strands.storage.search.keyword

Keyword search strategy using token-overlap scoring.

tokenize

def tokenize(text: str) -> set[str]

Defined in: src/strands/storage/search/keyword.py:13

Lowercase and split text into a set of word tokens, dropping empties.

Splits on any run of non-word characters (Unicode-aware). Ensures cross-SDK compatibility with the TypeScript /[^\p\{L}\p\{N}_]+/u regex.

Arguments:

• text - The text to tokenize.

Returns:

A set of lowercased word tokens.

token_overlap_score

def token_overlap_score(query_tokens: set[str], content: str) -> int

Defined in: src/strands/storage/search/keyword.py:28

Lexical relevance score: distinct query tokens present in the content.

A higher count means more of the query’s words are present. Returns 0 when there is no overlap.

Arguments:

• query_tokens - Pre-tokenized query terms.
• content - The content string to score against.

Returns:

Number of distinct query tokens found in the content.

KeywordSearchStrategy

class KeywordSearchStrategy()

Defined in: src/strands/storage/search/keyword.py:44

Keyword search strategy using token-overlap scoring.

Tokenizes the query and each stored entry (key + content), then scores by the number of distinct query tokens that appear. Works on any storage backend with list() and read() — no index or embedding model required.

This is the default search strategy for all shipped storage backends.

Example:

from strands.storage.search import KeywordSearchStrategy

strategy = KeywordSearchStrategy()
results = await strategy.search(storage, "dark mode toggle")

search

async def search(storage: Storage, query: str,
                 **kwargs: Any) -> builtins.list[StorageSearchResult]

Defined in: src/strands/storage/search/keyword.py:62

Search content in storage by keyword token-overlap scoring.

Arguments:

• storage - The storage to search over.
• query - A natural-language string query.
• **kwargs - Unused; accepted for protocol compatibility.

Returns:

Matched keys with relevance scores, ranked best-first.