monoco-toolkit 0.2.2__py3-none-any.whl → 0.2.4__py3-none-any.whl

monoco/features/issue/linter.py

@@ -1,68 +1,27 @@
-from typing import List, Optional, Tuple, Set
+from typing import List, Optional, Set, Tuple
 from pathlib import Path
 from rich.console import Console
 from rich.table import Table
 import typer
+import re
 
 from . import core
-from .models import IssueStatus, IssueStage
+from .validator import IssueValidator
+from monoco.core.lsp import Diagnostic, DiagnosticSeverity
 
 console = Console()
 
-
-def validate_issue(path: Path, meta: core.IssueMetadata, all_issue_ids: Set[str] = set(), issues_root: Optional[Path] = None) -> List[str]:
-    """
-    Validate a single issue's integrity.
+def check_integrity(issues_root: Path, recursive: bool = False) -> List[Diagnostic]:
     """
-    errors = []
-    
-    # A. Directory/Status Consistency
-    expected_status = meta.status.value
-    path_parts = path.parts
-    # We might be validating a temp file, so we skip path check if it's not in the tree?
-    # Or strict check? For "Safe Edit", the file might be in a temp dir. 
-    # So we probably only care about content/metadata integrity.
-    
-    # But wait, if we overwrite the file, it MUST be valid.
-    # Let's assume the validation is about the content itself (metadata logic).
-    
-    # B. Solution Compliance
-    if meta.status == IssueStatus.CLOSED and not meta.solution:
-        errors.append(f"[red]Solution Missing:[/red] {meta.id} is closed but has no [dim]solution[/dim] field.")
-        
-    # C. Link Integrity
-    if meta.parent:
-        if all_issue_ids and meta.parent not in all_issue_ids:
-             # Check workspace (fallback)
-             found = False
-             if issues_root:
-                 if core.find_issue_path(issues_root, meta.parent):
-                     found = True
-             
-             if not found:
-                 errors.append(f"[red]Broken Link:[/red] {meta.id} refers to non-existent parent [bold]{meta.parent}[/bold].")
-
-    # D. Lifecycle Guard (Backlog)
-    if meta.status == IssueStatus.BACKLOG and meta.stage != IssueStage.FREEZED:
-        errors.append(f"[red]Lifecycle Error:[/red] {meta.id} is backlog but stage is not [bold]freezed[/bold] (found: {meta.stage}).")
-
-    return errors
-
-def check_integrity(issues_root: Path, recursive: bool = False) -> List[str]:
+    Verify the integrity of the Issues directory using LSP Validator.
     """
-    Verify the integrity of the Issues directory.
-    Returns a list of error messages.
+    diagnostics = []
+    validator = IssueValidator(issues_root)
     
-    If recursive=True, performs workspace-level validation including:
-    - Cross-project ID collision detection
-    - Cross-project UID collision detection
-    """
-    errors = []
-    all_issue_ids = set()  # For parent reference validation (includes namespaced IDs)
-    id_to_projects = {}  # local_id -> [(project_name, meta, file)]
-    all_uids = {}  # uid -> (project, issue_id)
+    all_issue_ids = set()
     all_issues = []
     
+    # 1. Collection Phase (Build Index)
     # Helper to collect issues from a project
     def collect_project_issues(project_issues_root: Path, project_name: str = "local"):
         project_issues = []
@@ -81,92 +40,232 @@ def check_integrity(issues_root: Path, recursive: bool = False) -> List[str]:
                         local_id = meta.id
                         full_id = f"{project_name}::{local_id}" if project_name != "local" else local_id
                         
-                        # Track ID occurrences per project
-                        if local_id not in id_to_projects:
-                            id_to_projects[local_id] = []
-                        id_to_projects[local_id].append((project_name, meta, f))
-                        
-                        # Add IDs for reference validation
-                        all_issue_ids.add(local_id)  # Local ID
+                        all_issue_ids.add(local_id)
                         if project_name != "local":
-                            all_issue_ids.add(full_id)  # Namespaced ID
-                        
-                        # Check UID collision (if UID exists)
-                        if meta.uid:
-                            if meta.uid in all_uids:
-                                existing_project, existing_id = all_uids[meta.uid]
-                                errors.append(
-                                    f"[red]UID Collision:[/red] UID {meta.uid} is duplicated.\n"
-                                    f"  - {existing_project}::{existing_id}\n"
-                                    f"  - {project_name}::{local_id}"
-                                )
-                            else:
-                                all_uids[meta.uid] = (project_name, local_id)
+                            all_issue_ids.add(full_id)
                         
-                        project_issues.append((f, meta, project_name))
+                        project_issues.append((f, meta))
         return project_issues
 
