github2gerrit 0.1.4__py3-none-any.whl → 0.1.6__py3-none-any.whl

github2gerrit/cli.py

@@ -7,8 +7,12 @@ import json
 import logging
 import os
 import tempfile
+from collections.abc import Callable
 from pathlib import Path
+from typing import TYPE_CHECKING
 from typing import Any
+from typing import Protocol
+from typing import TypeVar
 from typing import cast
 from urllib.parse import urlparse
 
@@ -16,7 +20,9 @@ import click
 import typer
 
 from . import models
+from .config import _is_github_actions_context
 from .config import apply_config_to_env
+from .config import apply_parameter_derivation
 from .config import load_org_config
 from .core import Orchestrator
 from .core import SubmissionResult
@@ -31,6 +37,19 @@ from .models import GitHubContext
 from .models import Inputs
 
 
+def _is_verbose_mode() -> bool:
+    """Check if verbose mode is enabled via environment variable."""
+    return os.getenv("G2G_VERBOSE", "").lower() in ("true", "1", "yes")
+
+
+def _log_exception_conditionally(logger: logging.Logger, message: str, *args: Any) -> None:
+    """Log exception with traceback only if verbose mode is enabled."""
+    if _is_verbose_mode():
+        logger.exception(message, *args)
+    else:
+        logger.error(message, *args)
+
+
 class ConfigurationError(Exception):
     """Raised when configuration validation fails.
 
@@ -83,18 +102,36 @@ def _parse_github_target(url: str) -> tuple[str | None, str | None, int | None]:
 APP_NAME = "github2gerrit"
 
 
-class _SingleUsageGroup(click.Group):  # type: ignore[misc]
-    def format_usage(self, ctx: Any, formatter: Any) -> None:
+if TYPE_CHECKING:
+    BaseGroup = object
+else:
+    BaseGroup = click.Group
+
+
+class _FormatterProto(Protocol):
+    def write_usage(self, prog: str, args: str, prefix: str = ...) -> None: ...
+
+
+class _ContextProto(Protocol):
+    @property
+    def command_path(self) -> str: ...
+
+
+class _SingleUsageGroup(BaseGroup):
+    def format_usage(self, ctx: _ContextProto, formatter: _FormatterProto) -> None:
         # Force a simplified usage line without COMMAND [ARGS]...
-        formatter.write_usage(
-            ctx.command_path, "[OPTIONS] TARGET_URL", prefix="Usage: "
-        )
+        formatter.write_usage(ctx.command_path, "[OPTIONS] TARGET_URL", prefix="Usage: ")
+
 
+# Error message constants to comply with TRY003
+_MSG_MISSING_REQUIRED_INPUT = "Missing required input: {field_name}"
+_MSG_INVALID_FETCH_DEPTH = "FETCH_DEPTH must be a positive integer"
+_MSG_ISSUE_ID_MULTILINE = "Issue ID must be single line"
 
 app: typer.Typer = typer.Typer(
     add_completion=False,
     no_args_is_help=False,
-    cls=_SingleUsageGroup,
+    cls=cast(Any, _SingleUsageGroup),
 )
 
 
@@ -108,7 +145,15 @@ def _resolve_org(default_org: str | None) -> str:
     return ""
 
