Upload and File Issues

AcaTrove uses a unified upload system across publications, projects, labs, and profile documents. This article covers common upload problems and how to resolve them.

File Size Limits

Each upload context has a specific maximum file size:

| Context | Maximum Size | |---------|-------------| | Publications | 100 MB | | Project documents | 100 MB | | Lab datasets | 500 MB | | CVs and resumes | 10 MB | | Teaching materials | 50 MB | | Profile pictures | 1 MB |

If your file exceeds the limit, you will see "File too large" with the maximum allowed size. Compress the file or split it into smaller parts before uploading.

Supported File Formats

AcaTrove accepts different formats depending on the upload context:

• Publications: PDF, DOC, DOCX, TXT, MD, RTF, PPT, PPTX, XLS, XLSX, CSV, and scientific formats (TEX, BIB, IPYNB).
• CVs: PDF, DOCX, DOC, TXT, TEX.
• Profile pictures: PNG, JPG, JPEG, GIF, SVG, WEBP.
• General documents: PDF, DOC, DOCX, TXT, MD, RTF.

If you see "File type not allowed," check that your file extension matches the accepted types for the context you are uploading to. Renamed files with incorrect extensions will also be rejected.

Upload Failures

"Upload failed" or "Failed to upload file":

1. Check your network connection. Large file uploads are particularly sensitive to connection interruptions.
2. The upload component includes automatic retry logic. If the first attempt fails, it will retry automatically. Wait for the retry to complete before taking manual action.
3. If using drag-and-drop, try the manual file picker instead. Click the upload area to open the file browser.
4. Ensure you are not uploading more files at once than the component allows. Some contexts limit simultaneous uploads.

Upload stalls at a percentage:

1. This usually indicates a network timeout. Refresh the page and try again.
2. For files near the size limit, a wired connection is recommended over Wi-Fi.
3. Corporate firewalls and proxy servers can interfere with large uploads. If you are on an institutional network, try from an unrestricted connection.

Metadata Extraction Problems

When you upload a publication, AcaTrove attempts to extract metadata such as the title, authors, abstract, and DOI. This process uses AI-powered analysis of the document content.

If metadata is missing or incorrect:

1. Text-based PDFs yield the best extraction results. Scanned or image-heavy PDFs may produce incomplete metadata because the text layer is limited.
2. You can manually edit extracted metadata after upload. Navigate to the publication detail page and update any incorrect fields.
3. If AcaTrove cannot extract text from your PDF at all, the document may be image-only. Consider using OCR software to create a text-layer PDF before uploading.

Paper Chat After Upload

If you upload a document but Paper Chat cannot answer questions about it, the document may still be processing. AcaTrove chunks the document text and generates vector embeddings for semantic search. For large documents, this can take up to a minute. Wait briefly and try again.

If Paper Chat returns generic responses that do not reference your document, verify that the Research AI Panel scope is set to the correct publication. The scope indicator appears at the top of the panel.

Tips

• PDF is the most reliable format for publications. It provides the best metadata extraction, text parsing, and Paper Chat experience.
• Keep file names descriptive but concise. Avoid special characters in file names, as some systems may not handle them correctly.
• If an upload fails repeatedly, try a different browser. Chrome, Firefox, Safari, and Edge are all supported, but occasionally a browser-specific issue can interfere.