emdash-core 0.1.7__py3-none-any.whl

emdash_core/core/models.py

@@ -0,0 +1,265 @@
+"""Data models for EmDash entities."""
+
+from dataclasses import dataclass, field
+from datetime import datetime
+from pathlib import Path
+from typing import Optional
+
+
+# ============================================================================
+# Layer A: Code Structure Entities
+# ============================================================================
+
+
+@dataclass
+class FileEntity:
+    """Represents a source code file."""
+
+    path: str
+    name: str
+    extension: str
+    size_bytes: int
+    lines_of_code: int
+    hash: str
+    last_modified: Optional[datetime] = None
+
+    @classmethod
+    def from_path(cls, file_path: Path, content_hash: str) -> "FileEntity":
+        """Create a FileEntity from a file path."""
+        stat = file_path.stat()
+        with open(file_path, 'r', encoding='utf-8', errors='ignore') as f:
+            lines = len(f.readlines())
+
+        return cls(
+            path=str(file_path),
+            name=file_path.name,
+            extension=file_path.suffix,
+            size_bytes=stat.st_size,
+            lines_of_code=lines,
+            hash=content_hash,
+            last_modified=datetime.fromtimestamp(stat.st_mtime),
+        )
+
+
+@dataclass
+class ClassEntity:
+    """Represents a class definition."""
+
+    name: str
+    qualified_name: str  # e.g., "module.ClassName"
+    file_path: str
+    line_start: int
+    line_end: int
+    docstring: Optional[str] = None
+    is_abstract: bool = False
+    decorators: list[str] = field(default_factory=list)
+    base_classes: list[str] = field(default_factory=list)  # For inheritance
+    attributes: list[str] = field(default_factory=list)
+    methods: list[str] = field(default_factory=list)  # Qualified names of methods
+
+
+@dataclass
+class FunctionEntity:
+    """Represents a function or method definition."""
+
+    name: str
+    qualified_name: str  # e.g., "module.function" or "module.Class.method"
+    file_path: str
+    line_start: int
+    line_end: int
+    docstring: Optional[str] = None
+    parameters: list[str] = field(default_factory=list)
+    return_annotation: Optional[str] = None
+    is_async: bool = False
+    is_method: bool = False
+    is_static: bool = False
+    is_classmethod: bool = False
+    decorators: list[str] = field(default_factory=list)
+    cyclomatic_complexity: int = 1
+    calls: list[str] = field(default_factory=list)  # Qualified names of called functions
+
+
+@dataclass
+class ModuleEntity:
+    """Represents a Python module (imported)."""
+
+    name: str
+    import_path: str
+    is_external: bool = False  # stdlib or third-party
+    package: Optional[str] = None
+
+
+@dataclass
+class ImportStatement:
+    """Represents an import statement."""
+
+    file_path: str
+    line_number: int
+    module: str
+    imported_names: list[str] = field(default_factory=list)
+    alias: Optional[str] = None
+    import_type: str = "import"  # "import" or "from_import"
+
+
+# ============================================================================
+# Layer B: Git History Entities
+# ============================================================================
+
+
+@dataclass
+class CommitEntity:
+    """Represents a git commit."""
+
+    sha: str
+    message: str
+    timestamp: datetime
+    author_name: str
+    author_email: str
+    committer_name: str
+    committer_email: str
+    insertions: int = 0
+    deletions: int = 0
+    files_changed: int = 0
+    is_merge: bool = False
+    parent_shas: list[str] = field(default_factory=list)
+
+
+@dataclass
+class FileModification:
+    """Represents a file modification in a commit."""
+
+    commit_sha: str
+    file_path: str
+    change_type: str  # "added", "modified", "deleted", "renamed"
+    insertions: int = 0
+    deletions: int = 0
+    old_path: Optional[str] = None  # For renames
+
+
+@dataclass
+class AuthorEntity:
+    """Represents a code contributor."""
+
+    email: str
+    name: str
+    first_commit: Optional[datetime] = None
+    last_commit: Optional[datetime] = None
+    total_commits: int = 0
+    total_lines_added: int = 0
+    total_lines_deleted: int = 0
+
+
+@dataclass
+class PullRequestEntity:
+    """Represents a GitHub pull request."""
+
+    number: int
+    title: str
+    state: str  # "open", "closed", "merged"
+    created_at: datetime
+    author: str
+    description: Optional[str] = None  # Full PR body text
+    merged_at: Optional[datetime] = None
+    reviewers: list[str] = field(default_factory=list)
+    labels: list[str] = field(default_factory=list)
+    additions: int = 0
+    deletions: int = 0
+    files_changed: list[str] = field(default_factory=list)  # File paths modified
+    commit_shas: list[str] = field(default_factory=list)
+    base_branch: str = "main"
+    head_branch: str = ""
+    embedding: Optional[list[float]] = None  # Vector embedding for similarity
+
+
+@dataclass
+class TaskEntity:
+    """Represents a subtask extracted from a PR description."""
+
+    id: str  # Format: "pr_{number}_task_{index}"
+    pr_number: int
+    description: str
+    is_completed: bool
+    order: int  # Position in the task list
+
+
+@dataclass
+class RepositoryEntity:
+    """Represents a repository."""
+
+    url: str
+    name: str
+    owner: Optional[str] = None
+    default_branch: str = "main"
+    last_ingested: Optional[datetime] = None
+    ingestion_status: str = "pending"
+    commit_count: int = 0
+    file_count: int = 0
+    primary_language: str = "Python"
+
+
+# ============================================================================
+# Layer C: Analytics Entities
+# ============================================================================
+
+
+@dataclass
+class ClusterEntity:
+    """Represents a detected code cluster/community."""
+
+    cluster_id: int
+    algorithm: str  # "louvain", "label_propagation"
+    modularity_score: float
+    size: int  # Number of nodes in cluster
+    members: list[str] = field(default_factory=list)  # Qualified names
+
+
+# ============================================================================
+# Aggregate Models
+# ============================================================================
+
+
+@dataclass
+class CodebaseEntities:
+    """Aggregate of all entities extracted from a codebase."""
+
+    files: list[FileEntity] = field(default_factory=list)
+    classes: list[ClassEntity] = field(default_factory=list)
+    functions: list[FunctionEntity] = field(default_factory=list)
+    modules: list[ModuleEntity] = field(default_factory=list)
+    imports: list[ImportStatement] = field(default_factory=list)
+
+    @classmethod
+    def merge(cls, entities_list: list["FileEntities"]) -> "CodebaseEntities":
+        """Merge multiple FileEntities into a single CodebaseEntities."""
+        result = cls()
+        for file_entities in entities_list:
+            if file_entities:
+                if file_entities.file:
+                    result.files.append(file_entities.file)
+                result.classes.extend(file_entities.classes)
+                result.functions.extend(file_entities.functions)
+                result.modules.extend(file_entities.modules)
+                result.imports.extend(file_entities.imports)
+        return result
+
+
+@dataclass
+class FileEntities:
+    """Entities extracted from a single file."""
+
+    file: Optional[FileEntity] = None
+    classes: list[ClassEntity] = field(default_factory=list)
+    functions: list[FunctionEntity] = field(default_factory=list)
+    modules: list[ModuleEntity] = field(default_factory=list)
+    imports: list[ImportStatement] = field(default_factory=list)
+
+
+@dataclass
+class GitData:
+    """Aggregate of all Git history data."""
+
+    repository: RepositoryEntity
+    commits: list[CommitEntity] = field(default_factory=list)
+    modifications: list[FileModification] = field(default_factory=list)
+    authors: list[AuthorEntity] = field(default_factory=list)
+    pull_requests: list[PullRequestEntity] = field(default_factory=list)

