fix: set git config + misc daemon improvements (#677)

Author: caroljung-cg

src/codegen/cli/commands/run/run_daemon.py

@@ -46,7 +46,7 @@ def run_daemon(session: CodegenSession, function, diff_preview: int | None = Non
                     limited_diff = "\n".join(diff_lines[:diff_preview])
 
                     if truncated:
-                        limited_diff += "\n\n...\n\n[yellow]diff truncated to {diff_preview} lines, view the full change set on your local file system after using run with `--apply-local`[/yellow]"
+                        limited_diff += f"\n\n...\n\n[yellow]diff truncated to {diff_preview} lines[/yellow]"
 
                     panel = Panel(limited_diff, title="[bold]Diff Preview[/bold]", border_style="blue", padding=(1, 2), expand=False)
                     rich.print(panel)

src/codegen/cli/commands/start/main.py

@@ -1,3 +1,4 @@
+import platform as py_platform
 import subprocess
 from importlib.metadata import version
 from pathlib import Path
@@ -18,12 +19,11 @@
 
 
 @click.command(name="start")
-@click.option("--platform", "-t", type=click.Choice(["linux/amd64", "linux/arm64", "linux/amd64,linux/arm64"]), default="linux/amd64,linux/arm64", help="Target platform(s) for the Docker image")
 @click.option("--port", "-p", type=int, default=None, help="Port to run the server on")
 @click.option("--detached", "-d", is_flag=True, help="Run the server in detached mode")
 @click.option("--skip-build", is_flag=True, help="Skip building the Docker image")
 @click.option("--force", "-f", is_flag=True, help="Force start the server even if it is already running")
-def start_command(port: int | None, platform: str, detached: bool = False, skip_build: bool = False, force: bool = False) -> None:
+def start_command(port: int | None, detached: bool = False, skip_build: bool = False, force: bool = False) -> None:
     """Starts a local codegen server"""
     repo_path = Path.cwd().resolve()
     repo_config = RepoConfig.from_repo_path(str(repo_path))
@@ -42,15 +42,10 @@ def start_command(port: int | None, platform: str, detached: bool = False, skip_
 
     try:
         if not skip_build:
-            rich.print("[bold blue]Building Docker image...[/bold blue]")
-            _build_docker_image(codegen_root, platform)
-        rich.print("[bold blue]Starting Docker container...[/bold blue]")
+            _build_docker_image(codegen_root)
         _run_docker_container(repo_config, port, detached)
         rich.print(Panel(f"[green]Server started successfully![/green]\nAccess the server at: [bold]http://{_default_host}:{port}[/bold]", box=ROUNDED, title="Codegen Server"))
         # TODO: memory snapshot here
-    except subprocess.CalledProcessError as e:
-        rich.print(f"[bold red]Error:[/bold red] Failed to {e.cmd[0]} Docker container")
-        raise click.Abort()
     except Exception as e:
         rich.print(f"[bold red]Error:[/bold red] {e!s}")
         raise click.Abort()
@@ -75,7 +70,8 @@ def _handle_existing_container(repo_config: RepoConfig, container: DockerContain
     click.Abort()
 
 
-def _build_docker_image(codegen_root: Path, platform: str) -> None:
+def _build_docker_image(codegen_root: Path) -> None:
+    platform = _get_platform()
     build_cmd = [
         "docker",
         "buildx",
@@ -89,11 +85,31 @@ def _build_docker_image(codegen_root: Path, platform: str) -> None:
         "--load",
         str(codegen_root),
     ]
-    rich.print(f"build_cmd: {str.join(' ', build_cmd)}")
+    rich.print(
+        Panel(
+            f"{str.join(' ', build_cmd)}",
+            box=ROUNDED,
+            title="Running Build Command",
+            style="blue",
+            padding=(1, 1),
+        )
+    )
     subprocess.run(build_cmd, check=True)
 