-    # 1. Collect local issues
     all_issues.extend(collect_project_issues(issues_root, "local"))
-
-    # 2. If recursive, collect workspace member issues
+    
     if recursive:
         try:
             from monoco.core.config import get_config
             project_root = issues_root.parent
             conf = get_config(str(project_root))
-            
             for member_name, rel_path in conf.project.members.items():
                 member_root = (project_root / rel_path).resolve()
                 member_issues_dir = member_root / "Issues"
-                
                 if member_issues_dir.exists():
-                    all_issues.extend(collect_project_issues(member_issues_dir, member_name))
-        except Exception as e:
-            # Fail gracefully if workspace config is missing
+                    collect_project_issues(member_issues_dir, member_name)
+        except Exception:
             pass
 
-    # 3. Check for ID collisions within same project
-    for local_id, occurrences in id_to_projects.items():
-        # Group by project
-        projects_with_id = {}
-        for project_name, meta, f in occurrences:
-            if project_name not in projects_with_id:
-                projects_with_id[project_name] = []
-            projects_with_id[project_name].append((meta, f))
+    # 2. Validation Phase
+    for path, meta in all_issues:
+        content = path.read_text() # Re-read content for validation
         
-        # Check for duplicates within same project
-        for project_name, metas in projects_with_id.items():
-            if len(metas) > 1:
-                # Same ID appears multiple times in same project - this is an error
-                error_msg = f"[red]ID Collision:[/red] {local_id} appears {len(metas)} times in project '{project_name}':\n"
-                for idx, (meta, f) in enumerate(metas, 1):
-                    error_msg += f"  {idx}. uid: {meta.uid or 'N/A'} | created: {meta.created_at} | stage: {meta.stage} | status: {meta.status.value}\n"
-                error_msg += f"  [yellow]→ Action:[/yellow] Remove duplicate or use 'monoco issue move --to <target> --renumber' to resolve."
-                errors.append(error_msg)
-
-    # 4. Validation
-    for path, meta, project_name in all_issues:
-        # A. Directory/Status Consistency (Only check this for files in the tree)
-        expected_status = meta.status.value
-        path_parts = path.parts
-        if expected_status not in path_parts:
-             errors.append(f"[yellow]Placement Error:[/yellow] {meta.id} has status [cyan]{expected_status}[/cyan] but is not under a [dim]{expected_status}/[/dim] directory.")
+        # A. Run Core Validator
+        file_diagnostics = validator.validate(meta, content, all_issue_ids)
         
-        # Reuse common logic
-        errors.extend(validate_issue(path, meta, all_issue_ids, issues_root))
-
-    return errors
+        # Add context to diagnostics (Path)
+        for d in file_diagnostics:
+             d.source = f"{meta.id}" # Use ID as source context
+             d.data = {'path': path} # Attach path for potential fixers
+             diagnostics.append(d)
+             
+    return diagnostics
 
 
-def run_lint(issues_root: Path, recursive: bool = False):
-    errors = check_integrity(issues_root, recursive)
+def run_lint(issues_root: Path, recursive: bool = False, fix: bool = False, format: str = "table", file_path: Optional[str] = None):
+    """
+    Run lint with optional auto-fix and format selection.
+    
+    Args:
+        issues_root: Root directory of issues
+        recursive: Recursively scan workspace members
+        fix: Apply auto-fixes
+        format: Output format (table, json)
+        file_path: Optional path to a single file to validate (LSP mode)
+    """
+    # Single-file mode (for LSP integration)
+    if file_path:
+        file = Path(file_path).resolve()
+        if not file.exists():
+            console.print(f"[red]Error:[/red] File not found: {file_path}")
+            raise typer.Exit(code=1)
+        
+        # Parse and validate single file
+        try:
+            meta = core.parse_issue(file)
+            if not meta:
+                console.print(f"[red]Error:[/red] Failed to parse issue metadata from {file_path}")
+                raise typer.Exit(code=1)
+            
+            content = file.read_text()
+            validator = IssueValidator(issues_root)
+            
+            # For single-file mode, we need to build a minimal index
+            # We'll scan the entire workspace to get all issue IDs for reference validation
+            all_issue_ids = set()
+            for subdir in ["Epics", "Features", "Chores", "Fixes"]:
+                d = issues_root / subdir
+                if d.exists():
+                    for status in ["open", "closed", "backlog"]:
+                        status_dir = d / status
+                        if status_dir.exists():
+                            for f in status_dir.rglob("*.md"):
+                                try:
+                                    m = core.parse_issue(f)
+                                    if m:
+                                        all_issue_ids.add(m.id)
+                                except Exception:
+                                    pass
+            
+            diagnostics = validator.validate(meta, content, all_issue_ids)
+            
+            # Add context
+            for d in diagnostics:
+                d.source = meta.id
+                d.data = {'path': file}
+                
+        except Exception as e:
+            console.print(f"[red]Error:[/red] Validation failed: {e}")
+            raise typer.Exit(code=1)
+    else:
+        # Full workspace scan mode
+        diagnostics = check_integrity(issues_root, recursive)
+    
+    # Filter only Warnings and Errors
+    issues = [d for d in diagnostics if d.severity <= DiagnosticSeverity.Warning]
     
