mirror of
https://github.com/infiniflow/ragflow.git
synced 2025-12-08 12:32:30 +08:00
Feat: optimize the information displayed when .doc preview is unavailable (#11684)
### What problem does this PR solve? Feat: optimize the information displayed when .doc preview is unavailable #11605 ### Type of change - [X] New Feature (non-breaking change which adds functionality) #### Performance (Before) <img width="700" alt="image" src="https://github.com/user-attachments/assets/15cf69ee-3698-4e18-8e8f-bb75c321334d" /> #### Performance (After) 
This commit is contained in:
@ -173,8 +173,8 @@ def install_mineru() -> None:
|
|||||||
Logging is used to indicate status.
|
Logging is used to indicate status.
|
||||||
"""
|
"""
|
||||||
# Check if MinerU is enabled
|
# Check if MinerU is enabled
|
||||||
use_mineru = os.getenv("USE_MINERU", "").strip().lower()
|
use_mineru = os.getenv("USE_MINERU", "false").strip().lower()
|
||||||
if use_mineru == "false":
|
if use_mineru != "true":
|
||||||
logging.info("USE_MINERU=%r. Skipping MinerU installation.", use_mineru)
|
logging.info("USE_MINERU=%r. Skipping MinerU installation.", use_mineru)
|
||||||
return
|
return
|
||||||
|
|
||||||
|
|||||||
@ -1,7 +1,5 @@
|
|||||||
import message from '@/components/ui/message';
|
import message from '@/components/ui/message';
|
||||||
import { Spin } from '@/components/ui/spin';
|
import { Spin } from '@/components/ui/spin';
|
||||||
import { Authorization } from '@/constants/authorization';
|
|
||||||
import { getAuthorization } from '@/utils/authorization-util';
|
|
||||||
import request from '@/utils/request';
|
import request from '@/utils/request';
|
||||||
import classNames from 'classnames';
|
import classNames from 'classnames';
|
||||||
import mammoth from 'mammoth';
|
import mammoth from 'mammoth';
|
||||||
@ -16,22 +14,55 @@ export const DocPreviewer: React.FC<DocPreviewerProps> = ({
|
|||||||
className,
|
className,
|
||||||
url,
|
url,
|
||||||
}) => {
|
}) => {
|
||||||
// const url = useGetDocumentUrl();
|
|
||||||
const [htmlContent, setHtmlContent] = useState<string>('');
|
const [htmlContent, setHtmlContent] = useState<string>('');
|
||||||
const [loading, setLoading] = useState(false);
|
const [loading, setLoading] = useState(false);
|
||||||
|
|
||||||
const fetchDocument = async () => {
|
const fetchDocument = async () => {
|
||||||
|
if (!url) return;
|
||||||
|
|
||||||
setLoading(true);
|
setLoading(true);
|
||||||
|
|
||||||
const res = await request(url, {
|
const res = await request(url, {
|
||||||
method: 'GET',
|
method: 'GET',
|
||||||
responseType: 'blob',
|
responseType: 'blob',
|
||||||
headers: { [Authorization]: getAuthorization() },
|
|
||||||
onError: () => {
|
onError: () => {
|
||||||
message.error('Document parsing failed');
|
message.error('Document parsing failed');
|
||||||
console.error('Error loading document:', url);
|
console.error('Error loading document:', url);
|
||||||
},
|
},
|
||||||
});
|
});
|
||||||
|
|
||||||
try {
|
try {
|
||||||
const arrayBuffer = await res.data.arrayBuffer();
|
const blob: Blob = res.data;
|
||||||
|
const contentType: string =
|
||||||
|
blob.type || (res as any).headers?.['content-type'] || '';
|
||||||
|
|
||||||
|
// ---- Detect legacy .doc via MIME or URL ----
|
||||||
|
const cleanUrl = url.split(/[?#]/)[0].toLowerCase();
|
||||||
|
const isDocMime = /application\/msword/i.test(contentType);
|
||||||
|
const isLegacyDocByUrl =
|
||||||
|
cleanUrl.endsWith('.doc') && !cleanUrl.endsWith('.docx');
|
||||||
|
const isLegacyDoc = isDocMime || isLegacyDocByUrl;
|
||||||
|
|
||||||
|
if (isLegacyDoc) {
|
||||||
|
// Do not call mammoth and do not throw an error; instead, show a note in the preview area
|
||||||
|
setHtmlContent(`
|
||||||
|
<div class="flex h-full items-center justify-center">
|
||||||
|
<div class="border border-dashed border-border-normal rounded-xl p-8 max-w-2xl text-center">
|
||||||
|
<p class="text-2xl font-bold mb-4">
|
||||||
|
Preview not available for .doc files
|
||||||
|
</p>
|
||||||
|
<p class="italic text-sm text-muted-foreground leading-relaxed">
|
||||||
|
Mammoth does not support <code>.doc</code> documents.<br/>
|
||||||
|
Inline preview is unavailable.
|
||||||
|
</p>
|
||||||
|
</div>
|
||||||
|
</div>
|
||||||
|
`);
|
||||||
|
return;
|
||||||
|
}
|
||||||
|
|
||||||
|
// ---- Standard .docx preview path ----
|
||||||
|
const arrayBuffer = await blob.arrayBuffer();
|
||||||
const result = await mammoth.convertToHtml(
|
const result = await mammoth.convertToHtml(
|
||||||
{ arrayBuffer },
|
{ arrayBuffer },
|
||||||
{ includeDefaultStyleMap: true },
|
{ includeDefaultStyleMap: true },
|
||||||
@ -43,10 +74,12 @@ export const DocPreviewer: React.FC<DocPreviewerProps> = ({
|
|||||||
|
|
||||||
setHtmlContent(styledContent);
|
setHtmlContent(styledContent);
|
||||||
} catch (err) {
|
} catch (err) {
|
||||||
|
// Only errors from the mammoth conversion path should surface here
|
||||||
message.error('Document parsing failed');
|
message.error('Document parsing failed');
|
||||||
console.error('Error parsing document:', err);
|
console.error('Error parsing document:', err);
|
||||||
|
} finally {
|
||||||
|
setLoading(false);
|
||||||
}
|
}
|
||||||
setLoading(false);
|
|
||||||
};
|
};
|
||||||
|
|
||||||
useEffect(() => {
|
useEffect(() => {
|
||||||
@ -54,6 +87,7 @@ export const DocPreviewer: React.FC<DocPreviewerProps> = ({
|
|||||||
fetchDocument();
|
fetchDocument();
|
||||||
}
|
}
|
||||||
}, [url]);
|
}, [url]);
|
||||||
|
|
||||||
return (
|
return (
|
||||||
<div
|
<div
|
||||||
className={classNames(
|
className={classNames(
|
||||||
|
|||||||
Reference in New Issue
Block a user