Simplify column names in generated structs for single-table queries

[tunamaguro]

Problem

Currently, generated row structs always include table prefixes in column names, even for single-table queries. This makes the generated code unnecessarily verbose.

Current behavior

For a simple query like:

SELECT id, name, bio FROM authors WHERE id = $1

The generated struct is:

pub struct GetAuthorRow {
    pub authors_id: i64,
    pub authors_name: String, 
    pub authors_bio: Option<String>,
}

Expected behavior

For single-table queries, column names should be simplified:

pub struct GetAuthorRow {
    pub id: i64,
    pub name: String,
    pub bio: Option<String>,
}

For multi-table queries (JOINs), table prefixes should be preserved to avoid naming conflicts.

Implementation plan

1. Detect number of tables used in each query
2. For single-table queries: remove table prefix from column names
3. For multi-table queries: preserve current behavior with table prefixes
4. Update column name generation logic in src/query.rs

Benefits

• Cleaner, more readable generated code
• Reduced verbosity for common single-table operations
• Maintains compatibility for complex multi-table queries