-    if not errors:
+    if fix:
+        fixed_count = 0
+        console.print("[dim]Attempting auto-fixes...[/dim]")
+        
+        # We must track processed paths to avoid redundant writes if multiple errors exist
+        processed_paths = set()
+        
+        for d in issues:
+            path = d.data.get('path')
+            if not path: continue
+            
+            # Read fresh content iteration
+            pass
+
+        # Group diagnostics by file path
+        from collections import defaultdict
+        file_diags = defaultdict(list)
+        for d in issues:
+            if d.data.get('path'):
+                file_diags[d.data['path']].append(d)
+        
+        for path, diags in file_diags.items():
+            try:
+                content = path.read_text()
+                new_content = content
+                has_changes = False
+                
+                # Parse meta once for the file
+                try:
+                    meta = core.parse_issue(path)
+                except Exception:
+                    console.print(f"[yellow]Skipping fix for {path.name}: Cannot parse metadata[/yellow]")
+                    continue
+
+                # Apply fixes for this file
+                for d in diags:
+                    if "Structure Error" in d.message:
+                        expected_header = f"## {meta.id}: {meta.title}"
+                        
+                        # Check if strictly present
+                        if expected_header in new_content:
+                            continue
+                            
+                        # Strategy: Look for existing heading with same ID to replace
+                        # Matches: "## ID..." or "## ID ..."
+                        # Regex: ^##\s+ID\b.*$
+                        # We use meta.id which is safe.
+                        heading_regex = re.compile(rf"^##\s+{re.escape(meta.id)}.*$", re.MULTILINE)
+                        
+                        match_existing = heading_regex.search(new_content)
+                        
+                        if match_existing:
+                            # Replace existing incorrect heading
+                            # We use sub to replace just the first occurrence
+                            new_content = heading_regex.sub(expected_header, new_content, count=1)
+                            has_changes = True
+                        else:
+                            # Insert after frontmatter
+                            fm_match = re.search(r"^---(.*?)---", new_content, re.DOTALL | re.MULTILINE)
+                            if fm_match:
+                                end_pos = fm_match.end()
+                                header_block = f"\n\n{expected_header}\n"
+                                new_content = new_content[:end_pos] + header_block + new_content[end_pos:].lstrip()
+                                has_changes = True
+
+                    if "Review Requirement: Missing '## Review Comments' section" in d.message:
+                         if "## Review Comments" not in new_content:
+                             new_content = new_content.rstrip() + "\n\n## Review Comments\n\n- [ ] Self-Review\n"
+                             has_changes = True
+
+                if has_changes:
+                    path.write_text(new_content)
+                    fixed_count += 1
+                    console.print(f"[dim]Fixed: {path.name}[/dim]")
+            except Exception as e:
+                console.print(f"[red]Failed to fix {path.name}: {e}[/red]")
+
+        console.print(f"[green]Applied auto-fixes to {fixed_count} files.[/green]")
+        
+        # Re-run validation to verify
+        if file_path:
+            # Re-validate single file
+            file = Path(file_path).resolve()
+            meta = core.parse_issue(file)
+            content = file.read_text()
+            validator = IssueValidator(issues_root)
+            diagnostics = validator.validate(meta, content, all_issue_ids)
+            for d in diagnostics:
+                d.source = meta.id
+                d.data = {'path': file}
+        else:
+            diagnostics = check_integrity(issues_root, recursive)
+        issues = [d for d in diagnostics if d.severity <= DiagnosticSeverity.Warning]
+
+    # Output formatting
+    if format == "json":
+        import json
+        from pydantic import RootModel
+        # Use RootModel to export a list of models
+        print(RootModel(issues).model_dump_json(indent=2))
+        if any(d.severity == DiagnosticSeverity.Error for d in issues):
+            raise typer.Exit(code=1)
+        return
+
+    if not issues:
         console.print("[green]✔[/green] Issue integrity check passed. No integrity errors found.")
     else:
-        table = Table(title="Issue Integrity Issues", show_header=False, border_style="red")
-        for err in errors:
-            table.add_row(err)
+        table = Table(title="Issue Integrity Report", show_header=True, header_style="bold magenta", border_style="red")
+        table.add_column("Issue", style="cyan")
+        table.add_column("Severity", justify="center")
+        table.add_column("Line", justify="right", style="dim")
+        table.add_column("Message")
+        
+        for d in issues:
+            sev_style = "red" if d.severity == DiagnosticSeverity.Error else "yellow"
+            sev_label = "ERROR" if d.severity == DiagnosticSeverity.Error else "WARN"
+            line_str = str(d.range.start.line + 1) if d.range else "-"
+            table.add_row(
+                d.source or "Unknown",
+                f"[{sev_style}]{sev_label}[/{sev_style}]",
+                line_str,
+                d.message
+            )
+            
         console.print(table)
-        raise typer.Exit(code=1)
+        
+        if any(d.severity == DiagnosticSeverity.Error for d in issues):
+            raise typer.Exit(code=1)
+