 
+def _get_platform() -> str:
+    machine = py_platform.machine().lower()
+    if machine in ("x86_64", "amd64"):
+        return "linux/amd64"
+    elif machine in ("arm64", "aarch64"):
+        return "linux/arm64"
+    else:
+        rich.print(f"[yellow]Warning: Unknown architecture {machine}, defaulting to linux/amd64[/yellow]")
+        return "linux/amd64"
+
+
 def _run_docker_container(repo_config: RepoConfig, port: int, detached: bool) -> None:
+    rich.print("[bold blue]Starting Docker container...[/bold blue]")
     container_repo_path = f"/app/git/{repo_config.name}"
     name_args = ["--name", f"{repo_config.name}"]
     envvars = {
@@ -108,9 +124,17 @@ def _run_docker_container(repo_config: RepoConfig, port: int, detached: bool) ->
     entry_point = f"uv run --frozen uvicorn codegen.runner.servers.local_daemon:app --host {_default_host} --port {port}"
     port_args = ["-p", f"{port}:{port}"]
     detached_args = ["-d"] if detached else []
-    run_cmd = ["docker", "run", *detached_args, *port_args, *name_args, *mount_args, *envvars_args, CODEGEN_RUNNER_IMAGE, entry_point]
-
-    rich.print(f"run_cmd: {str.join(' ', run_cmd)}")
+    run_cmd = ["docker", "run", "--rm", *detached_args, *port_args, *name_args, *mount_args, *envvars_args, CODEGEN_RUNNER_IMAGE, entry_point]
+
+    rich.print(
+        Panel(
+            f"{str.join(' ', run_cmd)}",
+            box=ROUNDED,
+            title="Running Run Command",
+            style="blue",
+            padding=(1, 1),
+        )
+    )
     subprocess.run(run_cmd, check=True)
 
     if detached:

src/codegen/git/repo_operator/repo_operator.py

@@ -145,31 +145,25 @@ def git_cli(self) -> GitCLI:
         for level in levels:
             with git_cli.config_reader(level) as reader:
                 if reader.has_option("user", "name") and not username:
-                    username = reader.get("user", "name")
-                    user_level = level
+                    username = username or reader.get("user", "name")
+                    user_level = user_level or level
                 if reader.has_option("user", "email") and not email:
-                    email = reader.get("user", "email")
-                    email_level = level
-        if self.bot_commit:
-            self._set_bot_email(git_cli)
+                    email = email or reader.get("user", "email")
+                    email_level = email_level or level
+
+        # We need a username and email to commit, so if they're not set, set them to the bot's
+        if not username or self.bot_commit:
             self._set_bot_username(git_cli)
-        else:
-            # we need a username and email to commit, so if they're not set, set them to the bot's
-            # Case 1: username is not set: set it to the bot's
-            if not username:
-                self._set_bot_username(git_cli)
-            # Case 2: username is set to the bot's at the repo level, but something else is set at the user level: unset it
-            elif username != CODEGEN_BOT_NAME and user_level != "repository":
-                self._unset_bot_username(git_cli)
-            #  3: Caseusername is only set at the repo level: do nothing
-            else:
-                pass  # no-op to make the logic clearer
-            # Repeat for email
-            if not email:
-                self._set_bot_email(git_cli)
+        if not email or self.bot_commit:
+            self._set_bot_email(git_cli)
 
-            elif email != CODEGEN_BOT_EMAIL and email_level != "repository":
+        # If user config is set at a level above the repo level: unset it
+        if not self.bot_commit:
+            if username and username != CODEGEN_BOT_NAME and user_level != "repository":
+                self._unset_bot_username(git_cli)
+            if email and email != CODEGEN_BOT_EMAIL and email_level != "repository":
                 self._unset_bot_email(git_cli)
+
         return git_cli
 
     @property

src/codegen/runner/sandbox/runner.py

@@ -28,9 +28,9 @@ class SandboxRunner:
     codebase: CodebaseType
     executor: SandboxExecutor
 
-    def __init__(self, repo_config: RepoConfig) -> None:
+    def __init__(self, repo_config: RepoConfig, op: RepoOperator | None = None) -> None:
         self.repo = repo_config
-        self.op = RepoOperator(repo_config=self.repo, setup_option=SetupOption.PULL_OR_CLONE, bot_commit=True)
+        self.op = op or RepoOperator(repo_config=self.repo, setup_option=SetupOption.PULL_OR_CLONE, bot_commit=True)
         self.commit = self.op.git_cli.head.commit
 
     async def warmup(self) -> None:

src/codegen/runner/servers/local_daemon.py

@@ -4,6 +4,8 @@
 from fastapi import FastAPI
 
 from codegen.git.configs.constants import CODEGEN_BOT_EMAIL, CODEGEN_BOT_NAME
+from codegen.git.repo_operator.repo_operator import RepoOperator
+from codegen.git.schemas.enums import SetupOption
 from codegen.git.schemas.repo_config import RepoConfig
 from codegen.runner.enums.warmup_state import WarmupState
 from codegen.runner.models.apis import (
@@ -14,14 +16,15 @@
 )
 from codegen.runner.models.codemod import Codemod, CodemodRunResult
 from codegen.runner.sandbox.runner import SandboxRunner
+from codegen.shared.logging.get_logger import get_logger
 
