Skip to content

armgit / camel

Public Imported from GitHub
0

🐫 CAMEL: The first and the best multi-agent framework. Finding the Scaling Law of Agents. https://www.camel-ai.org

1.5 KB
Raw Blame
<a id="camel.retrievers.base_rerank"></a>

<a id="camel.retrievers.base_rerank.BaseRerankRetriever"></a>

## BaseRerankRetriever

```python
class BaseRerankRetriever(BaseRetriever):
```

Abstract base class for re-ranking retrievers.

A re-ranking retriever takes a query together with a list of already
retrieved results (for example, the output of a :class:`VectorRetriever`,
a :class:`BM25Retriever`, or a :class:`HybridRetriever`) and re-orders
those results by their semantic relevance to the query, typically using a
cross-encoder or a dedicated re-ranking API.

Unlike first-stage retrievers, re-rankers do not ingest content, so the
:meth:`process` method is not required and is left unimplemented.

<a id="camel.retrievers.base_rerank.BaseRerankRetriever.query"></a>

### query

```python
def query(
    self,
    query: str,
    retrieved_result: List[Dict[str, Any]],
    top_k: int = DEFAULT_TOP_K_RESULTS
):
```

Re-ranks the given retrieved results against the query.

**Parameters:**

- **query** (str): The query string used for re-ranking.
- **retrieved_result** (List[Dict[str, Any]]): The content to be re-ranked, typically the output of a first-stage retriever. Each item should contain a `__INLINE_CODE_0__` entry.
- **top_k** (int, optional): The number of top results to return after re-ranking. Must be a positive integer. (default: :obj:`DEFAULT_TOP_K_RESULTS`)

**Returns:**

  List[Dict[str, Any]]: The re-ranked results, each containing the
original data plus an updated `__INLINE_CODE_0__`.