Files
optimclaw/tests/workspace_integration.rs
T
30790439ee perf: build system prompt once per turn, skip tools on force-text (#583)
* perf: build system prompt once per turn, skip tools on force-text, fix nudge role (#565)

Three fixes to agentic loop prompt handling:

1. Build system prompt once per turn instead of every tool iteration.
   `build_system_prompt_with_tools` is now pub; callers pass the result
   via `ReasoningContext::system_prompt` to avoid rebuilding ~1,500 tokens
   per iteration.

2. Skip `## Available Tools` section when `force_text = true`. The
   dispatcher passes a no-tools prompt variant on the final iteration,
   saving ~460 tokens and removing misleading instructions.

3. Change nudge message from `Role::System` to `Role::User`. A second
   system message mid-conversation is unsupported by most providers.

Co-Authored-By: Claude Opus 4.6 (1M context) <[email protected]>

* fix: revert nudge role change to keep ChatMessage::system

Copilot review correctly identified that using Role::User for the nudge
breaks compact_messages_for_retry, which uses rposition for Role::User
to find the last real user message. Role::Assistant would cause
back-to-back assistant messages. Since no production issues were reported
with the original system role, revert to ChatMessage::system.

[skip-regression-check]

Co-Authored-By: Claude Opus 4.6 (1M context) <[email protected]>

* fix: address PR review — omit tool guidance when tools empty, rename shadowed var

- Conditionalize "Call tools…" guidelines and "## Tool Call Style" section
  in the system prompt so they are only included when tools are non-empty.
  Previously the force-text (no-tools) prompt still contained misleading
  tool-calling instructions. (Copilot review comment)

- Rename `system_prompt` → `cached_prompt` in dispatcher to avoid shadowing
  the earlier workspace identity `system_prompt` variable. (Copilot review)

- Add regression tests: `test_system_prompt_with_tools_contains_tool_guidance`
  and extended assertions in `test_system_prompt_without_tools_omits_tools_section`.

Co-Authored-By: Claude Opus 4.6 <[email protected]>

---------

Co-authored-by: Claude Opus 4.6 (1M context) <[email protected]>
Co-authored-by: [email protected] <[email protected]>
2026-03-07 09:15:00 +00:00

