE_PROVIDER="com.intellij.coverage.DefaultCoverageFileProvider" RUNNER="coverage.py" COVERAGE_BY_TEST_ENABLED="false" COVERAGE_TRACING_ENABLED="false" WORKING_DIRECTORY="$PROJECT_DIR$/tests" />
<SUITE FILE_PATH="coverage/filetranslate$test.coverage" NAME="test 覆盖结果" MODIFIED="1747472297913" SOURCE_PROVIDER="com.intellij.coverage.DefaultCoverageFileProvider" RUNNER="coverage.py" COVERAGE_BY_TEST_ENABLED="false" COVERAGE_TRACING_ENABLED="false" WORKING_DIRECTORY="$PROJECT_DIR$/tests" />
<SUITE FILE_PATH="coverage/filetranslate$convert.coverage" NAME="convert 覆盖结果" MODIFIED="1746963490689" SOURCE_PROVIDER="com.intellij.coverage.DefaultCoverageFileProvider" RUNNER="coverage.py" COVERAGE_BY_TEST_ENABLED="false" COVERAGE_TRACING_ENABLED="false" WORKING_DIRECTORY="$PROJECT_DIR$/docutranslate/utils" />
<SUITE FILE_PATH="coverage/PDFtranslate$PDFtranslater__1_.coverage" NAME="PDFtranslater (1) 覆盖结果" MODIFIED="1746633258205" SOURCE_PROVIDER="com.intellij.coverage.DefaultCoverageFileProvider" RUNNER="coverage.py" COVERAGE_BY_TEST_ENABLED="false" COVERAGE_TRACING_ENABLED="false" WORKING_DIRECTORY="$PROJECT_DIR$/pdftranslate_packages" />

1
docutranslate/agents/agent.py
View File

@@ -84,6 +84,7 @@ class Agent:
result = response.json()["choices"][0]["message"]["content"]
return result
except httpx.HTTPStatusError as e:
print(f"prompt：{prompt}\nsystem_prompt:{system_prompt}")
raise Exception(f"AI请求错误 (async): {e.response.status_code} - {e.response.text}") from e
except httpx.RequestError as e:
raise Exception(f"AI请求连接错误 (async): {e}") from e

34
docutranslate/static/index.html
View File

@@ -600,22 +600,24 @@
.then(htmlContent => {
iframe.onload = () => {
iframe.onload = null;
try {
const iframeWindow = iframe.contentWindow;
if (!iframeWindow) throw new Error("无法访问打印框架。");
iframeWindow.document.title = currentFileName + '_translated.pdf';
iframeWindow.focus();
iframeWindow.print();
} catch (err) {
console.error('打印PDF出错:', err);
statusMsg.textContent = '无法直接生成PDF。请预览HTML后，使用浏览器的打印功能 (Ctrl+P) 保存。';
statusMsg.className = 'error-message';
} finally {
setTimeout(() => {
downloadPdfBtn.disabled = false;
downloadPdfBtn.textContent = '下载 PDF';
}, 2000);
}
setTimeout(() => {
try {
const iframeWindow = iframe.contentWindow;
if (!iframeWindow) throw new Error("无法访问打印框架。");
iframeWindow.document.title = currentFileName + '_translated.pdf';
iframeWindow.focus();
iframeWindow.print();
} catch (err) {
console.error('打印PDF出错:', err);
statusMsg.textContent = '无法直接生成PDF。请预览HTML后，使用浏览器的打印功能 (Ctrl+P) 保存。';
statusMsg.className = 'error-message';
} finally {
setTimeout(() => {
downloadPdfBtn.disabled = false;
downloadPdfBtn.textContent = '下载 PDF';
}, 2000);
}
}, 500)
};
iframe.srcdoc = htmlContent;
})

20
docutranslate/utils/markdown_splitter.py
View File

@@ -11,10 +11,12 @@ class MarkdownBlockSplitter:
max_block_size: 每个块的最大字符数
"""
self.max_block_size = max_block_size
@staticmethod
def _get_bytes(text:str)->int:
def _get_bytes(text: str) -> int:
return len(text.encode('utf-8'))
#TODO: 修复分块有时候会有空白块的问题
def split_markdown(self, markdown_text: str) -> List[str]:
"""
将Markdown文本分割成指定大小的块
@@ -230,18 +232,22 @@ def split_markdown_text(markdown_text, max_block_size=5000):
可以通过''.join(chunks)重建的Markdown块列表
"""
splitter = MarkdownBlockSplitter(max_block_size=max_block_size)
return splitter.split_markdown(markdown_text)
chunks = splitter.split_markdown(markdown_text)
#过滤空白块
chunks=[chunk for chunk in chunks if chunk.strip()]
return chunks
def join_markdown_texts(markdown_texts:list[str])->str:
result=""
def join_markdown_texts(markdown_texts: list[str]) -> str:
result = ""
for text in markdown_texts:
#只有表格会收到多余空行的影响
# 只有表格会收到多余空行的影响
if text.lstrip().startswith("|"):
result=result+"\n"+text
result = result + "\n" + text
else:
result+="\n\n"+text
result += "\n\n" + text
return result
if __name__ == '__main__':
pass