 
-@app.command()  # type: ignore[misc]
+if TYPE_CHECKING:
+    F = TypeVar("F", bound=Callable[..., object])
+
+    def typed_app_command(*args: object, **kwargs: object) -> Callable[[F], F]: ...
+else:
+    typed_app_command = app.command
+
+
+@typed_app_command()
 def main(
     ctx: typer.Context,
     target_url: str | None = typer.Argument(
@@ -216,6 +261,14 @@ def main(
         envvar="ALLOW_DUPLICATES",
         help="Allow submitting duplicate changes without error.",
     ),
+    duplicates: str = typer.Option(
+        "open",
+        "--duplicates",
+        envvar="DUPLICATES",
+        help=(
+            'Gerrit statuses for duplicate detection (comma-separated). E.g. "open,merged,abandoned". Default: "open".'
+        ),
+    ),
     verbose: bool = typer.Option(
         False,
         "--verbose",
@@ -275,6 +328,8 @@ def main(
         os.environ["ISSUE_ID"] = issue_id
     if allow_duplicates:
         os.environ["ALLOW_DUPLICATES"] = "true"
+    if duplicates:
+        os.environ["DUPLICATES"] = duplicates
     # URL mode handling
     if target_url:
         org, repo, pr = _parse_github_target(target_url)
@@ -302,10 +357,7 @@ def main(
 def _setup_logging() -> logging.Logger:
     level_name = os.getenv("G2G_LOG_LEVEL", "INFO").upper()
     level = getattr(logging, level_name, logging.INFO)
-    fmt = (
-        "%(asctime)s %(levelname)-8s %(name)s "
-        "%(filename)s:%(lineno)d | %(message)s"
-    )
+    fmt = "%(asctime)s %(levelname)-8s %(name)s %(filename)s:%(lineno)d | %(message)s"
     logging.basicConfig(level=level, format=fmt)
     return logging.getLogger(APP_NAME)
 
@@ -344,9 +396,7 @@ def _build_inputs_from_env() -> Inputs:
         gerrit_ssh_privkey_g2g=_env_str("GERRIT_SSH_PRIVKEY_G2G"),
         gerrit_ssh_user_g2g=_env_str("GERRIT_SSH_USER_G2G"),
         gerrit_ssh_user_g2g_email=_env_str("GERRIT_SSH_USER_G2G_EMAIL"),
-        organization=_env_str(
-            "ORGANIZATION", _env_str("GITHUB_REPOSITORY_OWNER")
-        ),
+        organization=_env_str("ORGANIZATION", _env_str("GITHUB_REPOSITORY_OWNER")),
         reviewers_email=_env_str("REVIEWERS_EMAIL", ""),
         preserve_github_prs=_env_bool("PRESERVE_GITHUB_PRS", False),
         dry_run=_env_bool("DRY_RUN", False),
@@ -355,6 +405,7 @@ def _build_inputs_from_env() -> Inputs:
         gerrit_project=_env_str("GERRIT_PROJECT"),
         issue_id=_env_str("ISSUE_ID"),
         allow_duplicates=_env_bool("ALLOW_DUPLICATES", False),
+        duplicates_filter=_env_str("DUPLICATES", "open"),
     )
 
 
@@ -388,20 +439,23 @@ def _process_bulk(data: Inputs, gh: GitHubContext) -> None:
 
         log.info("Starting processing of PR #%d", pr_number)
         log.debug(
-            "Processing PR #%d in multi-PR mode with event_name=%s, "
-            "event_action=%s",
+            "Processing PR #%d in multi-PR mode with event_name=%s, event_action=%s",
             pr_number,
             gh.event_name,
             gh.event_action,
         )
 
         try:
-            check_for_duplicates(
-                per_ctx, allow_duplicates=data.allow_duplicates
-            )
+            if data.duplicates_filter:
+                os.environ["DUPLICATES"] = data.duplicates_filter
+            check_for_duplicates(per_ctx, allow_duplicates=data.allow_duplicates)
         except DuplicateChangeError as exc:
-            log.exception("Skipping PR #%d", pr_number)
-            typer.echo(f"Skipping PR #{pr_number}: {exc}")
+            _log_exception_conditionally(log, "Skipping PR #%d", pr_number)
+            log.warning(
+                "Skipping PR #%d due to duplicate detection: %s. Use --allow-duplicates to override this check.",
+                pr_number,
+                exc,
+            )
             continue
 
         try:
@@ -426,7 +480,7 @@ def _process_bulk(data: Inputs, gh: GitHubContext) -> None:
                         result_multi.change_numbers,
                     )
         except Exception as exc:
-            log.exception("Failed to process PR #%d", pr_number)
+            _log_exception_conditionally(log, "Failed to process PR #%d", pr_number)
             typer.echo(f"Failed to process PR #{pr_number}: {exc}")
             log.info("Continuing to next PR despite failure")
             continue
@@ -438,12 +492,8 @@ def _process_bulk(data: Inputs, gh: GitHubContext) -> None:
 
