agent: Improve initial file search quality (#29317)

This PR significantly improves the quality of the initial file search
that occurs when the model doesn't yet know the full path to a file it
needs to read/edit.

Previously, the assertions in file_search often failed on main as the
model attempted to guess full file paths. On this branch, it reliably
calls `find_path` (previously `path_search`) before reading files.

After getting the model to find paths first, I noticed it would try
using `grep` instead of `path_search`. This motivated renaming
`path_search` to `find_path` (continuing the analogy to unix commands)
and adding system prompt instructions about proper tool selection.

Note: I know the command is just called `find`, but that seemed too
general.

In my eval runs, the `file_search` example improved from 40% ± 10% to
98% ± 2%. The only assertion I'm seeing occasionally fail is "glob
starts with `**` or project". We can probably add some instructions in
that regard.

Release Notes:

- N/A

crates/assistant_tools/src/assistant_tools.rs

@@ -9,12 +9,12 @@ mod delete_path_tool;
 mod diagnostics_tool;
 mod edit_file_tool;
 mod fetch_tool;
+mod find_path_tool;
 mod grep_tool;
 mod list_directory_tool;
 mod move_path_tool;
 mod now_tool;
 mod open_tool;
-mod path_search_tool;
 mod read_file_tool;
 mod rename_tool;
 mod replace;
@@ -45,11 +45,11 @@ use crate::delete_path_tool::DeletePathTool;
 use crate::diagnostics_tool::DiagnosticsTool;
 use crate::edit_file_tool::EditFileTool;
 use crate::fetch_tool::FetchTool;
+use crate::find_path_tool::FindPathTool;
 use crate::grep_tool::GrepTool;
 use crate::list_directory_tool::ListDirectoryTool;
 use crate::now_tool::NowTool;
 use crate::open_tool::OpenTool;
-use crate::path_search_tool::PathSearchTool;
 use crate::read_file_tool::ReadFileTool;
 use crate::rename_tool::RenameTool;
 use crate::symbol_info_tool::SymbolInfoTool;
@@ -58,7 +58,7 @@ use crate::thinking_tool::ThinkingTool;
 
 pub use create_file_tool::CreateFileToolInput;
 pub use edit_file_tool::EditFileToolInput;
-pub use path_search_tool::PathSearchToolInput;
+pub use find_path_tool::FindPathToolInput;
 pub use read_file_tool::ReadFileToolInput;
 
 pub fn init(http_client: Arc<HttpClientWithUrl>, cx: &mut App) {
@@ -81,7 +81,7 @@ pub fn init(http_client: Arc<HttpClientWithUrl>, cx: &mut App) {
     registry.register_tool(OpenTool);
     registry.register_tool(CodeSymbolsTool);
     registry.register_tool(ContentsTool);
-    registry.register_tool(PathSearchTool);
+    registry.register_tool(FindPathTool);
     registry.register_tool(ReadFileTool);
     registry.register_tool(GrepTool);
     registry.register_tool(RenameTool);

crates/assistant_tools/src/batch_tool.rs

@@ -97,7 +97,7 @@ pub struct BatchToolInput {
     ///       }
     ///     },
     ///     {
-    ///       "name": "path_search",
+    ///       "name": "find_path",
     ///       "input": {
     ///         "glob": "**/*test*.rs"
     ///       }

crates/assistant_tools/src/find_path_tool.rs

@@ -12,7 +12,7 @@ use util::paths::PathMatcher;
 use worktree::Snapshot;
 
 #[derive(Debug, Serialize, Deserialize, JsonSchema)]
-pub struct PathSearchToolInput {
+pub struct FindPathToolInput {
     /// The glob to match against every path in the project.
     ///
     /// <example>
@@ -34,11 +34,11 @@ pub struct PathSearchToolInput {
 
 const RESULTS_PER_PAGE: usize = 50;
 
-pub struct PathSearchTool;
+pub struct FindPathTool;
 
-impl Tool for PathSearchTool {
+impl Tool for FindPathTool {
     fn name(&self) -> String {
-        "path_search".into()
+        "find_path".into()
     }
 
     fn needs_confirmation(&self, _: &serde_json::Value, _: &App) -> bool {
@@ -46,7 +46,7 @@ impl Tool for PathSearchTool {
     }
 
     fn description(&self) -> String {
-        include_str!("./path_search_tool/description.md").into()
+        include_str!("./find_path_tool/description.md").into()
     }
 
     fn icon(&self) -> IconName {
@@ -54,11 +54,11 @@ impl Tool for PathSearchTool {
     }
 
     fn input_schema(&self, format: LanguageModelToolSchemaFormat) -> Result<serde_json::Value> {
-        json_schema_for::<PathSearchToolInput>(format)
+        json_schema_for::<FindPathToolInput>(format)
     }
 
     fn ui_text(&self, input: &serde_json::Value) -> String {
-        match serde_json::from_value::<PathSearchToolInput>(input.clone()) {
+        match serde_json::from_value::<FindPathToolInput>(input.clone()) {
             Ok(input) => format!("Find paths matching “`{}`”", input.glob),
             Err(_) => "Search paths".to_string(),
         }
@@ -73,7 +73,7 @@ impl Tool for PathSearchTool {
         _window: Option<AnyWindowHandle>,
         cx: &mut App,
     ) -> ToolResult {
-        let (offset, glob) = match serde_json::from_value::<PathSearchToolInput>(input) {
+        let (offset, glob) = match serde_json::from_value::<FindPathToolInput>(input) {
             Ok(input) => (input.offset, input.glob),
             Err(err) => return Task::ready(Err(anyhow!(err))).into(),
         };
@@ -144,7 +144,7 @@ mod test {
     use util::path;
 
     #[gpui::test]
-    async fn test_path_search_tool(cx: &mut TestAppContext) {
+    async fn test_find_path_tool(cx: &mut TestAppContext) {
         init_test(cx);
 
         let fs = FakeFs::new(cx.executor());

crates/assistant_tools/src/find_path_tool/description.md

@@ -1,4 +1,4 @@
-Fast file pattern matching tool that works with any codebase size
+Fast file path pattern matching tool that works with any codebase size
 
 - Supports glob patterns like "**/*.js" or "src/**/*.ts"
 - Returns matching file paths sorted alphabetically

crates/assistant_tools/src/grep_tool.rs

@@ -20,6 +20,8 @@ use util::paths::PathMatcher;
 pub struct GrepToolInput {
     /// A regex pattern to search for in the entire project. Note that the regex
     /// will be parsed by the Rust `regex` crate.
+    ///
+    /// Do NOT specify a path here! This will only be matched against the code **content**.
     pub regex: String,
 
     /// A glob pattern for the paths of files to include in the search.

crates/assistant_tools/src/list_directory_tool/description.md

@@ -1 +1 @@
-Lists files and directories in a given path. Prefer the `grep` or `path_search` tools when searching the codebase.
+Lists files and directories in a given path. Prefer the `grep` or `find_path` tools when searching the codebase.