Merge branch 'main' into feat/cli/login

Author: aseembits93

codeflash/cli_cmds/cli.py

@@ -249,25 +249,29 @@ def project_root_from_module_root(module_root: Path, pyproject_file_path: Path)
 
 
 def handle_optimize_all_arg_parsing(args: Namespace) -> Namespace:
-    if hasattr(args, "all"):
-        import git
-
-        from codeflash.code_utils.git_utils import check_and_push_branch, get_repo_owner_and_name
-        from codeflash.code_utils.github_utils import require_github_app_or_exit
-
-        # Ensure that the user can actually open PRs on the repo.
-        try:
-            git_repo = git.Repo(search_parent_directories=True)
-        except git.exc.InvalidGitRepositoryError:
-            logger.exception(
-                "I couldn't find a git repository in the current directory. "
-                "I need a git repository to run --all and open PRs for optimizations. Exiting..."
-            )
-            apologize_and_exit()
-        if not args.no_pr and not check_and_push_branch(git_repo, git_remote=args.git_remote):
-            exit_with_message("Branch is not pushed...", error_on_exit=True)
-        owner, repo = get_repo_owner_and_name(git_repo)
-        if not args.no_pr:
+    if hasattr(args, "all") or (hasattr(args, "file") and args.file):
+        no_pr = getattr(args, "no_pr", False)
+
+        if not no_pr:
+            import git
+
+            from codeflash.code_utils.git_utils import check_and_push_branch, get_repo_owner_and_name
+            from codeflash.code_utils.github_utils import require_github_app_or_exit
+
+            # Ensure that the user can actually open PRs on the repo.
+            try:
+                git_repo = git.Repo(search_parent_directories=True)
+            except git.exc.InvalidGitRepositoryError:
+                mode = "--all" if hasattr(args, "all") else "--file"
+                logger.exception(
+                    f"I couldn't find a git repository in the current directory. "
+                    f"I need a git repository to run {mode} and open PRs for optimizations. Exiting..."
+                )
+                apologize_and_exit()
+            git_remote = getattr(args, "git_remote", None)
+            if not check_and_push_branch(git_repo, git_remote=git_remote):
+                exit_with_message("Branch is not pushed...", error_on_exit=True)
+            owner, repo = get_repo_owner_and_name(git_repo)
             require_github_app_or_exit(owner, repo)
     if not hasattr(args, "all"):
         args.all = None

codeflash/cli_cmds/cmd_init.py

@@ -172,10 +172,23 @@ def ask_run_end_to_end_test(args: Namespace) -> None:
         run_end_to_end_test(args, file_path)
 
 
-def is_valid_pyproject_toml(pyproject_toml_path: Path) -> tuple[bool, dict[str, Any] | None, str]:  # noqa: PLR0911
+def config_found(pyproject_toml_path: Union[str, Path]) -> tuple[bool, str]:
+    pyproject_toml_path = Path(pyproject_toml_path)
+
     if not pyproject_toml_path.exists():
-        return False, None, f"Configuration file not found: {pyproject_toml_path}"
+        return False, f"Configuration file not found: {pyproject_toml_path}"
+
+    if not pyproject_toml_path.is_file():
+        return False, f"Configuration file is not a file: {pyproject_toml_path}"
+
+    if pyproject_toml_path.suffix != ".toml":
+        return False, f"Configuration file is not a .toml file: {pyproject_toml_path}"
+
+    return True, ""
 
+
+def is_valid_pyproject_toml(pyproject_toml_path: Union[str, Path]) -> tuple[bool, dict[str, Any] | None, str]:
+    pyproject_toml_path = Path(pyproject_toml_path)
     try:
         config, _ = parse_config_file(pyproject_toml_path)
     except Exception as e:
@@ -207,6 +220,10 @@ def should_modify_pyproject_toml() -> tuple[bool, dict[str, Any] | None]:
 
     pyproject_toml_path = Path.cwd() / "pyproject.toml"
 
