Document Processing Issues

This guide helps you troubleshoot problems with uploading and processing documents in IntelliRepo.

Understanding Document Processing

When you upload a document, IntelliRepo:

1. Uploads the file to secure storage
2. Extracts text from the document
3. Chunks text into searchable segments
4. Embeds chunks for semantic search
5. Stores everything in the database

Each step can potentially fail. This guide covers common issues at each stage.

Document Status

Failed Uploads

Symptoms

• Status shows "Failed"
• Error message displayed
• Document not searchable

Common Causes

1. Unsupported File Format

Supported formats:

• PDF (.pdf)
• Microsoft Word (.docx)
• Plain text (.txt)
• Markdown ()
• HTML (.html, .htm)

Not supported:

• Legacy Word (.doc)
• Excel, PowerPoint
• Images (PNG, JPG)
• Scanned PDFs (image-only)

Solution: Convert to a supported format before uploading.

2. Corrupted File

Symptoms:

• File won't open on your computer
• Error when processing

Solution:

1. Try opening the file locally
2. If it doesn't open, the file is corrupted
3. Re-download or re-create the file
4. Upload the working version

3. Scanned/Image-Only PDF

Symptoms:

• PDF uploads but chat finds no content
• Status shows "Completed" but 0 chunks

Cause: The PDF contains images of text, not actual text.

How to check:

1. Open the PDF
2. Try to select/highlight text
3. If you can't select text, it's image-only

Solutions:

• Use a PDF with selectable text
• Run OCR on the scanned PDF first
• Use a text-based source document

4. File Too Large

Plan limits:

Solutions:

• Compress the file
• Split into multiple smaller files
• Remove unnecessary images/pages
• Upgrade your plan

5. Encoding Issues

Symptoms:

• Text file fails processing
• Strange characters in results

Cause: File is not UTF-8 encoded.

Solution:

1. Open in a text editor
2. Save with UTF-8 encoding
3. Re-upload

Processing Stuck

Symptoms

• Document shows "Processing" for more than 10 minutes
• No progress indication

Solutions

1. Refresh the page - Status may not have updated
2. Wait a bit longer - Very large files take more time
3. Delete and re-upload - If stuck for 15+ minutes

If Problems Persist

1. Note the file name and type
2. Check the file works locally
3. Contact support with details

Poor Text Extraction

Symptoms

• Search results are incomplete
• Answers miss obvious content
• Sources show garbled text

Causes and Solutions

Complex PDF Layouts

PDFs with multiple columns, tables, or sidebars may not extract perfectly.

Solutions:

• Simplify the document layout
• Use a single-column version
• Export as plain text if possible

Password-Protected PDFs

Protected PDFs cannot be processed.

Solution: Remove the password before uploading.

Forms and Interactive Elements

PDF forms, buttons, and interactive elements are ignored.

Solution: Flatten the PDF or save as a static document.

Very Large Tables

Complex tables may not extract with proper structure.

Solution: Consider extracting table data separately as text.

Zero Chunks After Processing

Symptoms

• Status shows "Completed"
• Chunk count shows 0
• Document not appearing in search results

Causes

1. Empty document - File contains no text
2. Image-only PDF - No extractable text
3. Unsupported content - Only images, diagrams, etc.

Solutions

1. Check that the document actually contains text
2. Try selecting text in the original file
3. Convert images to text using OCR before uploading

Slow Processing

Normal Processing Times

If Unusually Slow

1. Large files take longer - be patient
2. Peak usage times may affect speed
3. Multiple uploads are processed sequentially

Best Practices for Documents

Before Uploading

• Verify file opens correctly locally
• Confirm text is selectable (not scanned)
• Check file size is within limits
• Use supported format (.pdf, .docx, .txt, )

Document Quality Tips

• Use text-based PDFs whenever possible
• Keep formatting simple - single columns work best
• Include clear headings for better search
• Remove sensitive data before uploading

File Naming

• Use descriptive names: "Employee-Handbook-2024.pdf"
• Avoid special characters: use hyphens or underscores
• Include version/date if relevant

Re-Processing Documents

Currently, there's no "re-process" button. To re-process:

1. Delete the document
2. Upload it again

This is useful after:

• Fixing the source file
• System updates that improve processing

Still Having Issues?

If you've tried the above and still have problems:

1. Note the exact error (screenshot if possible)
2. Save the problematic file (we may ask for it)
3. Contact support with:
   • File name and type
   • File size
   • Error message
   • What you've tried

Contact Support