Description
Elasticsearch version: 8.12.0
elasticsearch-py version: 8.12.0
Description of the problem including expected versus actual behavior:
I am encountering a type mismatch error when attempting to use the Elasticsearch.search method with the fields parameter to retrieve a subset of stored fields from matched documents. According to the Elasticsearch documentation, passing a list of field names as strings to the fields parameter should be sufficient for this purpose Elasticsearch Fields Request.
However, when I specify the fields parameter as a list of strings, both Pylance and PyCharm raise type errors indicating that the expected type is t.Optional[t.Sequence[t.Mapping[str, t.Any]]], rather than t.Optional[t.Union[str, t.Sequence[str]]]. This discrepancy seems inconsistent with the type hints used in other methods like search_mvt or knn_search, where the fields parameter accepts a union of string or sequence of strings.
Despite the type checking errors reported by the IDEs, the actual execution of the search query does yield the expected results, with the returned hits including the requested fields. It appears that the type hints may need updating to reflect the correct usage of the fields parameter for the search method.
Steps to reproduce:
from elasticsearch import Elasticsearch
es_instance = Elasticsearch("http://localhost:9200")
query = {"match": {"title": {"query": "Framework Laptop"}}}
fields = ["title", "content"]
es_instance.search(index="products", query=query, fields=fields)In this snippet, fields is passed as a list of strings, but the type hint suggests that a sequence of mappings is expected. This leads to the type error in Pylance and PyCharm:
Could the type hint for the fields parameter in the search method be updated to align with the expected usage, or is there another way to address this issue without changing the type hint? Am I misunderstanding how to use the fields parameter?