+    found, _ = config_found(pyproject_toml_path)
+    if not found:
+        return True, None
+
     valid, config, _message = is_valid_pyproject_toml(pyproject_toml_path)
     if not valid:
         # needs to be re-configured
@@ -1407,7 +1424,7 @@ def ask_for_telemetry() -> bool:
     from rich.prompt import Confirm
 
     return Confirm.ask(
-        "⚡️ Would you like to enable telemetry to help us improve the Codeflash experience?",
+        "⚡️ Help us improve Codeflash by sharing anonymous usage data (e.g. errors encountered)?",
         default=True,
         show_default=True,
     )

codeflash/code_utils/config_parser.py

@@ -105,7 +105,7 @@ def parse_config_file(
         if lsp_mode:
             # don't fail in lsp mode if codeflash config is not found.
             return {}, config_file_path
-        msg = f"Could not find the 'codeflash' block in the config file {config_file_path}. Please run 'codeflash init' to create the config file."
+        msg = f"Could not find the 'codeflash' block in the config file {config_file_path}. Please run 'codeflash init' to add Codeflash config in the pyproject.toml config file."
         raise ValueError(msg) from e
     assert isinstance(config, dict)
 

codeflash/code_utils/formatter.py

@@ -97,6 +97,9 @@ def is_diff_line(line: str) -> bool:
 
 
 def format_generated_code(generated_test_source: str, formatter_cmds: list[str]) -> str:
+    formatter_name = formatter_cmds[0].lower() if formatter_cmds else "disabled"
+    if formatter_name == "disabled":  # nothing to do if no formatter provided
+        return re.sub(r"\n{2,}", "\n\n", generated_test_source)
     with tempfile.TemporaryDirectory() as test_dir_str:
         # try running formatter, if nothing changes (could be due to formatting failing or no actual formatting needed) return code with 2 or more newlines substituted with 2 newlines
         original_temp = Path(test_dir_str) / "original_temp.py"

codeflash/code_utils/git_worktree_utils.py

@@ -1,5 +1,6 @@
 from __future__ import annotations
 
+import configparser
 import subprocess
 import tempfile
 import time
@@ -18,14 +19,36 @@
 
 def create_worktree_snapshot_commit(worktree_dir: Path, commit_message: str) -> None:
     repository = git.Repo(worktree_dir, search_parent_directories=True)
-    with repository.config_writer() as cw:
+    username = None
+    no_username = False
+    email = None
+    no_email = False
+    with repository.config_reader(config_level="repository") as cr:
+        try:
+            username = cr.get("user", "name")
+        except (configparser.NoSectionError, configparser.NoOptionError):
+            no_username = True
+        try:
+            email = cr.get("user", "email")
+        except (configparser.NoSectionError, configparser.NoOptionError):
+            no_email = True
+    with repository.config_writer(config_level="repository") as cw:
         if not cw.has_option("user", "name"):
             cw.set_value("user", "name", "Codeflash Bot")
         if not cw.has_option("user", "email"):
             cw.set_value("user", "email", "bot@codeflash.ai")
 
     repository.git.add(".")
     repository.git.commit("-m", commit_message, "--no-verify")
+    with repository.config_writer(config_level="repository") as cw:
+        if username:
+            cw.set_value("user", "name", username)
+        elif no_username:
+            cw.remove_option("user", "name")
+        if email:
+            cw.set_value("user", "email", email)
+        elif no_email:
+            cw.remove_option("user", "email")
 
 
 def create_detached_worktree(module_root: Path) -> Optional[Path]:

codeflash/code_utils/instrument_existing_tests.py

@@ -684,27 +684,6 @@ def visit_ImportFrom(self, node: ast.ImportFrom) -> None:
                     )
 
 
-def instrument_source_module_with_async_decorators(
-    source_path: Path, function_to_optimize: FunctionToOptimize, mode: TestingMode = TestingMode.BEHAVIOR
-) -> tuple[bool, str | None]:
-    if not function_to_optimize.is_async:
-        return False, None
-
-    try:
-        with source_path.open(encoding="utf8") as f:
-            source_code = f.read()
-
-        modified_code, decorator_added = add_async_decorator_to_function(source_code, function_to_optimize, mode)
-
-        if decorator_added:
-            return True, modified_code
-
-    except Exception:
-        return False, None
-    else:
-        return False, None
-
-
 def inject_async_profiling_into_existing_test(
     test_path: Path,
     call_positions: list[CodePosition],
@@ -1288,25 +1267,29 @@ def leave_Module(self, original_node: cst.Module, updated_node: cst.Module) -> c
 
