Query Result

Query results are returned as an array of SearchResult objects:

export type SearchResult = {
    id: number;
    key: Record<string, unknown>;
    score: number;
    refs: Reference[];
};

Fields

• id

  The index position of the matching document in the original array. This allows you to retrieve the corresponding document from your data source.

• key

  An object containing the specified fields from the original document. For example, if ['slug', 'data.title'] were set as key_fields, the result would include { slug: "matching slug", data: { title: "matching title" } }. This provides context for the search result, allowing you to display relevant information to the user.

• score

  A TF-IDF-based relevance score, where higher scores indicate stronger matches. Scores decrease proportionally to edit distance in fuzzy searches. This helps in ranking the search results, allowing users to see the most relevant matches first.

• refs

  An array of Reference objects containing details on the match locations.

  • Structure:

    export type Reference = {
        token: string;
        path: Path;
        pos?: number;
        wordaround?: string;
        keyword_range?: [number, number];
        distance: number;
    };

Reference Fields

• token

  The specific search term that matched. It is useful for highlighting the matched term in the search results.

• path

  The field where the match was found. It helps in identifying which part of the document contained the match.

• pos

  The position of the match within the text (character index from the beginning of the field). For fuzzy matches, slight shifts may occur due to insertions/deletions. This can be used to highlight the exact location of the match within the text.

• wordaround

  A snippet of text surrounding the match for better context.

• keyword_range

  The keyword positions within wordaround. The first element indicates the start position of the keyword, and the second element indicates the end position of the keyword.

• distance

  The edit distance for fuzzy search matches. This ndicates how closely the search term matched the indexed content.

Example Usage

Here’s an example of how to handle query results:

const results: SearchResult[] = await search(index, "search term");
results.forEach(result => {
    console.log(`Matched Document ID: ${result.id}`);
    console.log(`Score: ${result.score}`);
    console.log(`Matched Key: ${JSON.stringify(result.key)}`);
    result.refs.forEach(ref => {
        console.log(`Matched Token: ${ref.token} at position ${ref.pos}`);
        const wa_pre = ref.wordaround.slice(0, ref.keyword_range[0]);
        const wa_kwd = ref.wordaround.slice(ref.keyword_range[0], ref.keyword_range[1]);
        const wa_post = ref.wordaround.slice(ref.keyword_range[1]);
        console.log(`Context: ${wa_pre}*${wa_kwd}*${wa_post}`);
    });
});

Error Handling

When processing query results, it is important to handle potential errors. If the search fails, a StaticSeekError will be returned. Ensure to check for this error and handle it appropriately in your application.