feat: basic Import resolution

.cargo/config.toml

@@ -1,2 +1,2 @@
 [build]
-rustflags = ["-Clink-arg=-fuse-ld=lld", "-Zthreads=16"]
+rustflags = ["-Clink-arg=-fuse-ld=lld", "-Zthreads=16", "-Ctarget-cpu=native"]

.gitignore

@@ -68,3 +68,4 @@ tests/integration/verified_codemods/codemod_data/repo_commits.json
 target/*
 .benchmarks/*
 **.snap.new
+flamegraph.svg

Cargo.toml

@@ -7,20 +7,21 @@ edition = "2024"
 
 [dependencies]
 clap = { version = "4.5.28", features = ["derive"] }
-codegen-sdk-analyzer = { path = "codegen-sdk-analyzer" }
+codegen-sdk-analyzer = { path = "codegen-sdk-analyzer", default-features = false }
 codegen-sdk-ast = { workspace = true}
 codegen-sdk-common = { workspace = true}
 anyhow = { workspace = true}
 salsa = { workspace = true}
-codegen-sdk-typescript = { workspace = true}
+codegen-sdk-typescript = { workspace = true, optional = true }
+codegen-sdk-python = { workspace = true, optional = true }
 env_logger = { workspace = true }
 log = { workspace = true }
 codegen-sdk-resolution = { workspace = true}
 sysinfo = "0.33.1"
 rkyv.workspace = true
 [features]
-python = [ "codegen-sdk-analyzer/python"] # TODO: Add python support
-typescript = [ "codegen-sdk-analyzer/typescript"]
+python = [ "codegen-sdk-analyzer/python", "codegen-sdk-python"]
+typescript = [ "codegen-sdk-analyzer/typescript", "codegen-sdk-typescript"]
 tsx = [ "codegen-sdk-analyzer/tsx"]
 jsx = [ "codegen-sdk-analyzer/jsx"]
 javascript = [ "codegen-sdk-analyzer/javascript"]
@@ -116,15 +117,16 @@ insta = "1.42.1"
 prettyplease = "0.2.29"
 syn = { version = "2.0.98", features = ["proc-macro", "full"] }
 derive_more = { version = "2.0.1", features = ["debug", "display"] }
-salsa = {git = "https://github.com/salsa-rs/salsa", branch = "master"}
+salsa = {git = "https://github.com/salsa-rs/salsa", rev ="dbb0e5f6ab2cd61e42b372f333ab694f24141cf1"}
 subenum = {git = "https://github.com/mrenow/subenum", branch = "main"}
 indicatif-log-bridge = "0.2.3"
 indicatif = { version = "0.17.11", features = ["rayon"] }
 crossbeam-channel = "0.5.11"
 rstest = "0.25.0"
 indextree = "4.7.3"
 thiserror = "2.0.11"
-
+indexmap = "2"
+smallvec = "1.11.0"
 [profile.dev]
 # codegen-backend = "cranelift"
 # split-debuginfo = "unpacked"
@@ -157,3 +159,7 @@ lto = false
 name = "parse"
 harness = false
 required-features = ["stable"]
+
+[profile.profiling]
+inherits = "release"
+debug = true

codegen-sdk-analyzer/src/codebase.rs

@@ -1,21 +1,18 @@
 use std::path::PathBuf;
 
 use anyhow::Context;
-use codegen_sdk_ast::Input;
 #[cfg(feature = "serialization")]
 use codegen_sdk_common::serialization::Cache;
-use codegen_sdk_resolution::CodebaseContext;
+use codegen_sdk_resolution::{CodebaseContext, Db};
 use discovery::FilesToParse;
 use notify_debouncer_mini::DebounceEventResult;
 use salsa::Setter;
 
