dGraph async Rust client which communicates with the server using gRPC build with Tonic.
Before using this client, it is highly recommended to go through [tour.dgraph.io] and [docs.dgraph.io] to understand how to run and work with Dgraph.
dgraph-tonic is available on crates.io. Add the following dependency to your Cargo.toml.
toml
[dependencies]
dgraph-tonic = "0.2"
Depending on the version of Dgraph that you are connecting to, you will have to use a different feature of this client (dgraph-1-0 is default version).
| Dgraph version | feature | |:--------------:|:-----------------:| | 1.0.X | dgraph-1-0 | | 1.1.X | dgraph-1-1 | | 1.2.X | dgraph-1-1 | | 20.03.0 | NOT SUPPORTED |
Note: Only API breakage from *dgraph-1-0 to dgraph-1-1 is in the function MutatedTxn.mutate(). This function returns a Assigned type in dgraph-1-0 but a Response type in dgraph-1-1.
Client object can be initialised by passing a vector of tonic::transport::Endpoints. Connecting to multiple dGraph servers in the same cluster allows for better distribution of workload.
The following code snippet shows it with just one endpoint.
rust
let client = Client::new(vec!["http://127.0.0.1:19080"]).await?;
Alternatively, secure client can be used:
rust
let client = Client::new_with_tls_client_auth(
vec!["https://dgraph.io"],
"/path/ca.crt",
"/path/client.crt",
"/path/client.key",
).await?;
All certs must be in PEM format.
To set the schema, create an instance of dgraph::Operation and use the Alter endpoint.
rust
let op = Operation {
schema: "name: string @index(exact) .".into(),
..Default::default()
};
let response = client.alter(op).await?;
Operation contains other fields as well, including DropAttr and DropAll. DropAll is useful if you wish to discard all the data, and start from a clean slate, without bringing the instance down. DropAttr is used to drop all the data related to a predicate.
Transaction is modeled with The Typestate Pattern in Rust. The typestate pattern is an API design pattern that encodes information about an object's run-time state in its compile-time type. This principle allows us to identify some type of errors, like mutation in read only transaction, during compilation. Transaction types are:
query and query_with_vars actions.query and query_with_vars actions.query and query_with_vars actions.Client provides several factory methods for transactions. Create new transaction incurs no network overhead.
rust
let txn = client.new_txn();
let read_only = txn.read_only();
let best_effort = txn.best_effort();
let mutated = txn.mutated();
Only for Mutated transaction must be always called txn.dicard().await? or txn.commit().await? function before txn variable is dropped.
txn.mutate(mu).await? runs a mutation. It takes in a Mutation object. You can set the data using JSON or RDF N-Quad format. There exist helper functions for JSON format (mu.set_set_json(), mu.set_delete_json())
Example:
```rust
struct Person { uid: String, name: String, }
let p = Person { uid: "_:alice".into(), name: "Alice".into(), };
let mut mu = Mutation::new(); mu.setsetjson(&p)?;
let mut txn = client.newmutatedtxn(); let assigned = txn.mutate(mu).await.expect("failed to create data"); ```
Note: Only API breakage from dgraph-1-0 to dgraph-1-1 is in the function MutatedTxn.mutate(). This function returns a Assigned type in dgraph-1-0 but a Response type in dgraph-1-1.
Sometimes, you only want to commit a mutation, without querying anything further. In such cases, you can use txn.mutate_and_commit_now(mu) to indicate that the mutation must be immediately committed. Txn object is being consumed in this case.
Mutation::with_ignored_index_conflict() can be applied on a Mutation object to not run conflict detection over the index, which would decrease the number of transaction conflicts and aborts. However, this would come at the cost of potentially inconsistent upsert operations. This flag is avaliable only in dgraph-1-0.
You can run a query by calling txn.query(q). You will need to pass in a GraphQL+- query string. If you want to pass an additional map of any variables that you might want to set in the query, call txn.query_with_vars(q, vars) with the variables map as second argument.
Let's run the following query with a variable \$a:
console
query all($a: string) {
all(func: eq(name, $a))
{
uid
name
}
}
Response provides function try_into() which can be used for transforming returned JSON into coresponding struct object which implements serde Deserialize traits.
```rust
struct Persons {
all: Vec
let q = r#"query all($a: string) { all(func: eq(name, $a)) { uid name } }"#;
let mut vars = HashMap::new(); vars.insert("$a", "Alice");
let resp: Response = client.newreadonlytxn().querywithvars(q, vars).await.expect("query"); let persons: Persons = resp.tryinto().except("Persons"); println!("Persons: {:?}", persons); ```
When running a schema query, the schema response is found in the Schema field of Response.
```rust let q = r#"schema(pred: [name]) { type index reverse tokenizer list count upsert lang }"#.to_string();
let resp = client.newreadonly_txn().query(q).await?; println!("{:#?}", resp.schema); ```
Avaibale since dgraph-1-1.
The txn.upsert(query, mutation) function allows you to run upserts consisting of one query and one or more mutations. Query variables could be defined with txn.upsert_with_vars(query, vars, mutation). Transaction in upsert is commited.
To know more about upsert, we highly recommend going through the docs at https://docs.dgraph.io/mutations/#upsert-block.
```rust let q = r#" query { user as var(func: eq(email, "wrong_email@dgraph.io")) }"#;
let mut mu = Mutation::new();
mu.setsetnquads(r#"uid(user)
let txn = client.newmutatedtxn(); // Upsert: If wrong_email found, update the existing data or else perform a new mutation. let response = txn.upsert(q, mu).await.expect("failed to upsert data");
```
You can upsert with one mutation or vector of mutations.
Avaibale since dgraph-1-1.
The upsert block allows specifying a conditional mutation block using an @if directive. The mutation is executed only when the specified condition is true. If the condition is false, the mutation is silently ignored.
See more about Conditional Upsert Here.
```rust let q = r#" query { user as var(func: eq(email, "wrong_email@dgraph.io")) }"#;
let mut mu = Mutation::new();
mu.setsetnquads(r#"uid(user)
let txn = client.newmutatedtxn(); let response = txn.upsert(q, vec![mu]).await.expect("failed to upsert data"); ```
A mutated transaction can be committed using the txn.commit() method. If your transaction consisted solely of calls to txn.query or txn.query_with_vars, and no calls to txn.mutate, then calling txn.commit is not necessary.
An error will be returned if other transactions running concurrently modify the same data that was modified in this transaction. It is up to the user to retry transactions when they fail.
```rust let mut txn = client.newmutatedtxn(); // Perform some queries and mutations.
let res = txn.commit().await; if res.is_err() { // Retry or handle error } ```
Tests require Dgraph running on localhost:19080. For the convenience there are two docker-compose.yaml files, depending on Dgraph you are testing against, prepared in the root directory:
bash
docker-compose -f docker-compose-1-X.yaml up -d
Since we are working with a database, tests also need to be run in a single thread to prevent aborts. Feature flags are used depending on version of Dgraph you are using. Eg.:
bash
cargo test --no-default-features --features dgraph-1-1 -- --test-threads=1
Contributions are welcome. Feel free to raise an issue, for feature requests, bug fixes and improvements.
These have to be done with both Dgraph 1.0 and Dgraph 1.1+:
Update the version and publish crate:
git tag v0.X.Xgit push origin v0.X.Xcargo publish