     _append_github_output(
         {
-            "gerrit_change_request_url": os.getenv(
-                "GERRIT_CHANGE_REQUEST_URL", ""
-            ),
-            "gerrit_change_request_num": os.getenv(
-                "GERRIT_CHANGE_REQUEST_NUM", ""
-            ),
+            "gerrit_change_request_url": os.getenv("GERRIT_CHANGE_REQUEST_URL", ""),
+            "gerrit_change_request_num": os.getenv("GERRIT_CHANGE_REQUEST_NUM", ""),
         }
     )
 
@@ -469,30 +519,20 @@ def _process_single(data: Inputs, gh: GitHubContext) -> None:
         except Exception as exc:
             log.debug("Execution failed; continuing to write outputs: %s", exc)
 
-            result = SubmissionResult(
-                change_urls=[], change_numbers=[], commit_shas=[]
-            )
+            result = SubmissionResult(change_urls=[], change_numbers=[], commit_shas=[])
         if result.change_urls:
-            os.environ["GERRIT_CHANGE_REQUEST_URL"] = "\n".join(
-                result.change_urls
-            )
+            os.environ["GERRIT_CHANGE_REQUEST_URL"] = "\n".join(result.change_urls)
             # Output Gerrit change URL(s) to console
             for url in result.change_urls:
                 log.info("Gerrit change URL: %s", url)
         if result.change_numbers:
-            os.environ["GERRIT_CHANGE_REQUEST_NUM"] = "\n".join(
-                result.change_numbers
-            )
+            os.environ["GERRIT_CHANGE_REQUEST_NUM"] = "\n".join(result.change_numbers)
 
         # Also write outputs to GITHUB_OUTPUT if available
         _append_github_output(
             {
-                "gerrit_change_request_url": os.getenv(
-                    "GERRIT_CHANGE_REQUEST_URL", ""
-                ),
-                "gerrit_change_request_num": os.getenv(
-                    "GERRIT_CHANGE_REQUEST_NUM", ""
-                ),
+                "gerrit_change_request_url": os.getenv("GERRIT_CHANGE_REQUEST_URL", ""),
+                "gerrit_change_request_num": os.getenv("GERRIT_CHANGE_REQUEST_NUM", ""),
                 "gerrit_commit_sha": os.getenv("GERRIT_COMMIT_SHA", ""),
             }
         )
@@ -504,13 +544,9 @@ def _process_single(data: Inputs, gh: GitHubContext) -> None:
         return
 
 
-def _prepare_local_checkout(
-    workspace: Path, gh: GitHubContext, data: Inputs
-) -> None:
+def _prepare_local_checkout(workspace: Path, gh: GitHubContext, data: Inputs) -> None:
     repo_full = gh.repository.strip() if gh.repository else ""
-    server_url = gh.server_url or os.getenv(
-        "GITHUB_SERVER_URL", "https://github.com"
-    )
+    server_url = gh.server_url or os.getenv("GITHUB_SERVER_URL", "https://github.com")
     server_url = (server_url or "https://github.com").rstrip("/")
     base_ref = gh.base_ref or ""
     pr_num_str: str = str(gh.pr_number) if gh.pr_number else "0"
@@ -522,6 +558,26 @@ def _prepare_local_checkout(
     run_cmd(["git", "init"], cwd=workspace)
     run_cmd(["git", "remote", "add", "origin", repo_url], cwd=workspace)
 
+    # Non-interactive SSH/Git environment for any network operations
+    env = {
+        "GIT_SSH_COMMAND": (
+            "ssh -F /dev/null "
+            "-o IdentitiesOnly=yes "
+            "-o IdentityAgent=none "
+            "-o BatchMode=yes "
+            "-o PreferredAuthentications=publickey "
+            "-o StrictHostKeyChecking=yes "
+            "-o PasswordAuthentication=no "
+            "-o PubkeyAcceptedKeyTypes=+ssh-rsa "
+            "-o ConnectTimeout=10"
+        ),
+        "SSH_AUTH_SOCK": "",
+        "SSH_AGENT_PID": "",
+        "SSH_ASKPASS": "/usr/bin/false",
+        "DISPLAY": "",
+        "SSH_ASKPASS_REQUIRE": "never",
+    }
+
     # Fetch base branch and PR head
     if base_ref:
         try:
@@ -535,15 +591,13 @@ def _prepare_local_checkout(
                     branch_ref,
                 ],
                 cwd=workspace,
+                env=env,
             )
         except Exception as exc:
             log.debug("Base branch fetch failed for %s: %s", base_ref, exc)
 