monoco/features/issue/models.py

@@ -147,22 +147,9 @@ class IssueMetadata(BaseModel):
         # status: backlog -> stage: freezed
         # status: closed -> stage: done
         # status: open -> stage: draft | doing | review | done (default draft)
-
-        if self.status == IssueStatus.BACKLOG:
-            self.stage = IssueStage.FREEZED
-        
-        elif self.status == IssueStatus.CLOSED:
-            # Enforce stage=done for closed issues
-            if self.stage != IssueStage.DONE:
-                self.stage = IssueStage.DONE
-            # Auto-fill closed_at if missing
-            if not self.closed_at:
-                self.closed_at = current_time()
         
-        elif self.status == IssueStatus.OPEN:
-            # Ensure valid stage for open status
-            if self.stage is None:
-                self.stage = IssueStage.DRAFT
+        # NOTE: We do NOT auto-correct state here anymore to allow Linter to detect inconsistencies.
+        # Auto-correction should be applied explicitly by 'create' or 'update' commands via core logic.
         
         return self
 

monoco/features/issue/resources/en/AGENTS.md

@@ -1,4 +1,6 @@
-### Issue Management
+# Issue Management (Agent Guidance)
+
+## Issue Management
 
 System for managing tasks using `monoco issue`.
 
@@ -7,3 +9,7 @@ System for managing tasks using `monoco issue`.
 - **Check**: `monoco issue lint` (Must run after manual edits)
 - **Lifecycle**: `monoco issue start|submit|delete <id>`
 - **Structure**: `Issues/{CapitalizedPluralType}/{lowercase_status}/` (e.g. `Issues/Features/open/`). Do not deviate.
+- **Rules**:
+  1. **Heading**: Must have `## {ID}: {Title}` (matches metadata).
+  2. **Checkboxes**: Min 2 using `- [ ]`, `- [x]`, `- [-]`, `- [/]`.
+  3. **Review**: `## Review Comments` section required for Review/Done stages.

monoco/features/issue/resources/en/SKILL.md

@@ -25,12 +25,20 @@ Use this skill to create and manage **Issues** (Universal Atoms) in Monoco proje
 
 ## Guidelines
 
-### Directory Structure
+### Directory Structure & Naming
 
 `Issues/{CapitalizedPluralType}/{lowercase_status}/`
 
-- `{TYPE}`: `Epics`, `Features`, `Chores`, `Fixes`
-- `{STATUS}`: `open`, `backlog`, `closed`
+- **Types**: `Epics`, `Features`, `Chores`, `Fixes`
+- **Statuses**: `open`, `backlog`, `closed`
+
+### Structural Integrity
+
+Issues are validated via `monoco issue lint`. key constraints:
+
+1. **Mandatory Heading**: `## {ID}: {Title}` must match front matter.
+2. **Min Checkboxes**: At least 2 checkboxes (AC/Tasks).
+3. **Review Protocol**: `## Review Comments` required for `review` or `done` stages.
 
