@cpod/sdk - v0.2.7
    Preparing search index...

    Function searchDocumentsStream

    • Search documents with streaming results for large result sets

      Streams search results in chunks using NDJSON format. Ideal for:

      • Large exports (thousands of documents)
      • Real-time progress tracking
      • Memory-efficient bulk operations

      Parameters

      • request: SearchRequest

        Search request parameters (same as searchDocuments)

        Request parameters for document search

        • query: string

          Search query string (required)

        • OptionalviewMode?: ViewMode

          View mode: files or projects

        • Optionalsources?: string[] | null

          Filter by specific sources

        • Optionalentities?: string[] | null

          Filter by entity names

        • OptionalfileHashes?: string[] | null

          Filter by specific file hashes

        • OptionalchunkIds?: string[] | null

          Filter by specific chunk IDs

        • Optionalpage?: number

          Page number (1-indexed, default: 1)

        • Optionallimit?: number

          Maximum results to return (1-100, default: 10)

        • OptionalsortBy?: "relevance" | "date" | "source" | "chunk_index"

          Sort order: "relevance" | "date" | "source" | "chunk_index" (default: "relevance")

        • OptionalsortOrder?: "asc" | "desc"

          Sort direction

        • OptionalincludeContent?: boolean

          Include full content in results (default: true)

        • OptionalincludeEntities?: boolean

          Include entities in results (default: true)

        • OptionalincludeSummary?: boolean

          Include AI summary in results

        • OptionaluserIds?: string[] | null

          Filter results to specific users' data (admin only)

        • OptionalbypassPermissions?: boolean

          Bypass source permission checks (admin only)

        • OptionalexactMatch?: boolean

          Return only exact phrase matches (case-insensitive). Default: false

      Returns AsyncGenerator<SearchChunk>

      SearchChunk objects containing:

      • results - Array of search results in this chunk
      • done - Boolean indicating if this is the final chunk

      401 - Not authenticated

      422 - Validation error

      import { searchDocumentsStream } from "@cpod/sdk";

      let totalProcessed = 0;

      for await (const chunk of searchDocumentsStream({
        query: "annual report",
        sources: ["financial"],
      })) {
        console.log(`Received ${chunk.results.length} results`);
        totalProcessed += chunk.results.length;

        // Process each result
        for (const doc of chunk.results) {
          await processDocument(doc);
        }

        if (chunk.done) {
          console.log(`Finished! Total: ${totalProcessed}`);
          break;
        }
      }

    Settings

    Member Visibility