Files
clawmates/crates/tc-runtime/tests/run_loop.rs
T
Omar SobhandClaude Fable 5 a8efada690 P5 exit: usage metering, credit billing, promo codes, 3-step wizard
- LlmEvent::Usage across all three providers (Scripted deterministic
  word-count accounting; Anthropic message_start/delta usage; OpenAI-compat
  stream_options include_usage)
- tc-billing: ceil(tokens/1000) min 1 credit; lots drain oldest-first under
  FOR UPDATE; balance clamps at zero while the usage ledger records the
  full obligation; promo codes redeem exactly once via CAS (migration 0006)
- Runtime charges every completed run (billing failure never fails a run);
  proven: 1 token in + 3 out -> 1 credit deducted
- API: GET /api/team/usage, POST /api/credits/redeem (409 on reuse, audited)
- Credits page: balance, 7-day usage meter with runway estimate, PromoRedeem
- /claws/new is the full §9 wizard: ?step=identity|access|slack deep-linked
  progress, accent swatches + name randomizer, access toggles, optional
  Slack step, explicit review-and-confirm (creation = live agent), animated
  provisioning state -> straight into chat
- E2E: chat decrements the visible balance and fills the usage meter;
  WELCOME500 adds exactly 500 once then refuses; wizard round trip

140 Rust + 63 frontend tests + 23 Playwright journeys.

Co-Authored-By: Claude Fable 5 <[email protected]>
2026-06-10 07:19:34 -05:00

276 lines
9.0 KiB
Rust

use std::sync::Arc;
use tc_domain::{
AccessPolicy, Agent, AgentId, AgentStatus, MessageRole, Role, RunState, User, UserId,
Workspace, WorkspaceId,
};
use tc_llm::ScriptedProvider;
use tc_runtime::{RunEventBody, Runtime, RuntimeConfig};
const SCENARIOS: &str = r#"
[[scenario]]
marker = "[[scenario:tool-time]]"
[[scenario.turns]]
events = [
{ type = "text", text = "Let me check the clock." },
{ type = "tool_use", name = "clock.now", input = {} },
]
[[scenario.turns]]
events = [
{ type = "text", text = " It is now known." },
]
"#;
async fn seeded(pool: &sqlx::PgPool) -> Agent {
let ws = Workspace {
id: WorkspaceId::new(),
name: "Acme".into(),
plan: "team".into(),
};
tc_db::repo::workspaces::insert(pool, &ws).await.unwrap();
let owner = User {
id: UserId::new(),
workspace_id: ws.id,
email: format!("{}@acme.test", UserId::new()),
role: Role::Owner,
display_name: "Owner".into(),
created_at: time::OffsetDateTime::UNIX_EPOCH,
};
tc_db::repo::users::insert(pool, &owner).await.unwrap();
let agent = Agent {
id: AgentId::new(),
workspace_id: ws.id,
name: "Scout".into(),
job_title: "Analyst".into(),
system_prompt: "You are concise.".into(),
avatar: String::new(),
accent: String::new(),
wallpaper: String::new(),
managed_by: owner.id,
status: AgentStatus::Online,
};
tc_db::repo::agents::insert(pool, &agent, &AccessPolicy::default())
.await
.unwrap();
agent
}
fn runtime(pool: sqlx::PgPool) -> Runtime {
Runtime::new(
pool,
Arc::new(ScriptedProvider::from_toml(SCENARIOS).unwrap()),
RuntimeConfig::basic("scripted", 1024),
)
}
async fn drain(
mut rx: tokio::sync::broadcast::Receiver<tc_runtime::RunEventEnvelope>,
) -> Vec<tc_runtime::RunEventEnvelope> {
let mut events = Vec::new();
while let Ok(envelope) = rx.recv().await {
let done = matches!(
envelope.event,
RunEventBody::RunCompleted { .. } | RunEventBody::Error { .. }
);
events.push(envelope);
if done {
break;
}
}
events
}
#[tokio::test]
async fn plain_message_streams_and_persists_a_reply() {
let pool = tc_testkit::test_pool().await;
let agent = seeded(&pool).await;
let session = tc_db::repo::sessions::create(&pool, agent.id, agent.workspace_id, "Chat")
.await
.unwrap();
let rt = runtime(pool.clone());
let started = rt.send_message(session.id, "hello there").await.unwrap();
let events = drain(started.events).await;
// Envelope sequence is strictly increasing from 1.
let seqs: Vec<i64> = events.iter().map(|e| e.seq).collect();
assert_eq!(seqs, (1..=seqs.len() as i64).collect::<Vec<_>>());
assert!(matches!(events[0].event, RunEventBody::RunStarted { .. }));
let text: String = events
.iter()
.filter_map(|e| match &e.event {
RunEventBody::TextDelta { delta } => Some(delta.as_str()),
_ => None,
})
.collect();
assert_eq!(text, "I received: hello there");
// Transcript persisted: user message + agent reply with that text.
let history = tc_db::repo::messages::history(&pool, session.id)
.await
.unwrap();
assert_eq!(history.len(), 2);
assert_eq!(history[0].message.role, MessageRole::User);
assert_eq!(history[1].message.role, MessageRole::Agent);
assert_eq!(
history[1].message.content["text"],
"I received: hello there"
);
let run = tc_db::repo::runs::get(&pool, started.run_id).await.unwrap();
assert_eq!(run.state, RunState::Completed);
}
#[tokio::test]