410 lines
11 KiB
Rust

#![cfg(feature = "postgres")]
//! Integration tests for the workspace module.
//!
//! Requires a running PostgreSQL with pgvector extension.
//! Set DATABASE_URL=postgres://localhost/ironclaw_test
use std::sync::Arc;
use ironclaw::workspace::{MockEmbeddings, SearchConfig, Workspace, paths};
fn get_pool() -> deadpool_postgres::Pool {
let database_url = std::env::var("DATABASE_URL")
.unwrap_or_else(|_| "postgres://localhost/ironclaw_test".to_string());
let config: tokio_postgres::Config = database_url.parse().expect("Invalid DATABASE_URL");
let mgr = deadpool_postgres::Manager::new(config, tokio_postgres::NoTls);
deadpool_postgres::Pool::builder(mgr)
.max_size(4)
.build()
.expect("Failed to create pool")
}
/// Try to get a connection, returning None if Postgres is unreachable.
/// Tests call this to skip gracefully in CI where no database is available.
async fn try_connect(pool: &deadpool_postgres::Pool) -> Option<()> {
match pool.get().await {
Ok(_) => Some(()),
Err(e) => {
eprintln!("skipping: database unavailable ({e})");
None
}
}
}
async fn cleanup_user(pool: &deadpool_postgres::Pool, user_id: &str) {
let conn = pool.get().await.expect("Failed to get connection");
conn.execute(
"DELETE FROM memory_documents WHERE user_id = $1",
&[&user_id],
)
.await
.ok();
}
#[tokio::test]
async fn test_workspace_write_and_read() {
let pool = get_pool();
if try_connect(&pool).await.is_none() {
return;
}
let user_id = "test_write_read";
cleanup_user(&pool, user_id).await;
let workspace = Workspace::new(user_id, pool.clone());
// Write a file
let doc = workspace
.write("README.md", "# Hello World\n\nThis is a test.")
.await
.expect("Failed to write");
assert_eq!(doc.path, "README.md");
assert!(doc.content.contains("Hello World"));
// Read it back
let doc2 = workspace.read("README.md").await.expect("Failed to read");
assert_eq!(doc2.content, "# Hello World\n\nThis is a test.");
// Cleanup
cleanup_user(&pool, user_id).await;
}
#[tokio::test]
async fn test_workspace_append() {
let pool = get_pool();
if try_connect(&pool).await.is_none() {
return;
}
let user_id = "test_append";
cleanup_user(&pool, user_id).await;
let workspace = Workspace::new(user_id, pool.clone());
// Write initial content
workspace
.write("notes.md", "Line 1")
.await
.expect("Failed to write");
// Append more
workspace
.append("notes.md", "Line 2")
.await
.expect("Failed to append");
// Read and verify
let doc = workspace.read("notes.md").await.expect("Failed to read");
assert_eq!(doc.content, "Line 1\nLine 2");
cleanup_user(&pool, user_id).await;
}
#[tokio::test]
async fn test_workspace_nested_paths() {
let pool = get_pool();
if try_connect(&pool).await.is_none() {
return;
}
let user_id = "test_nested";
cleanup_user(&pool, user_id).await;
let workspace = Workspace::new(user_id, pool.clone());
// Write nested files
workspace
.write("projects/alpha/README.md", "# Alpha")
.await
.expect("Failed to write alpha");
workspace
.write("projects/alpha/notes.md", "Notes here")
.await
.expect("Failed to write notes");
workspace
.write("projects/beta/README.md", "# Beta")
.await
.expect("Failed to write beta");
// List root
let root = workspace.list("").await.expect("Failed to list root");
assert_eq!(root.len(), 1); // just "projects/"
assert!(root[0].is_directory);
assert_eq!(root[0].name(), "projects");
// List projects
let projects = workspace
.list("projects")
.await
.expect("Failed to list projects");
assert_eq!(projects.len(), 2); // alpha/, beta/
// List alpha
let alpha = workspace
.list("projects/alpha")
.await
.expect("Failed to list alpha");
assert_eq!(alpha.len(), 2); // README.md, notes.md
cleanup_user(&pool, user_id).await;
}
#[tokio::test]
async fn test_workspace_delete() {
let pool = get_pool();
if try_connect(&pool).await.is_none() {
return;
}
let user_id = "test_delete";
cleanup_user(&pool, user_id).await;
let workspace = Workspace::new(user_id, pool.clone());
// Write and verify exists
workspace
.write("temp.md", "temporary")
.await
.expect("Failed to write");
assert!(workspace.exists("temp.md").await.expect("exists failed"));
// Delete
workspace.delete("temp.md").await.expect("Failed to delete");
// Verify gone
assert!(!workspace.exists("temp.md").await.expect("exists failed"));
cleanup_user(&pool, user_id).await;
}
#[tokio::test]
async fn test_workspace_memory_operations() {
let pool = get_pool();
if try_connect(&pool).await.is_none() {
return;
}
let user_id = "test_memory_ops";
cleanup_user(&pool, user_id).await;
let workspace = Workspace::new(user_id, pool.clone());
// Append to memory
workspace
.append_memory("User prefers dark mode")
.await
.expect("Failed to append memory");
workspace
.append_memory("User's timezone is PST")
.await
.expect("Failed to append memory");
// Read memory
let memory = workspace.memory().await.expect("Failed to get memory");
assert!(memory.content.contains("dark mode"));
assert!(memory.content.contains("PST"));