 # Configure logging at module level
 logging.basicConfig(
     level=logging.INFO,
     format="%(asctime)s - %(name)s - %(levelname)s - %(message)s",
     force=True,
 )
-logger = logging.getLogger(__name__)
+logger = get_logger(__name__)
 
 server_info: ServerInfo
 runner: SandboxRunner
@@ -37,13 +40,14 @@ async def lifespan(server: FastAPI):
         server_info = ServerInfo(repo_name=repo_config.full_name or repo_config.name)
 
         # Set the bot email and username
+        op = RepoOperator(repo_config=repo_config, setup_option=SetupOption.SKIP, bot_commit=True)
+        runner = SandboxRunner(repo_config=repo_config, op=op)
         logger.info(f"Configuring git user config to {CODEGEN_BOT_EMAIL} and {CODEGEN_BOT_NAME}")
-        runner = SandboxRunner(repo_config=repo_config)
         runner.op.git_cli.git.config("user.email", CODEGEN_BOT_EMAIL)
         runner.op.git_cli.git.config("user.name", CODEGEN_BOT_NAME)
 
         # Parse the codebase
-        logger.info(f"Starting up sandbox fastapi server for repo_name={repo_config.name}")
+        logger.info(f"Starting up fastapi server for repo_name={repo_config.name}")
         server_info.warmup_state = WarmupState.PENDING
         await runner.warmup()
         server_info.synced_commit = runner.commit.hexsha

src/codegen/sdk/codebase/codebase_context.py

@@ -31,6 +31,7 @@
 from codegen.sdk.typescript.external.ts_declassify.ts_declassify import TSDeclassify
 from codegen.shared.enums.programming_language import ProgrammingLanguage
 from codegen.shared.exceptions.control_flow import StopCodemodException
+from codegen.shared.logging.get_logger import get_logger
 from codegen.shared.performance.stopwatch_utils import stopwatch, stopwatch_with_sentry
 
 if TYPE_CHECKING:
@@ -51,9 +52,7 @@
     from codegen.sdk.core.node_id_factory import NodeId
     from codegen.sdk.core.parser import Parser
 
-import logging
-
-logger = logging.getLogger(__name__)
+logger = get_logger(__name__)
 
 
 # src/vs/platform/contextview/browser/contextMenuService.ts is ignored as there is a parsing error with tree-sitter

src/codegen/shared/logging/get_logger.py

@@ -2,9 +2,12 @@
 
 import colorlog
 
-handler = colorlog.StreamHandler()
-handler.setFormatter(
-    colorlog.ColoredFormatter(
+
+def get_logger(name: str, level: int = logging.INFO) -> logging.Logger:
+    # Force configure the root logger with a NullHandler to prevent duplicate logs
+    logging.basicConfig(handlers=[logging.NullHandler()], force=True)
+
+    formatter = colorlog.ColoredFormatter(
         "%(white)s%(asctime)s - %(name)s - %(log_color)s%(levelname)s%(reset)s%(white)s - %(message_log_color)s%(message)s",
         log_colors={
             "DEBUG": "cyan",
@@ -23,11 +26,16 @@
             }
         },
     )
-)
-
-
-def get_logger(name: str, level: int = logging.INFO) -> logging.Logger:
     logger = logging.getLogger(name)
+    if logger.hasHandlers():
+        for h in logger.handlers:
+            logger.removeHandler(h)
+
+    handler = colorlog.StreamHandler()
+    handler.setFormatter(formatter)
     logger.addHandler(handler)
+    # Ensure the logger propagates to the root logger
+    logger.propagate = False
+    # Set the level on the logger itself
     logger.setLevel(level)
     return logger