-use crate::{
-    ParsedFile,
-    database::{CodegenDatabase, Db},
-    parser::parse_file,
-};
+use crate::{ParsedFile, database::CodegenDatabase, parser::parse_file};
 mod discovery;
 mod parser;
+use parser::execute_op_with_progress;
+
 pub struct Codebase {
     db: CodegenDatabase,
     root: PathBuf,
@@ -26,8 +23,9 @@ pub struct Codebase {
 
 impl Codebase {
     pub fn new(root: PathBuf) -> Self {
+        let root = root.canonicalize().unwrap();
         let (tx, rx) = crossbeam_channel::unbounded();
-        let mut db = CodegenDatabase::new(tx);
+        let mut db = CodegenDatabase::new(tx, root.clone());
         db.watch_dir(PathBuf::from(&root)).unwrap();
         let codebase = Self { db, root, rx };
         codebase.sync();
@@ -47,8 +45,7 @@ impl Codebase {
                     // to kick in, just like any other update to a salsa input.
                     let contents = std::fs::read_to_string(path)
                         .with_context(|| format!("Failed to read file {}", event.path.display()))?;
-                    let input = Input::new(&self.db, contents);
-                    file.set_contents(&mut self.db).to(input);
+                    file.set_content(&mut self.db).to(contents);
                 }
                 Err(e) => {
                     log::error!(
@@ -68,7 +65,7 @@ impl Codebase {
 
     pub fn errors(&self) -> Vec<()> {
         let mut errors = Vec::new();
-        for file in self.discover().files(&self.db) {
+        for file in self.db.files() {
             if self.get_file(file.path(&self.db)).is_none() {
                 errors.push(());
             }
@@ -84,9 +81,29 @@ impl Codebase {
             files,
         )
     }
+    fn _db(&self) -> &dyn Db {
+        &self.db
+    }
+    pub fn execute_op_with_progress<T: Send + Sync>(
+        &self,
+        name: &str,
+        parallel: bool,
+        op: fn(&dyn Db, codegen_sdk_common::FileNodeId<'_>) -> T,
+    ) -> Vec<T> {
+        execute_op_with_progress(
+            self._db(),
+            codegen_sdk_resolution::files(self._db()),
+            name,
+            parallel,
+            op,
+        )
+    }
 }
 impl CodebaseContext for Codebase {
     type File<'a> = ParsedFile<'a>;
+    fn root_path(&self) -> PathBuf {
+        self.root.clone()
+    }
     fn files<'a>(&'a self) -> Vec<&'a Self::File<'a>> {
         let mut files = Vec::new();
         for file in self.discover().files(&self.db) {
@@ -96,13 +113,16 @@ impl CodebaseContext for Codebase {
         }
         files
     }
-    fn db(&self) -> &dyn salsa::Database {
+    fn db(&self) -> &dyn Db {
         &self.db
     }
     fn get_file<'a>(&'a self, path: PathBuf) -> Option<&'a Self::File<'a>> {
-        let file = self.db.files.get(&path);
-        if let Some(file) = file {
-            return parse_file(&self.db, file.clone()).file(&self.db).as_ref();
+        if let Ok(path) = path.canonicalize() {
+            let file = self.db.files.get(&path);
+            if let Some(_) = file {
+                let file_id = codegen_sdk_common::FileNodeId::new(&self.db, path);
+                return parse_file(&self.db, file_id).file(&self.db).as_ref();
+            }
         }
         None
     }

codegen-sdk-analyzer/src/codebase/discovery.rs

@@ -3,12 +3,14 @@ use std::path::PathBuf;
 use codegen_sdk_ast::*;
 #[cfg(feature = "serialization")]
 use codegen_sdk_common::serialize::Cache;
+use codegen_sdk_resolution::Db;
 use glob::glob;
 
-use crate::database::{CodegenDatabase, Db};
+use crate::database::CodegenDatabase;
 #[salsa::input]
 pub struct FilesToParse {
-    pub files: Vec<codegen_sdk_ast::input::File>,
+    pub files: codegen_sdk_common::hash::FxHashSet<codegen_sdk_cst::File>,
+    pub root: PathBuf,
 }
 pub fn log_languages() {
     for language in LANGUAGES.iter() {
@@ -40,7 +42,8 @@ pub fn collect_files(db: &CodegenDatabase, dir: &PathBuf) -> FilesToParse {
         .into_iter()
         .filter_map(|file| file.ok())
         .filter(|file| !file.is_dir() && !file.is_symlink())
+        .filter_map(|file| file.canonicalize().ok())
         .map(|file| db.input(file).unwrap())
         .collect();
-    FilesToParse::new(db, files)
+    FilesToParse::new(db, files, dir)
 }