save optimization patches metadata

Author: mohammedahmed18

codeflash/code_utils/git_utils.py

@@ -1,17 +1,19 @@
 from __future__ import annotations
 
+import json
 import os
 import shutil
 import subprocess
 import sys
 import tempfile
 import time
-from functools import cache
+from functools import cache, lru_cache
 from io import StringIO
 from pathlib import Path
 from typing import TYPE_CHECKING, Optional
 
 import git
+from filelock import FileLock
 from rich.prompt import Confirm
 from unidiff import PatchSet
 
@@ -20,6 +22,8 @@
 from codeflash.code_utils.config_consts import N_CANDIDATES
 
 if TYPE_CHECKING:
+    from typing import Any
+
     from git import Repo
 
 
@@ -199,6 +203,14 @@ def get_last_commit_author_if_pr_exists(repo: Repo | None = None) -> str | None:
 patches_dir = codeflash_cache_dir / "patches"
 
 
+@lru_cache(maxsize=1)
+def get_git_project_id() -> str:
+    """Return the first commit sha of the repo."""
+    repo: Repo = git.Repo(search_parent_directories=True)
+    root_commits = list(repo.iter_commits(rev="HEAD", max_parents=0))
+    return root_commits[0].hexsha
+
+
 def create_worktree_snapshot_commit(worktree_dir: Path, commit_message: str) -> None:
     repository = git.Repo(worktree_dir, search_parent_directories=True)
     repository.git.add(".")
@@ -257,20 +269,70 @@ def remove_worktree(worktree_dir: Path) -> None:
         logger.exception(f"Failed to remove worktree: {worktree_dir}")
 
 
-def create_diff_patch_from_worktree(worktree_dir: Path, files: list[str], fto_name: str) -> Path:
+def get_patches_dir_for_project() -> Path:
+    project_id = get_git_project_id() or ""
+    return Path(patches_dir / project_id)
+
+
+def get_patches_metadata() -> dict[str, Any]:
+    project_patches_dir = get_patches_dir_for_project()
+    meta_file = project_patches_dir / "metadata.json"
+    if meta_file.exists():
+        return json.loads(meta_file.read_text())
+    return {"id": get_git_project_id() or "", "patches": []}
+
+
+def save_patches_metadata(patch_metadata: dict) -> dict:
+    project_patches_dir = get_patches_dir_for_project()
+    meta_file = project_patches_dir / "metadata.json"
+    lock_file = project_patches_dir / "metadata.json.lock"
+
+    with FileLock(lock_file, timeout=10):
+        metadata = get_patches_metadata()
+
+        patch_metadata["id"] = time.strftime("%Y%m%d-%H%M%S")
+        metadata["patches"].append(patch_metadata)
+
+        meta_file.write_text(json.dumps(metadata, indent=2))
+
+    return patch_metadata
+
+
+def overwrite_patch_metadata(patches: list[dict]) -> bool:
+    project_patches_dir = get_patches_dir_for_project()
+    meta_file = project_patches_dir / "metadata.json"
+    lock_file = project_patches_dir / "metadata.json.lock"
+
+    with FileLock(lock_file, timeout=10):
+        metadata = get_patches_metadata()
+        metadata["patches"] = patches
+        meta_file.write_text(json.dumps(metadata, indent=2))
+    return True
+
+
+def create_diff_patch_from_worktree(
+    worktree_dir: Path, files: list[str], metadata_input: dict[str, Any]
+) -> dict[str, Any]:
     repository = git.Repo(worktree_dir, search_parent_directories=True)
     uni_diff_text = repository.git.diff(None, "HEAD", *files, ignore_blank_lines=True, ignore_space_at_eol=True)
 
     if not uni_diff_text:
         logger.warning("No changes found in worktree.")
-        return None
+        return {}
 
     if not uni_diff_text.endswith("\n"):
         uni_diff_text += "\n"
 
-    # write to patches_dir
-    patches_dir.mkdir(parents=True, exist_ok=True)
-    patch_path = patches_dir / f"{worktree_dir.name}.{fto_name}.patch"
+    project_patches_dir = get_patches_dir_for_project()
+    project_patches_dir.mkdir(parents=True, exist_ok=True)
+
+    patch_path = project_patches_dir / f"{worktree_dir.name}.{metadata_input['fto_name']}.patch"
     with patch_path.open("w", encoding="utf8") as f:
         f.write(uni_diff_text)
-    return patch_path
+
+    final_metadata = {}
+    if metadata_input:
+        metadata_input["patch_path"] = str(patch_path)
+        final_metadata = save_patches_metadata(metadata_input)
+
+    return final_metadata

codeflash/lsp/beta.py

@@ -11,7 +11,11 @@
 
 from codeflash.api.cfapi import get_codeflash_api_key, get_user_id
 from codeflash.cli_cmds.cli import process_pyproject_config
-from codeflash.code_utils.git_utils import create_diff_patch_from_worktree
+from codeflash.code_utils.git_utils import (
+    create_diff_patch_from_worktree,
+    get_patches_metadata,
+    overwrite_patch_metadata,
+)
 from codeflash.code_utils.shell_utils import save_api_key_to_rc
 from codeflash.discovery.functions_to_optimize import filter_functions, get_functions_within_git_diff
 from codeflash.either import is_successful
@@ -216,6 +220,29 @@ def provide_api_key(server: CodeflashLanguageServer, params: ProvideApiKeyParams
         return {"status": "error", "message": "something went wrong while saving the api key"}
 
