Function parse_with_schema 

pub fn parse_with_schema<T: DeserializeOwned + SchemaInfo>(
    input: &str,
) -> Result<T>

Parses an LLM response using schema information for guided parsing.

This function leverages compile-time schema information from the SchemaInfo trait to make better parsing decisions. It’s particularly useful for:

• Field name normalization (camelCase → snake_case)
• Better error messages with type context
• Schema-aware type coercion

Examples

use tryparse::parse_with_schema;
use tryparse::schema::SchemaInfo;
use serde::Deserialize;

#[derive(Deserialize)]
struct User {
    user_name: String,
    age: u32,
}

// Implement SchemaInfo manually (or use derive macro with feature flag)
impl SchemaInfo for User {
    fn schema() -> tryparse::schema::Schema {
        tryparse::schema::Schema::Object {
            name: "User".to_string(),
            fields: vec![
                tryparse::schema::Field::new("user_name", tryparse::schema::Schema::String)
                    .with_alias("userName")
                    .with_alias("UserName"),
                tryparse::schema::Field::new("age", tryparse::schema::Schema::Int),
            ],
        }
    }
}

// Parse with camelCase field names - schema guides normalization
let response = r#"{"userName": "Alice", "age": 30}"#;
let user: User = parse_with_schema(response).unwrap();
assert_eq!(user.user_name, "Alice");

Errors

Returns ParseError::NoCandidates if no valid JSON could be extracted. Returns ParseError::DeserializeFailed if deserialization fails for all candidates.