 ### Path Transitions
 
@@ -49,3 +57,31 @@ Use `monoco issue`:
 5. **Modification**: `monoco issue start/submit/delete <id>`
 
 6. **Commit**: `monoco issue commit` (Atomic commit for issue files)
+7. **Validation**: `monoco issue lint` (Enforces compliance)
+
+## Validation Rules (FEAT-0082)
+
+To ensure data integrity, all Issue tickets must follow these strict rules:
+
+### 1. Structural Consistency
+
+- Must contain a Level 2 Heading matching exactly: `## {ID}: {Title}`.
+- Example: `## FEAT-0082: Issue Ticket Validator`
+
+### 2. Content Completeness
+
+- **Checkboxes**: Minimum of 2 checkboxes required (one for AC, one for Tasks).
+- **Review Comments**: If `stage` is `review` or `done`, a `## Review Comments` section is mandatory and must not be empty.
+
+### 3. Checkbox Syntax & Hierarchy
+
+- Use only `- [ ]`, `- [x]`, `- [-]`, or `- [/]`.
+- **Inheritance**: If nested checkboxes exist, the parent state must reflect child states (e.g., if any child is `[/]`, parent must be `[/]`; if all children are `[x]`, parent must be `[x]`).
+
+### 4. State Matrix
+
+The `status` (folder) and `stage` (front matter) must be compatible:
+
+- **open**: Draft, Doing, Review, Done
+- **backlog**: Draft, Doing, Review
+- **closed**: Done

monoco/features/issue/resources/zh/AGENTS.md

@@ -1,4 +1,6 @@
-### Issue 管理
+# Issue 管理 (Agent 指引)
+
+## Issue 管理
 
 使用 `monoco issue` 管理任务的系统。
 
@@ -6,4 +8,8 @@
 - **状态**: `monoco issue open|close|backlog <id>`
 - **检查**: `monoco issue lint` (手动编辑后必须运行)
 - **生命周期**: `monoco issue start|submit|delete <id>`
-- **结构**: `Issues/{CapitalizedPluralType}/{lowercase_status}/` (例如 `Issues/Features/open/`)。不要偏离此结构。
+- **结构**: `Issues/{CapitalizedPluralType}/{lowercase_status}/` (如 `Issues/Features/open/`)。
+- **强制规则**:
+  1. **标题**: 必须包含 `## {ID}: {Title}` 标题（与 Front Matter 一致）。
+  2. **内容**: 至少 2 个 Checkbox，使用 `- [ ]`, `- [x]`, `- [-]`, `- [/]`。
+  3. **评审**: `review`/`done` 阶段必须包含 `## Review Comments` 章节且内容不为空。

monoco/features/issue/resources/zh/SKILL.md

@@ -38,7 +38,7 @@ Monoco 不仅仅复刻 Jira，而是基于 **"思维模式 (Mindset)"** 重新
 - **Focus**: "How" (为了支撑系统运转，必须做什么)。
 - **Prefix**: `CHORE-`
 
-_(取代了 Task 概念)_
+> 注：取代了传统的 Task 概念。
 
 #### 🐞 FIX (修复)
 
@@ -47,7 +47,7 @@ _(取代了 Task 概念)_
 - **Focus**: "Fix" (恢复原状)。
 - **Prefix**: `FIX-`
 
-_(取代了 Bug 概念)_
+> 注：取代了传统的 Bug 概念。
 
 ---
 
@@ -82,4 +82,33 @@ _(取代了 Bug 概念)_
 
 5. **Modification**: `monoco issue start/submit/delete <id>`
 
-6. **Commit**: `monoco issue commit` (Atomic commit for issue files)
+6. **Commit**: `monoco issue commit` (原子化提交 Issue 文件)
+7. **Validation**: `monoco issue lint` (强制执行合规性检查)
+
+## 合规与结构校验 (Validation Rules)
+
+为了确保数据严谨性，所有 Issue Ticket 必须遵循以下强制规则：
+
+### 1. 结构一致性 (Structural Consistency)
+
+- 必须包含一个二级标题 (`##`)，内容必须与 Front Matter 中的 ID 和 Title 严格匹配。
+- 格式：`## {ID}: {Title}`
+- 示例：`## FEAT-0082: Issue Ticket Validator`
+
+### 2. 内容完整性 (Content Completeness)
+
+- **Checkbox 数量**: 每个 Ticket 必须包含至少 2 个 Checkbox（通常代表 AC 和 Tasks）。
+- **评审记录**: 当 `stage` 为 `review` 或 `done` 时，必须包含 `## Review Comments` 标题且内容不能为空。
+
+### 3. Checkbox 语法与层级 (Checkbox Matrix)
+
+- 仅限使用: `- [ ]`, `- [x]`, `- [-]`, `- [/]`。
+- **层级继承**: 若存在嵌套 Checkbox，父项状态必须正确反映子项的聚合结果（例如：任一子项为 `[/]` 则父项必为 `[/]`；子项全选则父项为 `[x]`）。
+
+### 4. 状态矩阵 (State Matrix)
+
+`status` (物理存放目录) 与 `stage` (Front Matter 字段) 必须兼容：
+
+- **open**: Draft, Doing, Review, Done
+- **backlog**: Draft, Doing, Review
+- **closed**: Done