emdash_core/core/review_config.py

@@ -0,0 +1,57 @@
+"""Configuration loader for PR review settings."""
+
+from __future__ import annotations
+
+import json
+import os
+from pathlib import Path
+from typing import Optional
+
+from pydantic import BaseModel, Field
+
+
+class ReviewConfig(BaseModel):
+    """Configuration for PR review command."""
+
+    max_files: int = Field(default=50, ge=1, le=500)
+    max_tokens: int = Field(default=50000, ge=1000, le=200000)
+
+
+def _find_config_path() -> Optional[Path]:
+    """Find the review config file path."""
+    env_path = os.getenv("EMDASH_CONFIG_PATH")
+    if env_path:
+        return Path(env_path)
+
+    cwd = Path.cwd()
+    candidates = [
+        cwd / "emdash.config.json",
+        cwd / ".emdash-rules" / "config.json",
+    ]
+    for path in candidates:
+        if path.is_file():
+            return path
+
+    return None
+
+
+def load_review_config() -> ReviewConfig:
+    """Load review configuration from file if present."""
+    path = _find_config_path()
+    if not path or not path.is_file():
+        return ReviewConfig()
+
+    try:
+        with open(path, "r", encoding="utf-8") as handle:
+            raw = json.load(handle)
+    except (json.JSONDecodeError, OSError):
+        return ReviewConfig()
+
+    review = raw.get("review", {}) if isinstance(raw, dict) else {}
+    if not isinstance(review, dict):
+        return ReviewConfig()
+
+    return ReviewConfig(
+        max_files=review.get("max_files", 50),
+        max_tokens=review.get("max_tokens", 50000),
+    )