 
+@server.feature("onPatchApplied")
+def on_patch_applied(_server: CodeflashLanguageServer, params: dict[str, str]) -> dict[str, str]:
+    # first remove the patch from the metadata
+    patch_id = params["patch_id"]
+    metadata = get_patches_metadata()
+
+    deleted_patch_file = None
+    new_patches = []
+    for patch in metadata["patches"]:
+        if patch["id"] == patch_id:
+            deleted_patch_file = patch["patch_path"]
+            continue
+        new_patches.append(patch)
+
+    overwrite_patch_metadata(new_patches)
+    # then remove the patch file
+    if deleted_patch_file:
+        patch_path = Path(deleted_patch_file)
+        patch_path.unlink(missing_ok=True)
+        return {"status": "success"}
+    return {"status": "error", "message": "Patch not found"}
+
+
 @server.feature("performFunctionOptimization")
 @server.thread()
 def perform_function_optimization(  # noqa: PLR0911
@@ -317,24 +344,34 @@ def perform_function_optimization(  # noqa: PLR0911
 
         # generate a patch for the optimization
         relative_file_paths = [code_string.file_path for code_string in code_context.read_writable_code.code_strings]
-        patch_file = create_diff_patch_from_worktree(
+
+        speedup = original_code_baseline.runtime / best_optimization.runtime
+
+        # get the original file path in the actual project (not in the worktree)
+        original_args, _ = server.optimizer.original_args_and_test_cfg
+        relative_file_path = current_function.file_path.relative_to(server.args.project_root)
+        original_file_path = Path(original_args.project_root / relative_file_path).resolve()
+
+        metadata = create_diff_patch_from_worktree(
             server.optimizer.current_worktree,
             relative_file_paths,
-            server.optimizer.current_function_optimizer.function_to_optimize.qualified_name,
+            metadata_input={
+                "fto_name": function_to_optimize_qualified_name,
+                "explanation": best_optimization.explanation_v2,
+                "file_path": str(original_file_path),
+                "speedup": speedup,
+            },
         )
 
-        optimized_source = best_optimization.candidate.source_code.markdown
-        speedup = original_code_baseline.runtime / best_optimization.runtime
-
         server.show_message_log(f"Optimization completed for {params.functionName} with {speedup:.2f}x speedup", "Info")
 
         return {
             "functionName": params.functionName,
             "status": "success",
             "message": "Optimization completed successfully",
             "extra": f"Speedup: {speedup:.2f}x faster",
-            "optimization": optimized_source,
-            "patch_file": str(patch_file),
+            "patch_file": metadata["patch_path"],
+            "patch_id": metadata["id"],
             "explanation": best_optimization.explanation_v2,
         }
     finally:

codeflash/optimization/function_optimizer.py

@@ -631,14 +631,13 @@ def determine_best_candidate(
                                     executor=self.executor,
                                 )
                             )
-                    else:
-                        tree.add(
-                            f"Summed runtime: {humanize_runtime(best_test_runtime)} "
-                            f"(measured over {candidate_result.max_loop_count} "
-                            f"loop{'s' if candidate_result.max_loop_count > 1 else ''})"
-                        )
-                        tree.add(f"Speedup percentage: {perf_gain * 100:.1f}%")
-                        tree.add(f"Speedup ratio: {perf_gain + 1:.3f}X")
+                    tree.add(
+                        f"Summed runtime: {humanize_runtime(best_test_runtime)} "
+                        f"(measured over {candidate_result.max_loop_count} "
+                        f"loop{'s' if candidate_result.max_loop_count > 1 else ''})"
+                    )
+                    tree.add(f"Speedup percentage: {perf_gain * 100:.1f}%")
+                    tree.add(f"Speedup ratio: {perf_gain + 1:.3f}X")
                     console.print(tree)
                     if self.args.benchmark and benchmark_tree:
                         console.print(benchmark_tree)

codeflash/optimization/optimizer.py

@@ -343,16 +343,15 @@ def run(self) -> None:
                             optimizations_found += 1
                             # create a diff patch for successful optimization
                             if self.current_worktree:
-                                read_writable_code = best_optimization.unwrap().code_context.read_writable_code
+                                best_opt = best_optimization.unwrap()
+                                read_writable_code = best_opt.code_context.read_writable_code
                                 relative_file_paths = [
                                     code_string.file_path for code_string in read_writable_code.code_strings
                                 ]
-                                patch_path = create_diff_patch_from_worktree(
-                                    self.current_worktree,
-                                    relative_file_paths,
-                                    self.current_function_optimizer.function_to_optimize.qualified_name,
+                                metadata = create_diff_patch_from_worktree(
+                                    self.current_worktree, relative_file_paths, metadata_input={}
                                 )
-                                self.patch_files.append(patch_path)
+                                self.patch_files.append(metadata["patch_path"])
                                 if i < len(functions_to_optimize) - 1:
                                     create_worktree_snapshot_commit(
                                         self.current_worktree,