async fn tool_scenario_executes_real_tool_and_records_steps() {
let pool = tc_testkit::test_pool().await;
let agent = seeded(&pool).await;
let session = tc_db::repo::sessions::create(&pool, agent.id, agent.workspace_id, "Chat")
.await
.unwrap();
let rt = runtime(pool.clone());
let started = rt
.send_message(session.id, "time? [[scenario:tool-time]]")
.await
.unwrap();
let events = drain(started.events).await;
let step_started = events.iter().find_map(|e| match &e.event {
RunEventBody::StepStarted { tool, .. } => Some(tool.clone()),
_ => None,
});
assert_eq!(step_started.as_deref(), Some("clock.now"));
let step_finished = events.iter().find_map(|e| match &e.event {
RunEventBody::StepFinished { status, output, .. } => Some((status.clone(), output.clone())),
_ => None,
});
let (status, output) = step_finished.expect("step finished event");
assert_eq!(status, "ok");
assert!(output["now"].is_string(), "clock output: {output}");
// Both text legs land in one agent message.
let history = tc_db::repo::messages::history(&pool, session.id)
.await
.unwrap();
let reply = &history[1];
let text = reply.message.content["text"].as_str().unwrap();
assert!(text.contains("Let me check the clock."), "got: {text}");
assert!(text.contains("It is now known."), "got: {text}");
// The step trace is persisted on the reply (the "N steps" UI source).
assert_eq!(reply.steps.len(), 1);
assert_eq!(reply.steps[0].tool_name.as_deref(), Some("clock.now"));
assert_eq!(reply.steps[0].status, tc_domain::StepStatus::Ok);
}
#[tokio::test]
async fn journal_matches_what_subscribers_saw() {
let pool = tc_testkit::test_pool().await;
let agent = seeded(&pool).await;
let session = tc_db::repo::sessions::create(&pool, agent.id, agent.workspace_id, "Chat")
.await
.unwrap();
let rt = runtime(pool.clone());
let started = rt.send_message(session.id, "ping").await.unwrap();
let live = drain(started.events).await;
// Replay from the journal reproduces the exact live sequence (the
// property reconnect/resumeFrom depends on).
let journal = tc_db::repo::run_events::list_after(&pool, started.run_id, 0)
.await
.unwrap();
assert_eq!(journal.len(), live.len());
for (persisted, observed) in journal.iter().zip(live.iter()) {
assert_eq!(persisted.seq, observed.seq);
let live_json = serde_json::to_value(&observed.event).unwrap();
assert_eq!(persisted.event_type, live_json["type"].as_str().unwrap());
}
}
#[tokio::test]
async fn unknown_tool_fails_the_step_but_not_the_run() {
let pool = tc_testkit::test_pool().await;
let agent = seeded(&pool).await;
let session = tc_db::repo::sessions::create(&pool, agent.id, agent.workspace_id, "Chat")
.await
.unwrap();
let scenarios = r#"
[[scenario]]
marker = "[[scenario:bad-tool]]"
[[scenario.turns]]
events = [ { type = "tool_use", name = "no.such.tool", input = {} } ]
[[scenario.turns]]
events = [ { type = "text", text = "I could not use that tool." } ]
"#;
let rt = Runtime::new(
pool.clone(),
Arc::new(ScriptedProvider::from_toml(scenarios).unwrap()),
RuntimeConfig::basic("scripted", 1024),
);
let started = rt
.send_message(session.id, "x [[scenario:bad-tool]]")
.await
.unwrap();
let events = drain(started.events).await;
let failed_step = events.iter().any(
|e| matches!(&e.event, RunEventBody::StepFinished { status, .. } if status == "error"),
);
assert!(failed_step, "expected an errored step");
// The error went back to the model as a tool result and it continued.
let run = tc_db::repo::runs::get(&pool, started.run_id).await.unwrap();
assert_eq!(run.state, RunState::Completed);
}
#[tokio::test]
async fn completed_runs_are_metered_and_decrement_credits() {
let pool = tc_testkit::test_pool().await;
let agent = seeded(&pool).await;
tc_db::repo::credits::add_lot(&pool, agent.workspace_id, 10, "seed")
.await
.unwrap();
let rt = runtime(pool.clone());
let session = tc_db::repo::sessions::create(&pool, agent.id, agent.workspace_id, "Bill")
.await
.unwrap();
let started = rt.send_message(session.id, "ping").await.unwrap();
drain(started.events).await;
// Scripted accounting: 1 word in, "I received: ping" = 3 words out →
// 4 tokens → 1 credit.
let mut metered = false;
for _ in 0..100 {
let row = sqlx::query!(
r#"SELECT tokens_in, tokens_out, credits FROM usage_events
WHERE workspace_id = $1"#,
agent.workspace_id.as_uuid(),
)
.fetch_optional(&pool)
.await
.unwrap();
if let Some(usage) = row {
assert_eq!(usage.tokens_in, 1);
assert_eq!(usage.tokens_out, 3);
assert_eq!(
tc_db::repo::credits::balance(&pool, agent.workspace_id)
.await
.unwrap(),
9
);
metered = true;
break;
}
tokio::time::sleep(std::time::Duration::from_millis(50)).await;
}
assert!(metered, "usage event never recorded");
}