# entid
A Rust library for generating and validating type-safe, prefixed entity identifiers based on UUIDs and ULIDs.
[](https://crates.io/crates/entid)
[](https://docs.rs/entid)
[](https://opensource.org/licenses/MIT)
## Features
- **Type-safe entity IDs**: Create distinct ID types for different entities
- **Multiple identifier formats**: Support for both UUID and ULID
- **Prefix support**: Automatically add entity-specific prefixes to IDs
- **Performance optimized**: Thread-safe caching of string representations
- **Serde compatible**: Seamless serialization and deserialization
- **Comprehensive error handling**: Clear error types for all operations
- **Zero-cost abstractions**: Minimal runtime overhead
- **Derive macro for implementing the `Prefix` trait**: Optional
## Installation
Add this to your `Cargo.toml`:
```toml
[dependencies]
entid = "0.3.0"
```
To use the derive macro for implementing the `Prefix` trait, enable the `derive` feature:
```toml
[dependencies]
entid = { version = "0.3.0", features = ["derive"] }
```
## Usage
### Basic Example with UUID
```rust
use entid::{EntityId, Prefix, UuidIdentifier, UuidEntityId};
// Define your entity types with custom prefixes
struct User;
impl Prefix for User {
fn prefix() -> &'static str {
"user"
}
fn delimiter() -> &'static str {
"_"
}
}
struct Post;
impl Prefix for Post {
fn prefix() -> &'static str {
"post"
}
// Optional: Override the default delimiter
fn delimiter() -> &'static str {
"-"
}
}
fn main() {
// Generate random IDs with UUID
let user_id = UuidEntityId::<User>::generate();
let post_id = EntityId::<Post, UuidIdentifier>::generate();
// Print the IDs
println!("User ID: {}", user_id); // e.g., "user_6ba7b810-9dad-11d1-80b4-00c04fd430c8"
println!("Post ID: {}", post_id); // e.g., "post-123e4567-e89b-12d3-a456-426614174000"
// Parse existing IDs
let parsed_user_id = UuidEntityId::<User>::new("user_6ba7b810-9dad-11d1-80b4-00c04fd430c8").unwrap();
// Type safety prevents mixing different entity IDs
// This won't compile:
// let wrong: UuidEntityId<Post> = user_id;
}
```
### Using the Derive Macro
With the `derive` feature enabled, you can use the derive macro to implement the `Prefix` trait:
```rust
use entid::{Prefix, UuidEntityId, UlidEntityId};
#[derive(Prefix)]
#[entid(prefix = "user", delimiter = "_")]
struct User;
#[derive(Prefix)]
#[entid(prefix = "post", delimiter = "-")]
struct Post;
// The delimiter is optional and defaults to "_"
#[derive(Prefix)]
#[entid(prefix = "comment")]
struct Comment;
fn main() {
let user_id = UuidEntityId::<User>::generate();
println!("User ID: {}", user_id); // e.g., "user_6ba7b810-9dad-11d1-80b4-00c04fd430c8"
let post_id = UlidEntityId::<Post>::generate();
println!("Post ID: {}", post_id); // e.g., "post-01H1VECZJYJ1QV2V0D0000JJDX"
let comment_id = UuidEntityId::<Comment>::generate();
println!("Comment ID: {}", comment_id); // e.g., "comment_6ba7b810-9dad-11d1-80b4-00c04fd430c8"
}
```
### Using ULID Instead of UUID
```rust
use entid::{EntityId, Prefix, UlidIdentifier, UlidEntityId};
struct Product;
impl Prefix for Product {
fn prefix() -> &'static str {
"prod"
}
}
fn main() {
// Generate a ULID-based ID
let product_id = UlidEntityId::<Product>::generate();
// ULIDs are lexicographically sortable by creation time
let product_ids: Vec<UlidEntityId<Product>> = (0..10)
.map(|_| UlidEntityId::<Product>::generate())
.collect();
// Sorting will order by creation time
let mut sorted_ids = product_ids.clone();
sorted_ids.sort();
// Get the timestamp from a ULID (not available with UUID)
if let Some(timestamp_ms) = product_id.timestamp_ms() {
println!("Product ID created at: {} ms since epoch", timestamp_ms);
}
}
```
### Using Deterministic UUIDs (v5)
```rust
use entid::{EntityId, Prefix, UuidIdentifier, Uuid};
struct ApiKey;
impl Prefix for ApiKey {
fn prefix() -> &'static str {
"key"
}
}
fn main() {
// Create a namespace for your application
let namespace = Uuid::parse_str("6ba7b810-9dad-11d1-80b4-00c04fd430c8").unwrap();
// Create a UUID v5 identifier
let uuid_id = UuidIdentifier::new_v5(&namespace, "user@example.com");
// Create an entity ID from the identifier
let api_key = EntityId::<ApiKey, UuidIdentifier>::from_identifier(uuid_id);
// Same input produces the same ID
let uuid_id2 = UuidIdentifier::new_v5(&namespace, "user@example.com");
