Skip to main content

Document Upload

Introduction​

In whyhow, documents serve as the foundation for creating knowledge graphs. Whether you have structured data in CSV format or unstructured data in PDF files, you can easily upload and manage your documents within the platform. This guide will walk you through uploading documents to your workspace.

Uploading a Document​

  1. From the central console, navigate to the desired workspace where you want to upload the document.
  2. Click on the "Upload Document" button.
  3. Select the document you wish to upload in the file dialogue. Whyhow supports various file formats, including CSV, JSON, PDF and TXT.
  4. The platform will display a progress bar indicating the upload status. Wait for the upload to complete.

Add Documents

Document Processing​

After the document is uploaded, whyhow automatically processes it to extract relevant information and create chunks.

  • The platform identifies columns and rows for structured data (e.g., CSV files, JSON files), treating each row as a chunk.
  • For unstructured data (e.g., PDF files, TXT files), the platform applies natural language processing techniques to split the document into meaningful chunks based on paragraphs or sections.

The chunking process may take a few moments, depending on the size and complexity of the document.

Viewing and Managing Documents​

Once the document is processed, you can view and manage it within the workspace.

  1. In the workspace, you will see a list of all the uploaded documents.
  2. Click on a document to view its details, including the document name, file type, upload date, and number of chunks.
  3. From the document details page, you can:
    • Preview the document content
    • See the list of extracted chunks
    • Edit the document metadata (e.g., name, description)
    • Delete the document if no longer needed

Best Practices​

To ensure smooth document upload and processing, consider the following best practices:

  • Ensure that your documents are in a supported file format.
  • For structured data, ensure the CSV file has a header row defining the column names.
  • For unstructured data, provide clean and well-formatted documents to improve the accuracy of chunking and information extraction.
  • Organize your documents into relevant workspaces to keep your projects structured and manageable.
  • Regularly review and update your documents to maintain data quality and consistency.

Next Steps​

You can create schemas and generate knowledge graphs with uploaded and processed documents. Here are the following steps:

  1. Review the extracted chunks and verify the information is correctly identified.
  2. Create a schema that defines your knowledge graph's entities, relationships, and patterns.
  3. Use the schema and document chunks to generate a knowledge graph.
  4. Explore and visualize the generated graph to gain insights and discover connections.

Refer to the respective guides in the documentation for detailed instructions on each step.

If you encounter any issues during the document upload process or have questions, please get in touch with our support team at team@whyhow.ai.