     if pr_num_str:
-        pr_ref = (
-            f"refs/pull/{pr_num_str}/head:"
-            f"refs/remotes/origin/pr/{pr_num_str}/head"
-        )
+        pr_ref = f"refs/pull/{pr_num_str}/head:refs/remotes/origin/pr/{pr_num_str}/head"
         run_cmd(
             [
                 "git",
@@ -553,6 +607,7 @@ def _prepare_local_checkout(
                 pr_ref,
             ],
             cwd=workspace,
+            env=env,
         )
         run_cmd(
             [
@@ -563,6 +618,7 @@ def _prepare_local_checkout(
                 f"refs/remotes/origin/pr/{pr_num_str}/head",
             ],
             cwd=workspace,
+            env=env,
         )
 
 
@@ -571,21 +627,19 @@ def _load_effective_inputs() -> Inputs:
     data = _build_inputs_from_env()
 
     # Load per-org configuration and apply to environment before validation
-    org_for_cfg = (
-        data.organization
-        or os.getenv("ORGANIZATION")
-        or os.getenv("GITHUB_REPOSITORY_OWNER")
-    )
+    org_for_cfg = data.organization or os.getenv("ORGANIZATION") or os.getenv("GITHUB_REPOSITORY_OWNER")
     cfg = load_org_config(org_for_cfg)
+
+    # Apply dynamic parameter derivation for missing Gerrit parameters
+    cfg = apply_parameter_derivation(cfg, org_for_cfg, save_to_config=True)
+
     apply_config_to_env(cfg)
 
     # Refresh inputs after applying configuration to environment
     data = _build_inputs_from_env()
 
     # Derive reviewers from local git config if running locally and unset
-    if not os.getenv("REVIEWERS_EMAIL") and (
-        os.getenv("G2G_TARGET_URL") or not os.getenv("GITHUB_EVENT_NAME")
-    ):
+    if not os.getenv("REVIEWERS_EMAIL") and (os.getenv("G2G_TARGET_URL") or not os.getenv("GITHUB_EVENT_NAME")):
         try:
             from .gitutils import enumerate_reviewer_emails
 
@@ -609,6 +663,7 @@ def _load_effective_inputs() -> Inputs:
                     gerrit_project=data.gerrit_project,
                     issue_id=data.issue_id,
                     allow_duplicates=data.allow_duplicates,
+                    duplicates_filter=data.duplicates_filter,
                 )
                 log.info("Derived reviewers: %s", data.reviewers_email)
         except Exception as exc:
@@ -637,24 +692,14 @@ def _append_github_output(outputs: dict[str, str]) -> None:
 
 
 def _augment_pr_refs_if_needed(gh: GitHubContext) -> GitHubContext:
-    if (
-        os.getenv("G2G_TARGET_URL")
-        and gh.pr_number
-        and (not gh.head_ref or not gh.base_ref)
-    ):
+    if os.getenv("G2G_TARGET_URL") and gh.pr_number and (not gh.head_ref or not gh.base_ref):
         try:
             client = build_client()
             repo = get_repo_from_env(client)
             pr_obj = get_pull(repo, int(gh.pr_number))
-            base_ref = str(
-                getattr(getattr(pr_obj, "base", object()), "ref", "") or ""
-            )
-            head_ref = str(
-                getattr(getattr(pr_obj, "head", object()), "ref", "") or ""
-            )
-            head_sha = str(
-                getattr(getattr(pr_obj, "head", object()), "sha", "") or ""
-            )
+            base_ref = str(getattr(getattr(pr_obj, "base", object()), "ref", "") or "")
+            head_ref = str(getattr(getattr(pr_obj, "head", object()), "ref", "") or "")
+            head_sha = str(getattr(getattr(pr_obj, "head", object()), "sha", "") or "")
             if base_ref:
                 os.environ["GITHUB_BASE_REF"] = base_ref
                 log.info("Resolved base_ref via GitHub API: %s", base_ref)
@@ -677,7 +722,7 @@ def _process() -> None:
     try:
         _validate_inputs(data)
     except ConfigurationError as exc:
-        log.exception("Configuration validation failed")
+        _log_exception_conditionally(log, "Configuration validation failed")
         typer.echo(f"Configuration validation failed: {exc}", err=True)
         raise typer.Exit(code=2) from exc
 
@@ -692,21 +737,17 @@ def _process() -> None:
 
     # Bulk mode for URL/workflow_dispatch
     sync_all = _env_bool("SYNC_ALL_OPEN_PRS", False)
-    if sync_all and (
-        gh.event_name == "workflow_dispatch" or os.getenv("G2G_TARGET_URL")
-    ):
+    if sync_all and (gh.event_name == "workflow_dispatch" or os.getenv("G2G_TARGET_URL")):
         _process_bulk(data, gh)
         return
 
     if not gh.pr_number:
         log.error(
-            "PR_NUMBER is empty. This tool requires a valid pull request "
-            "context. Current event: %s",
+            "PR_NUMBER is empty. This tool requires a valid pull request context. Current event: %s",
             gh.event_name,
         )
         typer.echo(
-            "PR_NUMBER is empty. This tool requires a valid pull request "
-            f"context. Current event: {gh.event_name}",
+            f"PR_NUMBER is empty. This tool requires a valid pull request context. Current event: {gh.event_name}",
             err=True,
         )
         raise typer.Exit(code=2)
@@ -720,13 +761,16 @@ def _process() -> None:
     # Check for duplicates in single-PR mode (before workspace setup)
     if gh.pr_number and not _env_bool("SYNC_ALL_OPEN_PRS", False):
         try:
+            if data.duplicates_filter:
+                os.environ["DUPLICATES"] = data.duplicates_filter
             check_for_duplicates(gh, allow_duplicates=data.allow_duplicates)
         except DuplicateChangeError as exc:
-            log.exception("Duplicate change detected")
-            typer.echo(f"Error: {exc}", err=True)
-            typer.echo(
-                "Use --allow-duplicates to override this check.", err=True
+            _log_exception_conditionally(
+                log,
+                "Duplicate detection blocked submission for PR #%d",
+                gh.pr_number,
             )
+            log.info("Use --allow-duplicates to override this check.")
             raise typer.Exit(code=3) from exc
 
     _process_single(data, gh)
@@ -745,9 +789,7 @@ def _load_event(path: Path | None) -> dict[str, Any]:
     if not path or not path.exists():
         return {}
     try:
-        return cast(
-            dict[str, Any], json.loads(path.read_text(encoding="utf-8"))
-        )
+        return cast(dict[str, Any], json.loads(path.read_text(encoding="utf-8")))
     except Exception as exc:
         log.warning("Failed to parse GITHUB_EVENT_PATH: %s", exc)
         return {}
@@ -814,31 +856,79 @@ def _read_github_context() -> GitHubContext:
 
 def _validate_inputs(data: Inputs) -> None:
     if data.use_pr_as_commit and data.submit_single_commits:
-        msg = (
-            "USE_PR_AS_COMMIT and SUBMIT_SINGLE_COMMITS cannot be enabled at "
-            "the same time"
-        )
+        msg = "USE_PR_AS_COMMIT and SUBMIT_SINGLE_COMMITS cannot be enabled at the same time"
         raise ConfigurationError(msg)
 
-    # Presence checks for required fields used by existing action
-    for field_name in (
-        "gerrit_known_hosts",
-        "gerrit_ssh_privkey_g2g",
-        "gerrit_ssh_user_g2g",
-        "gerrit_ssh_user_g2g_email",
-    ):
+    # Context-aware validation: different requirements for GH Actions vs CLI
+    is_github_actions = _is_github_actions_context()
+
+    # SSH private key is always required
+    required_fields = ["gerrit_ssh_privkey_g2g"]
+
+    # Gerrit parameters can be derived in GH Actions if organization available
+    # In local CLI context, we're more strict about explicit configuration
+    if is_github_actions:
+        # In GitHub Actions: allow derivation if organization is available
+        if not data.organization:
+            required_fields.extend(
+                [
+                    "gerrit_ssh_user_g2g",
+                    "gerrit_ssh_user_g2g_email",
+                ]
+            )
+    else:
+        # In local CLI: require explicit values or organization + derivation
+        # This prevents unexpected behavior when running locally
+        missing_gerrit_params = [
+            field for field in ["gerrit_ssh_user_g2g", "gerrit_ssh_user_g2g_email"] if not getattr(data, field)
+        ]
+        if missing_gerrit_params:
+            if data.organization:
+                log.info(
+                    "Local CLI usage: Gerrit parameters can be derived from "
+                    "organization '%s'. Missing: %s. Consider setting "
+                    "G2G_ENABLE_DERIVATION=true to enable derivation.",
+                    data.organization,
+                    ", ".join(missing_gerrit_params),
+                )
+                # Allow derivation in local mode only if explicitly enabled
+                if not _env_bool("G2G_ENABLE_DERIVATION", False):
+                    required_fields.extend(missing_gerrit_params)
+            else:
+                required_fields.extend(missing_gerrit_params)
+
+    for field_name in required_fields:
         if not getattr(data, field_name):
             log.error("Missing required input: %s", field_name)
-            raise ConfigurationError(f"Missing required input: {field_name}")  # noqa: TRY003
+            if field_name in [
+                "gerrit_ssh_user_g2g",
+                "gerrit_ssh_user_g2g_email",
+            ]:
+                if data.organization:
+                    if is_github_actions:
+                        log.error(
+                            "These fields can be derived automatically from "
+                            "organization '%s' if G2G_ENABLE_DERIVATION=true",
+                            data.organization,
+                        )
+                    else:
+                        log.error(
+                            "These fields can be derived from organization '%s'",
+                            data.organization,
+                        )
+                        log.error("Set G2G_ENABLE_DERIVATION=true to enable")
+                else:
+                    log.error("These fields require either explicit values or an ORGANIZATION for derivation")
+            raise ConfigurationError(_MSG_MISSING_REQUIRED_INPUT.format(field_name=field_name))
 
     # Validate fetch depth is a positive integer
     if data.fetch_depth <= 0:
         log.error("Invalid FETCH_DEPTH: %s", data.fetch_depth)
-        raise ConfigurationError("FETCH_DEPTH must be a positive integer")  # noqa: TRY003
+        raise ConfigurationError(_MSG_INVALID_FETCH_DEPTH)
 
     # Validate Issue ID is a single line string if provided
     if data.issue_id and ("\n" in data.issue_id or "\r" in data.issue_id):
-        raise ConfigurationError("Issue ID must be single line")  # noqa: TRY003
+        raise ConfigurationError(_MSG_ISSUE_ID_MULTILINE)
 
 
 def _log_effective_config(data: Inputs, gh: GitHubContext) -> None:
@@ -848,10 +938,8 @@ def _log_effective_config(data: Inputs, gh: GitHubContext) -> None:
     log.info("  SUBMIT_SINGLE_COMMITS: %s", data.submit_single_commits)
     log.info("  USE_PR_AS_COMMIT: %s", data.use_pr_as_commit)
     log.info("  FETCH_DEPTH: %s", data.fetch_depth)
-    log.info(
-        "  GERRIT_KNOWN_HOSTS: %s",
-        "<provided>" if data.gerrit_known_hosts else "<missing>",
-    )
+    known_hosts_status = "<provided>" if data.gerrit_known_hosts else "<will auto-discover>"
+    log.info("  GERRIT_KNOWN_HOSTS: %s", known_hosts_status)
     log.info("  GERRIT_SSH_PRIVKEY_G2G: %s", safe_privkey)
     log.info("  GERRIT_SSH_USER_G2G: %s", data.gerrit_ssh_user_g2g)
     log.info("  GERRIT_SSH_USER_G2G_EMAIL: %s", data.gerrit_ssh_user_g2g_email)