mxaccessgw/clients/rust/RustClientDesign.md

# Rust Client Detailed Design

## Purpose

Provide an async Rust client crate for MXAccess Gateway, plus a test CLI and
unit tests. The Rust client should use `tonic` and `tokio`.

Follow the [Rust Style Guide](../../docs/style-guides/RustStyleGuide.md) for handwritten
code and the [Protobuf Style Guide](../../docs/style-guides/ProtobufStyleGuide.md) for
generated contract inputs.

## Crate Layout

Actual layout — the `mxgateway-client` library crate is the workspace root,
with the `mxgw` test CLI as a workspace member:

```text
clients/rust/                 # `mxgateway-client` library crate (workspace root)
  Cargo.toml
  build.rs
  src/
    lib.rs
    client.rs
    session.rs
    galaxy.rs
    options.rs
    auth.rs
    value.rs
    version.rs
    error.rs
    generated.rs
  crates/
    mxgw-cli/                 # `mxgw` test CLI (workspace member)
      Cargo.toml
      src/main.rs
  tests/
    client_behavior.rs
    proto_fixtures.rs
```

Dependencies:

- `tonic`
- `prost`
- `prost-types`
- `tokio`
- `tokio-stream`
- `thiserror`
- `clap`
- `serde`
- `serde_json`

## Library API

Suggested API:

```rust
pub struct GatewayClient { /* tonic channel + generated client */ }

pub struct ClientOptions {
    pub endpoint: String,
    pub api_key: String,
    pub plaintext: bool,
    pub ca_file: Option<PathBuf>,
    pub server_name_override: Option<String>,
    pub connect_timeout: Duration,
    pub call_timeout: Duration,
}

impl GatewayClient {
    pub async fn connect(options: ClientOptions) -> Result<Self, Error>;
    pub async fn open_session(&self, options: OpenSessionOptions) -> Result<Session, Error>;
    pub async fn invoke(&self, request: MxCommandRequest) -> Result<MxCommandReply, Error>;
}
```

Session:

```rust
pub struct Session {
    pub id: String,
}

impl Session {
    pub async fn register(&self, client_name: &str) -> Result<i32, Error>;
    pub async fn add_item(&self, server_handle: i32, item: &str) -> Result<i32, Error>;
    pub async fn add_item2(&self, server_handle: i32, item: &str, context: &str) -> Result<i32, Error>;
    pub async fn advise(&self, server_handle: i32, item_handle: i32) -> Result<(), Error>;
    pub async fn add_item_bulk(&self, server_handle: i32, tag_addresses: Vec<String>) -> Result<Vec<SubscribeResult>, Error>;
    pub async fn advise_item_bulk(&self, server_handle: i32, item_handles: Vec<i32>) -> Result<Vec<SubscribeResult>, Error>;
    pub async fn remove_item_bulk(&self, server_handle: i32, item_handles: Vec<i32>) -> Result<Vec<SubscribeResult>, Error>;
    pub async fn un_advise_item_bulk(&self, server_handle: i32, item_handles: Vec<i32>) -> Result<Vec<SubscribeResult>, Error>;
    pub async fn subscribe_bulk(&self, server_handle: i32, tag_addresses: Vec<String>) -> Result<Vec<SubscribeResult>, Error>;
    pub async fn unsubscribe_bulk(&self, server_handle: i32, item_handles: Vec<i32>) -> Result<Vec<SubscribeResult>, Error>;
    pub async fn write(&self, server_handle: i32, item_handle: i32, value: MxValue, user_id: i32) -> Result<(), Error>;
    pub async fn write_bulk(&self, server_handle: i32, entries: Vec<WriteBulkEntry>, user_id: i32) -> Result<Vec<BulkWriteResult>, Error>;
    pub async fn write2_bulk(&self, server_handle: i32, entries: Vec<Write2BulkEntry>, timestamp: prost_types::Timestamp, user_id: i32) -> Result<Vec<BulkWriteResult>, Error>;
    pub async fn write_secured_bulk(&self, server_handle: i32, entries: Vec<WriteSecuredBulkEntry>, current_user_id: i32, verifier_user_id: i32) -> Result<Vec<BulkWriteResult>, Error>;
    pub async fn write_secured2_bulk(&self, server_handle: i32, entries: Vec<WriteSecured2BulkEntry>, timestamp: prost_types::Timestamp, current_user_id: i32, verifier_user_id: i32) -> Result<Vec<BulkWriteResult>, Error>;
    pub async fn read_bulk(&self, server_handle: i32, tags: &[String], timeout_ms: u32) -> Result<Vec<ReadBulkResult>, Error>;
    pub async fn events(&self) -> Result<impl Stream<Item = Result<MxEvent, Error>>, Error>;
    pub async fn close(&self) -> Result<(), Error>;
}
```

The five bulk-write helpers (`write_bulk`, `write2_bulk`, `write_secured_bulk`,
`write_secured2_bulk`) and `read_bulk` mirror the worker's bulk command shapes
in `mxaccess_gateway.proto` and use the same correlation-id discipline as the
unary helpers — `session::next_correlation_id` is `pub` so that consumers
constructing raw `MxCommandRequest`/`CloseSessionRequest` payloads outside
the `Session` helpers (notably the `mxgw` test CLI's `ping` and
`close-session` subcommands) share the same id generation.