 
 def add_async_decorator_to_function(
-    source_code: str, function: FunctionToOptimize, mode: TestingMode = TestingMode.BEHAVIOR
-) -> tuple[str, bool]:
-    """Add async decorator to an async function definition.
+    source_path: Path, function: FunctionToOptimize, mode: TestingMode = TestingMode.BEHAVIOR
+) -> bool:
+    """Add async decorator to an async function definition and write back to file.
 
     Args:
     ----
-        source_code: The source code to modify.
+        source_path: Path to the source file to modify in-place.
         function: The FunctionToOptimize object representing the target async function.
         mode: The testing mode to determine which decorator to apply.
 
     Returns:
     -------
-        Tuple of (modified_source_code, was_decorator_added).
+        Boolean indicating whether the decorator was successfully added.
 
     """
     if not function.is_async:
-        return source_code, False
+        return False
 
     try:
+        # Read source code
+        with source_path.open(encoding="utf8") as f:
+            source_code = f.read()
+
         module = cst.parse_module(source_code)
 
         # Add the decorator to the function
@@ -1318,10 +1301,17 @@ def add_async_decorator_to_function(
             import_transformer = AsyncDecoratorImportAdder(mode)
             module = module.visit(import_transformer)
 
-        return sort_imports(code=module.code, float_to_top=True), decorator_transformer.added_decorator
+        modified_code = sort_imports(code=module.code, float_to_top=True)
     except Exception as e:
         logger.exception(f"Error adding async decorator to function {function.qualified_name}: {e}")
-        return source_code, False
+        return False
+    else:
+        if decorator_transformer.added_decorator:
+            with source_path.open("w", encoding="utf8") as f:
+                f.write(modified_code)
+            logger.debug(f"Applied async {mode.value} instrumentation to {source_path}")
+            return True
+        return False
 
 
 def create_instrumented_source_module_path(source_path: Path, temp_dir: Path) -> Path:

codeflash/context/unused_definition_remover.py

@@ -469,22 +469,32 @@ def remove_unused_definitions_by_function_names(code: str, qualified_function_na
         qualified_function_names: Set of function names to keep. For methods, use format 'classname.methodname'
 
     """
-    module = cst.parse_module(code)
-    # Collect all definitions (top level classes, variables or function)
-    definitions = collect_top_level_definitions(module)
+    try:
+        module = cst.parse_module(code)
+    except Exception as e:
+        logger.debug(f"Failed to parse code with libcst: {type(e).__name__}: {e}")
+        return code
 
-    # Collect dependencies between definitions using the visitor pattern
-    dependency_collector = DependencyCollector(definitions)
-    module.visit(dependency_collector)
+    try:
+        # Collect all definitions (top level classes, variables or function)
+        definitions = collect_top_level_definitions(module)
 
-    # Mark definitions used by specified functions, and their dependencies recursively
-    usage_marker = QualifiedFunctionUsageMarker(definitions, qualified_function_names)
-    usage_marker.mark_used_definitions()
+        # Collect dependencies between definitions using the visitor pattern
+        dependency_collector = DependencyCollector(definitions)
+        module.visit(dependency_collector)
 
-    # Apply the recursive removal transformation
-    modified_module, _ = remove_unused_definitions_recursively(module, definitions)
+        # Mark definitions used by specified functions, and their dependencies recursively
+        usage_marker = QualifiedFunctionUsageMarker(definitions, qualified_function_names)
+        usage_marker.mark_used_definitions()
 
-    return modified_module.code if modified_module else ""
+        # Apply the recursive removal transformation
+        modified_module, _ = remove_unused_definitions_recursively(module, definitions)
+
+        return modified_module.code if modified_module else ""  # noqa: TRY300
+    except Exception as e:
+        # If any other error occurs during processing, return the original code
+        logger.debug(f"Error processing code to remove unused definitions: {type(e).__name__}: {e}")
+        return code
 
 
 def print_definitions(definitions: dict[str, UsageInfo]) -> None: