tapps-agents 3.5.38__py3-none-any.whl → 3.5.40__py3-none-any.whl

tapps_agents/agents/reviewer/tools/scoped_mypy.py

@@ -0,0 +1,284 @@
+"""
+Scoped Mypy Executor - ENH-002-S2
+
+Runs mypy with file-level scoping for ~70% performance improvement.
+Uses --follow-imports=skip and --no-site-packages; filters results to target file only.
+"""
+
+from __future__ import annotations
+
+import asyncio
+import logging
+import re
+import subprocess  # nosec B404 - fixed args, no shell
+import sys
+import time
+from dataclasses import dataclass
+from pathlib import Path
+from typing import Any
+
+logger = logging.getLogger(__name__)
+
+
+class MypyTimeoutError(Exception):
+    """Mypy execution timed out."""
+
+    def __init__(self, timeout: int) -> None:
+        self.timeout = timeout
+        super().__init__(f"Mypy timed out after {timeout}s")
+
+
+@dataclass(frozen=True)
+class MypyIssue:
+    """Single mypy type checking issue."""
+
+    file_path: Path
+    line: int
+    column: int
+    severity: str
+    message: str
+    error_code: str | None
+
+    def to_dict(self) -> dict[str, Any]:
+        """Convert to dictionary for serialization."""
+        return {
+            "file_path": str(self.file_path),
+            "line": self.line,
+            "column": self.column,
+            "severity": self.severity,
+            "message": self.message,
+            "error_code": self.error_code,
+        }
+
+
+@dataclass(frozen=True)
+class MypyResult:
+    """Result of scoped mypy execution."""
+
+    issues: tuple[MypyIssue, ...]
+    duration_seconds: float
+    files_checked: int
+    success: bool
+
+    def to_dict(self) -> dict[str, Any]:
+        """Convert to dictionary for serialization."""
+        return {
+            "issues": [i.to_dict() for i in self.issues],
+            "duration_seconds": self.duration_seconds,
+            "files_checked": self.files_checked,
+            "success": self.success,
+        }
+
+
+@dataclass(frozen=True)
+class ScopedMypyConfig:
+    """Configuration for scoped mypy execution."""
+
+    enabled: bool = True
+    timeout: int = 10
+    flags: tuple[str, ...] = (
+        "--follow-imports=skip",
+        "--no-site-packages",
+        "--show-column-numbers",
+        "--show-error-codes",
+        "--no-error-summary",
+        "--no-color-output",
+        "--no-incremental",
+    )
+
+
+class ScopedMypyExecutor:
+    """
+    Execute mypy with scoped imports for performance.
+
+    Uses --follow-imports=skip, --no-site-packages and filters results
+    to the target file only. Target: <10s vs 30-60s unscoped.
+    """
+
+    def __init__(self, config: ScopedMypyConfig | None = None) -> None:
+        self.config = config or ScopedMypyConfig()
+        self._logger = logging.getLogger(__name__)
+
+    def get_scoped_flags(self) -> list[str]:
+        """Return mypy flags for scoped execution."""
+        return list(self.config.flags)
+
+    def parse_output(self, raw_output: str, file_path: Path) -> list[MypyIssue]:
+        """
+        Parse mypy stdout and return issues for the target file only.
+
+        Mypy format: file.py:line:col: severity: message [error-code]
+        or file.py:line: severity: message [error-code]
+        """
+        issues: list[MypyIssue] = []
+        target_name = file_path.name
+        target_resolved = str(file_path.resolve()).replace("\\", "/")
+
+        for line in raw_output.splitlines():
+            line = line.strip()
+            if not line or "error:" not in line.lower():
+                continue
+            # Match file:line:col: severity: message [code] or file:line: severity: message [code]
+            match = re.match(
+                r"^(.+?):(\d+):(?:\d+:)?\s*(error|warning|note):\s*(.+)$",
+                line,
+                re.IGNORECASE,
+            )
+            if not match:
+                continue
+            path_part, line_str, severity, rest = match.groups()
+            path_part = path_part.replace("\\", "/")
+            if target_name not in path_part and target_resolved not in path_part:
+                try:
+                    if Path(path_part).resolve() != file_path.resolve():
+                        continue
+                except Exception:
+                    continue
+            try:
+                line_num = int(line_str)
+            except ValueError:
+                continue
+            col_num = 0
+            if ":" in line:
+                col_match = re.match(r"^.+?:\d+:(\d+):", line)
+                if col_match:
+                    try:
+                        col_num = int(col_match.group(1))
+                    except ValueError:
+                        pass
+            error_code = None
+            if "[" in rest and "]" in rest:
+                start = rest.rfind("[")
+                end = rest.rfind("]")
+                if start < end:
+                    error_code = rest[start + 1 : end].strip()
+                    rest = rest[:start].strip()
+            issues.append(
+                MypyIssue(
+                    file_path=file_path,
+                    line=line_num,
+                    column=col_num,
+                    severity=severity.strip().lower(),
+                    message=rest.strip(),
+                    error_code=error_code,
+                )
+            )
+        return issues
+
+    async def execute_scoped(
+        self,
+        file_path: Path,
+        *,
+        timeout: int | None = None,
+    ) -> MypyResult:
+        """
+        Run mypy on a single file with scoped flags and filter to that file.
+
+        On timeout, returns a result with success=False and empty issues
+        (graceful fallback; caller can run full mypy if needed).
+        """
+        timeout_sec = timeout if timeout is not None else self.config.timeout
+        if not file_path.is_file():
+            return MypyResult(
+                issues=(), duration_seconds=0.0, files_checked=0, success=False
+            )
+        cmd = [sys.executable, "-m", "mypy"] + self.get_scoped_flags() + [str(file_path)]
+        cwd = file_path.parent if file_path.parent.exists() else None
+        start = time.monotonic()
+        try:
+            proc = await asyncio.create_subprocess_exec(
+                *cmd,
+                stdout=asyncio.subprocess.PIPE,
+                stderr=asyncio.subprocess.PIPE,
+                cwd=cwd,
+            )
+            try:
+                stdout, stderr = await asyncio.wait_for(
+                    proc.communicate(), timeout=timeout_sec
+                )
+            except TimeoutError:
+                proc.kill()
+                await proc.wait()
+                elapsed = time.monotonic() - start
+                self._logger.warning(
+                    "mypy timed out for %s after %.2fs", file_path, elapsed
+                )
+                raise MypyTimeoutError(timeout_sec)
+            elapsed = time.monotonic() - start
+            out = (stdout or b"").decode("utf-8", errors="replace")
+            issues = self.parse_output(out, file_path)
+            return MypyResult(
+                issues=tuple(issues),
+                duration_seconds=elapsed,
+                files_checked=1,
+                success=proc.returncode == 0 or not issues,
+            )
+        except MypyTimeoutError as e:
+            raise e from None
+        except FileNotFoundError:
+            self._logger.debug("mypy not found")
+            return MypyResult(
+                issues=(), duration_seconds=0.0, files_checked=0, success=False
+            )
+        except Exception as e:
+            self._logger.warning("mypy failed for %s: %s", file_path, e)
+            return MypyResult(
+                issues=(), duration_seconds=0.0, files_checked=0, success=False
+            )
+
+    def run_scoped_sync(
+        self,
+        file_path: Path,
+        *,
+        timeout: int | None = None,
+    ) -> MypyResult:
+        """
+        Run mypy synchronously with scoped flags (for use from scoring.py).
+        Uses subprocess.run to avoid event loop issues in sync callers.
+        """
+        timeout_sec = timeout if timeout is not None else self.config.timeout
+        if not file_path.is_file():
+            return MypyResult(
+                issues=(), duration_seconds=0.0, files_checked=0, success=False
+            )
+        cmd = [sys.executable, "-m", "mypy"] + self.get_scoped_flags() + [str(file_path)]
+        cwd = file_path.parent if file_path.parent.exists() else None
+        start = time.monotonic()
+        try:
+            result = subprocess.run(  # nosec B603
+                cmd,
+                capture_output=True,
+                text=True,
+                encoding="utf-8",
+                errors="replace",
+                timeout=timeout_sec,
+                cwd=cwd,
+            )
+            elapsed = time.monotonic() - start
+            out = (result.stdout or "").strip()
+            issues = self.parse_output(out, file_path)
+            return MypyResult(
+                issues=tuple(issues),
+                duration_seconds=elapsed,
+                files_checked=1,
+                success=result.returncode == 0 or not issues,
+            )
+        except subprocess.TimeoutExpired:
+            elapsed = time.monotonic() - start
+            self._logger.warning("mypy timed out for %s after %.2fs", file_path, elapsed)
+            return MypyResult(
+                issues=(),
+                duration_seconds=elapsed,
+                files_checked=0,
+                success=False,
+            )
+        except FileNotFoundError:
+            self._logger.debug("mypy not found")
+            return MypyResult(
+                issues=(), duration_seconds=0.0, files_checked=0, success=False
+            )
+        except Exception as e:
+            self._logger.warning("mypy failed for %s: %s", file_path, e)
+            return MypyResult(
+                issues=(), duration_seconds=0.0, files_checked=0, success=False
+            )

tapps_agents/beads/__init__.py

@@ -2,6 +2,7 @@
 Beads (bd) integration: optional task-tracking for agents.
 
 Use is_available(project_root) before run_bd. See docs/BEADS_INTEGRATION.md.
+Task specs: .tapps-agents/task-specs/ for hydration/dehydration.
 """
 
 from .client import (
@@ -13,6 +14,9 @@ from .client import (
     run_bd,
 )
 from .parse import parse_bd_id_from_stdout
+from .specs import TaskSpec, load_task_spec, load_task_specs, save_task_spec
+
+from .hydration import HydrationReport, dehydrate_from_beads, hydrate_to_beads
 
 __all__ = [
     "BeadsRequiredError",
@@ -22,4 +26,11 @@ __all__ = [
     "require_beads",
     "resolve_bd_path",
     "run_bd",
+    "TaskSpec",
+    "HydrationReport",
+    "dehydrate_from_beads",
+    "hydrate_to_beads",
+    "load_task_spec",
+    "load_task_specs",
+    "save_task_spec",
 ]

tapps_agents/beads/hydration.py

@@ -0,0 +1,213 @@
+"""
+Hydration engine: create Beads issues from task specs, update specs from Beads.
+
+Hydrate: for specs without beads_issue, run bd create and save id; recreate
+dependency graph with bd dep add. Dehydrate: run bd list, update spec files
+with current status. Handles missing bd gracefully (log, no crash).
+"""
+
+from __future__ import annotations
+
+import json
+import logging
+from dataclasses import dataclass
+from pathlib import Path
+
+from .client import is_available, run_bd
+from .parse import parse_bd_id_from_stdout
+from .specs import TaskSpec, load_task_specs, save_task_spec
+
+logger = logging.getLogger(__name__)
+
+
+@dataclass
+class HydrationReport:
+    """Summary of a hydration run."""
+
+    created: int = 0
+    skipped: int = 0
+    failed: int = 0
+    deps_added: int = 0
+    bd_unavailable: bool = False
+    dry_run: bool = False
+
+
+def hydrate_to_beads(
+    project_root: Path | None = None,
+    *,
+    dry_run: bool = False,
+) -> HydrationReport:
+    """
+    Create Beads issues for task specs that don't have beads_issue; recreate deps.
+
+    For each spec in .tapps-agents/task-specs/ without beads_issue, runs
+    bd create and stores the returned id in the spec file. Then runs
+    bd dep add for each dependency. If bd is not available, logs and returns
+    without raising.
+
+    Args:
+        project_root: Project root (default: cwd).
+        dry_run: If True, do not run bd or write files; report what would be done.
+
+    Returns:
+        HydrationReport with created/skipped/failed/deps_added counts.
+    """
+    project_root = project_root or Path.cwd()
+    report = HydrationReport(dry_run=dry_run)
+
+    if not is_available(project_root):
+        logger.warning("Hydration skipped: bd not available")
+        report.bd_unavailable = True
+        return report
+
+    specs = load_task_specs(project_root)
+    if not specs:
+        return report
+
+    # Map spec.id -> beads_issue id after creation
+    spec_to_bd: dict[str, str] = {}
+    for spec in specs:
+        if spec.beads_issue:
+            spec_to_bd[spec.id] = spec.beads_issue
+            report.skipped += 1
+            continue
+        if dry_run:
+            report.created += 1
+            continue
+        title = (spec.title or spec.id)[:200].strip()
+        desc = (spec.description or "")[:500].replace("\n", " ").strip()
+        args = ["create", title]
+        if desc:
+            args.extend(["-d", desc])
+        try:
+            r = run_bd(project_root, args)
+            if r.returncode != 0:
+                logger.warning("Hydration: bd create failed for %s: %s", spec.id, r.stderr)
+                report.failed += 1
+                continue
+            bd_id = parse_bd_id_from_stdout(r.stdout)
+            if bd_id:
+                spec_to_bd[spec.id] = bd_id
+                spec.beads_issue = bd_id
+                save_task_spec(spec, project_root)
+                report.created += 1
+            else:
+                logger.warning("Hydration: could not parse bd id for %s", spec.id)
+                report.failed += 1
+        except Exception as e:
+            logger.warning("Hydration: bd create failed for %s: %s", spec.id, e)
+            report.failed += 1
+
+    # Recreate dependency graph: bd dep add child parent (parent blocks child)
+    if dry_run:
+        for spec in specs:
+            if spec.dependencies and spec_to_bd.get(spec.id):
+                report.deps_added += len([d for d in spec.dependencies if spec_to_bd.get(d)])
+        return report
+
+    for spec in specs:
+        child_bd = spec_to_bd.get(spec.id)
+        if not child_bd or not spec.dependencies:
+            continue
+        for dep_id in spec.dependencies:
+            parent_bd = spec_to_bd.get(dep_id)
+            if not parent_bd:
+                continue
+            try:
+                r = run_bd(project_root, ["dep", "add", child_bd, parent_bd])
+                if r.returncode == 0:
+                    report.deps_added += 1
+                else:
+                    logger.warning(
+                        "Hydration: bd dep add %s %s failed: %s",
+                        child_bd,
+                        parent_bd,
+                        r.stderr,
+                    )
+            except Exception as e:
+                logger.warning("Hydration: bd dep add %s %s failed: %s", child_bd, parent_bd, e)
+
+    return report
+
+
+def dehydrate_from_beads(project_root: Path | None = None) -> int:
+    """
+    Update task spec files with current status from Beads.
+
+    Runs bd list (or equivalent), maps beads_issue id to status, and updates
+    each spec file. If bd is not available, logs and returns 0.
+
+    Args:
+        project_root: Project root (default: cwd).
+
+    Returns:
+        Number of spec files updated.
+    """
+    project_root = project_root or Path.cwd()
+
+    if not is_available(project_root):
+        logger.warning("Dehydration skipped: bd not available")
+        return 0
+
+    # Try bd list --json; fallback to parsing stdout if format differs
+    try:
+        r = run_bd(project_root, ["list", "--json"])
+    except Exception as e:
+        logger.warning("Dehydration: bd list failed: %s", e)
+        return 0
+
+    if r.returncode != 0:
+        logger.warning("Dehydration: bd list failed: %s", r.stderr)
+        return 0
+
+    bd_status_by_id: dict[str, str] = {}
+    try:
+        raw = json.loads(r.stdout or "[]")
+        if isinstance(raw, list):
+            for item in raw:
+                if isinstance(item, dict):
+                    bid = item.get("id") or item.get("bd_id") or item.get("issue_id")
+                    status = item.get("status") or item.get("state") or "todo"
+                    if bid:
+                        bd_status_by_id[str(bid)] = str(status).lower()
+        elif isinstance(raw, dict):
+            for bid, info in raw.items():
+                if isinstance(info, dict):
+                    status = info.get("status") or info.get("state") or "todo"
+                else:
+                    status = str(info)
+                bd_status_by_id[str(bid)] = str(status).lower()
+    except json.JSONDecodeError:
+        # Fallback: parse line-based output for "id status" or "id\tstatus"
+        for line in (r.stdout or "").splitlines():
+            parts = line.strip().split()
+            if len(parts) >= 2:
+                bd_status_by_id[parts[0]] = parts[1].lower()
+            elif len(parts) == 1:
+                bd_status_by_id[parts[0]] = "todo"
+
+    specs = load_task_specs(project_root)
+    updated = 0
+    for spec in specs:
+        if not spec.beads_issue:
+            continue
+        new_status = bd_status_by_id.get(spec.beads_issue)
+        if not new_status:
+            continue
+        # Map bd status to our status if needed
+        if new_status in ("todo", "open", "pending"):
+            mapped = "todo"
+        elif new_status in ("in-progress", "in_progress", "wip"):
+            mapped = "in-progress"
+        elif new_status in ("done", "closed", "completed"):
+            mapped = "done"
+        elif new_status == "blocked":
+            mapped = "blocked"
+        else:
+            mapped = spec.status
+        if spec.status != mapped:
+            spec = spec.model_copy(update={"status": mapped})
+            save_task_spec(spec, project_root)
+            updated += 1
+
+    return updated