Refactor duplicate IndexOutput struct

[leaf-potato]

What type of enhancement is this?

Refactor

What does the enhancement do?

When I read the relevant code of the Mito engine, I found that two identical structures are defined:

greptimedb/src/mito2/src/sst/index.rs

Lines 68 to 75 in 975b8c6

	pub struct FulltextIndexOutput {
	/// Size of the index.
	pub index_size: ByteCount,
	/// Number of rows in the index.
	pub row_count: RowCount,
	/// Available columns in the index.
	pub columns: Vec<ColumnId>,
	}

greptimedb/src/mito2/src/sst/index.rs

Lines 57 to 64 in 975b8c6

	pub struct InvertedIndexOutput {
	/// Size of the index.
	pub index_size: ByteCount,
	/// Number of rows in the index.
	pub row_count: RowCount,
	/// Available columns in the index.
	pub columns: Vec<ColumnId>,
	}

Perhaps defining a base structure is sufficient.

Implementation challenges

#[derive(Debug, Clone, Default)]
pub struct IndexBaseOutput {
    /// Size of the index.
    pub index_size: ByteCount,
    /// Number of rows in the index.
    pub row_count: RowCount,
    /// Available columns in the index.
    pub columns: Vec<ColumnId>,
}

pub type InvertedIndexOutput = IndexBaseOutput;
pub type FulltextIndexOutput = IndexBaseOutput;

1. Define IndexBaseOutput struct.
2. type IndexBaseOutput to InvertedIndexOutput and FulltextIndexOutput.

In the future, if InvertedIndexOutput or FulltextIndexOutput has non-public fields, it can be modified like this:

struct InvertedIndexOutput {
    base_output: IndexBaseOutput,
    .....
}

[evenyag]

@zhongzc What do you think?

[zhongzc]

I have no objection, you are free to change it this way. My original idea was that these two indexes have the opportunity to present different Output, and it would be more convenient to separate them in case fields are added later.

[leaf-potato]

I have no objection, you are free to change it this way.

Thanks, I'll submit a PR later.

[evenyag]

Closed by #4592