// Entries should be separated by double newline
assert!(memory.content.contains("\n\n"));
cleanup_user(&pool, user_id).await;
}
#[tokio::test]
async fn test_workspace_daily_log() {
let pool = get_pool();
if try_connect(&pool).await.is_none() {
return;
}
let user_id = "test_daily_log";
cleanup_user(&pool, user_id).await;
let workspace = Workspace::new(user_id, pool.clone());
// Append to daily log (timestamped)
workspace
.append_daily_log("Started working on feature X")
.await
.expect("Failed to append daily log");
// Read today's log
let log = workspace
.today_log()
.await
.expect("Failed to get today log");
assert!(log.content.contains("feature X"));
// Should have timestamp prefix like [HH:MM:SS]
assert!(log.content.contains("["));
cleanup_user(&pool, user_id).await;
}
#[tokio::test]
async fn test_workspace_fts_search() {
let pool = get_pool();
if try_connect(&pool).await.is_none() {
return;
}
let user_id = "test_fts_search";
cleanup_user(&pool, user_id).await;
let workspace = Workspace::new(user_id, pool.clone());
// Write some documents
workspace
.write(
"docs/authentication.md",
"# Authentication\n\nThe system uses JWT tokens for authentication.",
)
.await
.expect("write failed");
workspace
.write(
"docs/database.md",
"# Database\n\nWe use PostgreSQL with pgvector for vector search.",
)
.await
.expect("write failed");
workspace
.write(
"docs/api.md",
"# API\n\nThe REST API uses JSON for request and response bodies.",
)
.await
.expect("write failed");
// Search for JWT (FTS only since no embeddings)
let results = workspace
.search_with_config("JWT authentication", SearchConfig::default().fts_only())
.await
.expect("search failed");
assert!(!results.is_empty(), "Should find results for JWT");
assert!(
results[0].content.contains("JWT"),
"Top result should contain JWT"
);
// Search for PostgreSQL
let results = workspace
.search_with_config("PostgreSQL database", SearchConfig::default().fts_only())
.await
.expect("search failed");
assert!(!results.is_empty(), "Should find results for PostgreSQL");
assert!(
results[0].content.contains("PostgreSQL"),
"Top result should contain PostgreSQL"
);
cleanup_user(&pool, user_id).await;
}
#[tokio::test]
async fn test_workspace_hybrid_search_with_mock_embeddings() {
let pool = get_pool();
if try_connect(&pool).await.is_none() {
return;
}
let user_id = "test_hybrid_search";
cleanup_user(&pool, user_id).await;
// Create workspace with mock embeddings (1536 dimensions to match OpenAI)
let embeddings = Arc::new(MockEmbeddings::new(1536));
let workspace = Workspace::new(user_id, pool.clone()).with_embeddings(embeddings);
// Write documents
workspace
.write(
"memory.md",
"The user prefers dark mode and vim keybindings.",
)
.await
.expect("write failed");
workspace
.write(
"prefs.md",
"Settings: theme=dark, editor=vim, font=monospace",
)
.await
.expect("write failed");
// Hybrid search
let results = workspace
.search("dark theme preference", 5)
.await
.expect("search failed");
assert!(!results.is_empty(), "Should find results");
// At least one result should be a hybrid match (found by both FTS and vector)
// or we should have results from either method
cleanup_user(&pool, user_id).await;
}
#[tokio::test]
async fn test_workspace_list_all() {
let pool = get_pool();
if try_connect(&pool).await.is_none() {
return;
}
let user_id = "test_list_all";
cleanup_user(&pool, user_id).await;
let workspace = Workspace::new(user_id, pool.clone());
// Write files at various depths
workspace.write("README.md", "root").await.unwrap();
workspace.write("docs/intro.md", "intro").await.unwrap();
workspace.write("docs/api/rest.md", "rest").await.unwrap();
workspace.write("src/main.md", "main").await.unwrap();
// List all
let all = workspace.list_all().await.expect("list_all failed");
assert_eq!(all.len(), 4);
assert!(all.contains(&"README.md".to_string()));
assert!(all.contains(&"docs/intro.md".to_string()));
assert!(all.contains(&"docs/api/rest.md".to_string()));
assert!(all.contains(&"src/main.md".to_string()));
cleanup_user(&pool, user_id).await;
}
#[tokio::test]
async fn test_workspace_system_prompt() {
let pool = get_pool();
if try_connect(&pool).await.is_none() {
return;
}
let user_id = "test_system_prompt";
cleanup_user(&pool, user_id).await;
let workspace = Workspace::new(user_id, pool.clone());
// Write identity files
workspace
.write(paths::AGENTS, "You are a helpful assistant.")
.await
.unwrap();
workspace
.write(paths::SOUL, "Be kind and thorough.")
.await
.unwrap();
workspace.write(paths::USER, "Name: Alice").await.unwrap();
// Get system prompt
let prompt = workspace
.system_prompt()
.await
.expect("system_prompt failed");
assert!(
prompt.contains("helpful assistant"),
"Should include AGENTS.md"
);
assert!(
prompt.contains("kind and thorough"),
"Should include SOUL.md"
);
assert!(prompt.contains("Alice"), "Should include USER.md");
cleanup_user(&pool, user_id).await;
}