## Authentication

Use a `tonic` interceptor or request extension layer to add:

```text
authorization: Bearer <api key>
```

Use `SecretString` or equivalent if a dependency is acceptable. Always redact
API keys in `Debug` output.

## TLS

Support:

- plaintext channel for local development,
- native or rustls TLS depending on project preference,
- custom CA file,
- domain override.

## Streaming

Expose event streams as a `Stream<Item = Result<MxEvent, Error>>`. Dropping the
stream should cancel the underlying gRPC stream.

Do not buffer unboundedly in the client. If a helper channel is used, make it
bounded.

## Error Handling

Use `thiserror`:

```rust
pub enum Error {
    InvalidEndpoint { endpoint: String, detail: String },
    InvalidArgument { name: String, detail: String },
    Transport(tonic::transport::Error),
    Authentication { message: String, status: Box<tonic::Status> },
    Authorization { message: String, status: Box<tonic::Status> },
    Timeout { message: String, status: Box<tonic::Status> },
    Cancelled { message: String, status: Box<tonic::Status> },
    Unavailable { message: String, status: Box<tonic::Status> },
    Status(Box<tonic::Status>),
    Command(Box<CommandError>),
    ProtocolStatus { operation: &'static str, code: ProtocolStatusCode, message: String },
    MalformedReply { detail: String },
}
```

`Unavailable` classifies the transient `Code::Unavailable` /
`Code::ResourceExhausted` statuses so callers can decide whether to retry
without unwrapping the raw status. `MalformedReply` surfaces OK replies
whose payload does not carry the data the command contract requires (for
example, an `AddItem` reply missing the item handle, or a `WriteBulk` reply
carrying the wrong payload arm). `InvalidEndpoint` is returned when the
endpoint URL fails to parse or its TLS material cannot be loaded.

Preserve raw command replies in `CommandError` where applicable.

## Test CLI

Binary: `mxgw`.

Use `clap` derive.

Commands (see `clients/rust/README.md` for full argument lists):

```text
mxgw version
mxgw ping
mxgw open-session
mxgw close-session --session-id <id>
mxgw register --session-id <id> --client-name <name>
mxgw add-item --session-id <id> --server-handle <h> --item <tag>
mxgw advise --session-id <id> --server-handle <h> --item-handle <h>
mxgw subscribe-bulk --session-id <id> --server-handle <h> --items <a,b,c>
mxgw unsubscribe-bulk --session-id <id> --server-handle <h> --item-handles <1,2,3>
mxgw read-bulk --session-id <id> --server-handle <h> --items <a,b,c> --timeout-ms 1500
mxgw write --session-id <id> --server-handle 1 --item-handle 1 --value-type int32 --value 123
mxgw write2 --session-id <id> --server-handle 1 --item-handle 1 --value-type int32 --value 123 --timestamp <rfc3339>
mxgw write-bulk --session-id <id> --server-handle <h> --item-handles <1,2> --value-type int32 --values <1,2>
mxgw write2-bulk --session-id <id> --server-handle <h> --item-handles <1,2> --value-type int32 --values <1,2> --timestamp <rfc3339>
mxgw write-secured-bulk --session-id <id> --server-handle <h> --item-handles <1,2> --value-type int32 --values <1,2>
mxgw write-secured2-bulk --session-id <id> --server-handle <h> --item-handles <1,2> --value-type int32 --values <1,2> --timestamp <rfc3339>
mxgw stream-events --session-id <id> --json
mxgw bench-read-bulk --duration-seconds 30 --bulk-size 6 --json
mxgw smoke --endpoint http://localhost:5000 --api-key-env MXGATEWAY_API_KEY --item TestChildObject.TestInt
mxgw galaxy test-connection
mxgw galaxy last-deploy-time
mxgw galaxy discover-hierarchy
mxgw galaxy watch [--last-seen-deploy-time <rfc3339>] [--max-events N]
```

JSON output should use `serde_json`.

## Unit Tests

Use a fake `tonic` server started on a local ephemeral port, or abstract the
generated client behind a trait for unit tests.

Required tests:

- generated client compiles from proto,
- auth metadata injection,
- TLS/plaintext endpoint construction,
- value conversion,
- command request construction,
- error mapping from `tonic::Status`,
- event stream order,
- stream cancellation,
- CLI parsing,
- JSON redaction.

## Integration Tests

Skip unless:

```text
MXGATEWAY_INTEGRATION=1
```

Use `tokio::test`. Run bounded smoke flow and ensure `CloseSession` is attempted
with `drop` fallback docs, but do not rely on `Drop` for async close.

## Related Documentation

- [Client Libraries Detailed Design](../../docs/ClientLibrariesDesign.md)
- [Client Proto Generation](../../docs/ClientProtoGeneration.md)
- [Client Packaging](../../docs/ClientPackaging.md)
- [Rust Style Guide](../../docs/style-guides/RustStyleGuide.md)