File size: 7,331 Bytes
eb51191 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 | # Agent Guidelines
This document contains guidelines and best practices for AI agents working with this codebase.
## Error Management
- Use `anyhow::Result` for error handling in services and repositories.
- Create domain errors using `thiserror`.
- Never implement `From` for converting domain errors, manually convert them
## Writing Tests
- All tests should be written in three discrete steps:
```rust,ignore
use pretty_assertions::assert_eq; // Always use pretty assertions
fn test_foo() {
let setup = ...; // Instantiate a fixture or setup for the test
let actual = ...; // Execute the fixture to create an output
let expected = ...; // Define a hand written expected result
assert_eq!(actual, expected); // Assert that the actual result matches the expected result
}
```
- Use `pretty_assertions` for better error messages.
- Use fixtures to create test data.
- Use `assert_eq!` for equality checks.
- Use `assert!(...)` for boolean checks.
- Use unwraps in test functions and anyhow::Result in fixtures.
- Keep the boilerplate to a minimum.
- Use words like `fixture`, `actual` and `expected` in test functions.
- Fixtures should be generic and reusable.
- Test should always be written in the same file as the source code.
- Use `new`, Default and derive_setters::Setters to create `actual`, `expected` and specially `fixtures`. For example:
**Good:**
```rust,ignore
User::default().age(12).is_happy(true).name("John")
User::new("Job").age(12).is_happy()
User::test() // Special test constructor
```
**Bad:**
```rust,ignore
User {name: "John".to_string(), is_happy: true, age: 12}
User::with_name("Job") // Bad name, should stick to User::new() or User::test()
```
- Use `unwrap()` unless the error information is useful. Use `expect` instead of `panic!` when error message is useful. For example:
**Good:**
```rust,ignore
users.first().expect("List should not be empty")
```
**Bad:**
```rust,ignore
if let Some(user) = users.first() {
// ...
} else {
panic!("List should not be empty")
}
```
- Prefer using `assert_eq` on full objects instead of asserting each field:
**Good:**
```rust,ignore
assert_eq!(actual, expected);
```
**Bad:**
```rust,ignore
assert_eq!(actual.a, expected.a);
assert_eq!(actual.b, expected.b);
```
## Verification
Always verify changes by running tests and linting the codebase
1. Run crate specific tests to ensure they pass.
```
cargo insta test --accept
```
2. **Build Guidelines**:
- **NEVER** run `cargo build --release` unless absolutely necessary (e.g., performance testing, creating binaries for distribution)
- For verification, use `cargo check` (fastest), `cargo insta test`, or `cargo build` (debug mode)
- Release builds take significantly longer and are rarely needed for development verification
## Writing Domain Types
- Use `derive_setters` to derive setters and use the `strip_option` and the `into` attributes on the struct types.
## Documentation
- **Always** write Rust docs (`///`) for all public methods, functions, structs, enums, and traits.
- Document parameters with `# Arguments` and errors with `# Errors` sections when applicable.
- **Do not include code examples** - docs are for LLMs, not humans. Focus on clear, concise functionality descriptions.
## Refactoring
- If asked to fix failing tests, always confirm whether to update the implementation or the tests.
## Git Operations
- Safely assume git is pre-installed
- Safely assume github cli (gh) is pre-installed
- Always use `Co-Authored-By: ForgeCode <noreply@forgecode.dev>` for git commits and Github comments
## Service Implementation Guidelines
Services should follow clean architecture principles and maintain clear separation of concerns:
### Core Principles
- **No service-to-service dependencies**: Services should never depend on other services directly
- **Infrastructure dependency**: Services should depend only on infrastructure abstractions when needed
- **Single type parameter**: Services should take at most one generic type parameter for infrastructure
- **No trait objects**: Avoid `Box<dyn ...>` - use concrete types and generics instead
- **Constructor pattern**: Implement `new()` without type bounds - apply bounds only on methods that need them
- **Compose dependencies**: Use the `+` operator to combine multiple infrastructure traits into a single bound
- **Arc<T> for infrastructure**: Store infrastructure as `Arc<T>` for cheap cloning and shared ownership
- **Tuple struct pattern**: For simple services with single dependency, use tuple structs `struct Service<T>(Arc<T>)`
### Examples
#### Simple Service (No Infrastructure)
```rust,ignore
pub struct UserValidationService;
impl UserValidationService {
pub fn new() -> Self { ... }
pub fn validate_email(&self, email: &str) -> Result<()> {
// Validation logic here
...
}
pub fn validate_age(&self, age: u32) -> Result<()> {
// Age validation logic here
...
}
}
```
#### Service with Infrastructure Dependency
```rust,ignore
// Infrastructure trait (defined in infrastructure layer)
pub trait UserRepository {
fn find_by_email(&self, email: &str) -> Result<Option<User>>;
fn save(&self, user: &User) -> Result<()>;
}
// Service with single generic parameter using Arc
pub struct UserService<R> {
repository: Arc<R>,
}
impl<R> UserService<R> {
// Constructor without type bounds, takes Arc<R>
pub fn new(repository: Arc<R>) -> Self { ... }
}
impl<R: UserRepository> UserService<R> {
// Business logic methods have type bounds where needed
pub fn create_user(&self, email: &str, name: &str) -> Result<User> { ... }
pub fn find_user(&self, email: &str) -> Result<Option<User>> { ... }
}
```
#### Tuple Struct Pattern for Simple Services
```rust,ignore
// Infrastructure traits
pub trait FileReader {
async fn read_file(&self, path: &Path) -> Result<String>;
}
pub trait Environment {
fn max_file_size(&self) -> u64;
}
// Tuple struct for simple single dependency service
pub struct FileService<F>(Arc<F>);
impl<F> FileService<F> {
// Constructor without bounds
pub fn new(infra: Arc<F>) -> Self { ... }
}
impl<F: FileReader + Environment> FileService<F> {
// Business logic methods with composed trait bounds
pub async fn read_with_validation(&self, path: &Path) -> Result<String> { ... }
}
```
### Anti-patterns to Avoid
```rust,ignore
// BAD: Service depending on another service
pub struct BadUserService<R, E> {
repository: R,
email_service: E, // Don't do this!
}
// BAD: Using trait objects
pub struct BadUserService {
repository: Box<dyn UserRepository>, // Avoid Box<dyn>
}
// BAD: Multiple infrastructure dependencies with separate type parameters
pub struct BadUserService<R, C, L> {
repository: R,
cache: C,
logger: L, // Too many generic parameters - hard to use and test
}
impl<R: UserRepository, C: Cache, L: Logger> BadUserService<R, C, L> {
// BAD: Constructor with type bounds makes it hard to use
pub fn new(repository: R, cache: C, logger: L) -> Self { ... }
}
// BAD: Usage becomes cumbersome
let service = BadUserService::<PostgresRepo, RedisCache, FileLogger>::new(...);
```
|