Document Lineage

What this page is for

Browse documents within a collection, grouped by source, with quality aggregation and lineage metadata.

Endpoint

• GET /api/collections/:collectionId/documents

How it works

Every chunk stores its source document metadata in the doc.* payload fields. The documents endpoint aggregates chunks by source_id to provide a document-level view of your collection, including chunk counts and average quality scores.

Document metadata

Each document summary includes:

Field	Description
sourceId	Unique identifier (MD5 of source URL)
title	Document title
sourceType	web, manual, or file
sourceUrl	Original source URL
filename	Original filename (file uploads only)
mimeType	File MIME type (file uploads only)
chunkCount	Number of chunks from this document
avgQualityScore	Average quality score across chunks
ingestDate	When the document was first ingested
lastModifiedAt	Most recent modification timestamp

Example

curl http://localhost:3000/api/collections/<collectionId>/documents \
  -H "X-User-ID: user-1"

Response

{
  "documents": [
    {
      "sourceId": "a1b2c3d4",
      "title": "API Authentication Guide",
      "sourceType": "web",
      "sourceUrl": "https://example.com/api-auth-guide",
      "filename": null,
      "mimeType": null,
      "chunkCount": 8,
      "avgQualityScore": 85.5,
      "ingestDate": "2025-01-10T09:00:00.000Z",
      "lastModifiedAt": "2025-01-15T14:30:00.000Z"
    }
  ],
  "total": 1
}

Verify

• Response lists all documents in the collection grouped by source.
• chunkCount matches the actual number of chunks per document.
• avgQualityScore reflects quality scores set via the collection editor.

Next steps

• Collections — manage chunks within a collection.
• Publishing — publish new content to a collection.