Skip to content

Search engine interface

Provide a Streamlit-based search interface for the Any-to-Any Terminology Mapper.

This module defines an interactive web application for querying a terminology retrieval system backed by ChromaDB. It exposes a simple search UI with multiple tabs, supports cached retriever loading, and displays ranked candidate concepts with key metadata and confidence scores.

The application is intended for local or internal use as a lightweight search front end for terminology exploration and validation.

cached_load_retriever(retriever_name='embeddinggemma-300M')

Load and cache a retriever instance for the Streamlit application.

This function wraps retriever loading with Streamlit's resource cache so that the same retriever is reused across reruns, avoiding repeated model or database initialization overhead.

Parameters:

Name Type Description Default
retriever_name str

Name of the retriever configuration to load from the retriever registry.

 'embeddinggemma-300M'

Returns:

Type Description
ChromaDBRetriever

A loaded ChromaDBRetriever instance corresponding to the requested retriever name.

Raises:

Type Description
KeyError

If the provided retriever name is not found in the registry.
Exception

Propagates errors raised during retriever initialization.
Source code in aatm\search_ui.py 
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
@st.cache_resource
def cached_load_retriever(
    retriever_name: str = "embeddinggemma-300M",
) -> "ChromaDBRetriever":
    """Load and cache a retriever instance for the Streamlit application.

    This function wraps retriever loading with Streamlit's resource cache so
    that the same retriever is reused across reruns, avoiding repeated model or
    database initialization overhead.

    Args:
        retriever_name: Name of the retriever configuration to load from the
            retriever registry.

    Returns:
        A loaded ``ChromaDBRetriever`` instance corresponding to the requested
            retriever name.

    Raises:
        KeyError: If the provided retriever name is not found in the registry.
        Exception: Propagates errors raised during retriever initialization.
    """
    retriever = load_retriever(retriever_name)
    return retriever