/salinic

Modular search which features a unified API that allows you to plug in different search backends

Primary LanguagePythonApache License 2.0Apache-2.0

Tests

Salinic

Salinic - provides modular search. It features a unified API that allows you to plug in different search backends. Currently, it supports Xapian and Solr backends.

Usage

Declare your search schema:

from typing import Optional
from typing_extensions import Annotated

from salinic.field import IdField, KeywordField, TextField
from salinic.schema import Schema


class MyModel(Schema):
    """Index schema"""
    id: Annotated[str, IdField(primary_key=True)]
    user_id: str
    parent_id: str
    title: Annotated[str, TextField()]
    text: Annotated[Optional[str], TextField()] = None
    tags: Annotated[Optional[list[str]], KeywordField()] = []

Index your documents:

    from salinic import IndexRW, create_engine

    engine = create_engine("xapian:////search_index")
    index = IndexRW(engine, schema=MyModel)

    for document in all_your_documents():
        model = MyModel(
            id=str(document.id),
            user_id=str(document.user_id),
            parent_id=document.parent_id,
            title=document.title,
            text=document.text,
            tags=document.tags
        )
        index.add(model)

Search your documents:

    from salinic import IndexRO, create_engine

    engine = create_engine("xapian:////search_index")
    index = IndexRO(engine, schema=MyModel)

    sq = Search(MyModel).query(" your query string ")

    for found in index.search(sq):
        print(found)  # found is instance of MyModel

The only modification of your for changing to different search engine backend, is the first argument of the create_engine method e.g. from "xapian:////search_index" to "solr://localhost:8983/my-index-name".