let api_key2 = EntityId::<ApiKey, UuidIdentifier>::from_identifier(uuid_id2);
assert_eq!(api_key, api_key2);
}
```
### Error Handling
```rust
use entid::{EntityId, EntityIdError, IdentifierError, Prefix, UuidIdentifier};
struct User;
impl Prefix for User {
fn prefix() -> &'static str {
"user"
}
}
fn parse_id(input: &str) -> Result<(), Box<dyn std::error::Error>> {
// Parse an entity ID string
match EntityId::<User, UuidIdentifier>::new(input) {
Ok(id) => {
println!("Successfully parsed ID: {}", id);
Ok(())
},
Err(EntityIdError::InvalidFormat) => {
// Handle invalid format (missing prefix or delimiter)
println!("Invalid ID format: {}", input);
Err(Box::new(EntityIdError::InvalidFormat))
},
Err(EntityIdError::InvalidIdentifier) => {
// Handle invalid identifier (not a valid UUID/ULID)
println!("Invalid identifier part in ID: {}", input);
Err(Box::new(EntityIdError::InvalidIdentifier))
}
}
}
// Parse a raw identifier string
fn parse_raw_identifier(input: &str) -> Result<UuidIdentifier, IdentifierError> {
UuidIdentifier::parse(input)
}
```
### Using with Serde
```rust
use entid::{EntityId, Prefix, UlidIdentifier};
use serde::{Serialize, Deserialize};
struct Order;
impl Prefix for Order {
fn prefix() -> &'static str {
"order"
}
}
#[derive(Serialize, Deserialize)]
struct OrderRecord {
id: EntityId<Order, UlidIdentifier>,
customer_name: String,
amount: f64,
}
fn main() {
let order = OrderRecord {
id: EntityId::<Order, UlidIdentifier>::generate(),
customer_name: "John Doe".to_string(),
amount: 123.45,
};
// Serialize to JSON
let json = serde_json::to_string(&order).unwrap();
println!("JSON: {}", json);
// Deserialize from JSON
let deserialized: OrderRecord = serde_json::from_str(&json).unwrap();
assert_eq!(order.id, deserialized.id);
}
```
### Using with Databases
```rust
use entid::{EntityId, Prefix, UuidIdentifier};
struct Customer;
impl Prefix for Customer {
fn prefix() -> &'static str {
"cust"
}
}
// Example with a hypothetical database library
fn store_in_db(customer_id: &EntityId<Customer, UuidIdentifier>, name: &str) {
// The ID will be stored as a string like "cust_123e4567-e89b-12d3-a456-426614174000"
let id_str = customer_id.as_str();
// You can also access the raw identifier if needed
let uuid = customer_id.identifier().uuid();
// Database operations...
}
fn retrieve_from_db(id_str: &str) -> Result<EntityId<Customer, UuidIdentifier>, entid::EntityIdError> {
// Parse the ID string back into an EntityId
EntityId::<Customer, UuidIdentifier>::new(id_str)
}
```
## Advanced Usage
### Creating Monotonic ULIDs
```rust
use entid::{EntityId, Prefix, UlidIdentifier};
struct Task;
impl Prefix for Task {
fn prefix() -> &'static str {
"task"
}
}
fn main() {
// Create a ULID-based entity ID
let task1 = EntityId::<Task, UlidIdentifier>::generate();
// Create a monotonic ULID (ensures ordering even within the same millisecond)
let ulid2 = UlidIdentifier::monotonic_from(Some(task1.identifier()));
let task2 = EntityId::<Task, UlidIdentifier>::from_identifier(ulid2);
// task2 is guaranteed to sort after task1
assert!(task2 > task1);
}
```
### Custom Validation
```rust
use entid::{EntityId, Prefix, UuidIdentifier};
struct ApiKey;
impl Prefix for ApiKey {
fn prefix() -> &'static str {
"key"
}
}
// Extend EntityId with custom validation logic
impl EntityId<ApiKey, UuidIdentifier> {
pub fn is_valid_for_environment(&self, env: &str) -> bool {
// Custom validation logic based on the UUID version
match env {
"production" => self.identifier().version() == Some(uuid::Version::Sha1),
_ => true,
}
}
}
```
## Choosing Between UUID and ULID
### UUID Advantages
- Industry standard with wide adoption
- Multiple versions for different use cases (v1, v3, v4, v5)
- Well-supported in databases and other systems
### ULID Advantages
- Lexicographically sortable (sorts by creation time)
- URL-safe (no special characters)
- Shorter string representation (26 characters vs 36 for UUID)
- Built-in timestamp component
## Performance Considerations
- String representations are cached using `OnceLock` for thread-safe lazy initialization
- The `EntityId` type implements `Hash`, `PartialEq`, and `Eq` for efficient use in collections
- Memory usage is optimized by using `PhantomData` for type parameters
## License
This project is licensed under the MIT License - see the LICENSE file for details.