emdash_core/db/__init__.py

@@ -0,0 +1,67 @@
+"""Database layer for EmDash.
+
+Provides an abstraction over database backends with Supabase as the default implementation.
+"""
+
+import os
+from typing import Optional
+
+from .models import (
+    Feature,
+    FeatureAssignee,
+    FeaturePR,
+    FeatureStatus,
+    PRStatus,
+    Project,
+    TeamMember,
+)
+from .provider import DatabaseProvider
+from .providers.supabase import SupabaseProvider
+from .auth import AuthProvider, User, get_auth
+
+__all__ = [
+    # Models
+    "Feature",
+    "FeatureAssignee",
+    "FeaturePR",
+    "FeatureStatus",
+    "PRStatus",
+    "Project",
+    "TeamMember",
+    "User",
+    # Provider
+    "DatabaseProvider",
+    "SupabaseProvider",
+    # Auth
+    "AuthProvider",
+    "get_auth",
+    # Factory
+    "get_provider",
+]
+
+_provider_instance: Optional[DatabaseProvider] = None
+
+
+def get_provider() -> DatabaseProvider:
+    """Get the configured database provider instance.
+
+    Returns a singleton instance of the database provider.
+    Currently only supports Supabase.
+
+    Returns:
+        DatabaseProvider instance
+
+    Raises:
+        ValueError: If required environment variables are not set
+    """
+    global _provider_instance
+
+    if _provider_instance is None:
+        provider_type = os.getenv("EMDASH_DB_PROVIDER", "supabase")
+
+        if provider_type == "supabase":
+            _provider_instance = SupabaseProvider()
+        else:
+            raise ValueError(f"Unknown database provider: {provider_type}")
+
+    return _provider_instance

emdash_core/db/auth.py

@@ -0,0 +1,134 @@
+"""Supabase authentication with GitHub OAuth."""
+
+import os
+from dataclasses import dataclass
+from typing import Optional
+
+from supabase import create_client, Client
+
+
+@dataclass
+class User:
+    """Authenticated user from Supabase Auth."""
+
+    id: str
+    email: Optional[str]
+    github_handle: Optional[str]
+    avatar_url: Optional[str]
+    access_token: Optional[str] = None
+
+
+class AuthProvider:
+    """Handle Supabase authentication with GitHub OAuth."""
+
+    def __init__(self, url: Optional[str] = None, key: Optional[str] = None):
+        self.url = url or os.getenv("SUPABASE_URL")
+        self.key = key or os.getenv("SUPABASE_KEY")
+
+        if not self.url or not self.key:
+            raise ValueError("SUPABASE_URL and SUPABASE_KEY must be set")
+
+        self.client: Client = create_client(self.url, self.key)
+
+    def get_github_login_url(self, redirect_to: Optional[str] = None) -> str:
+        """Get the GitHub OAuth login URL.
+
+        Args:
+            redirect_to: URL to redirect after login (for web apps)
+
+        Returns:
+            GitHub OAuth URL to redirect user to
+        """
+        options = {}
+        if redirect_to:
+            options["redirect_to"] = redirect_to
+
+        response = self.client.auth.sign_in_with_oauth(
+            {"provider": "github", "options": options}
+        )
+        return response.url
+
+    def sign_in_with_github_token(self, access_token: str) -> Optional[User]:
+        """Sign in with an existing GitHub access token.
+
+        Args:
+            access_token: GitHub OAuth access token
+
+        Returns:
+            User if successful, None otherwise
+        """
+        try:
+            response = self.client.auth.sign_in_with_id_token(
+                {"provider": "github", "token": access_token}
+            )
+            return self._session_to_user(response)
+        except Exception:
+            return None
+
+    def get_session(self) -> Optional[User]:
+        """Get the current authenticated user from session.
+
+        Returns:
+            User if authenticated, None otherwise
+        """
+        try:
+            response = self.client.auth.get_session()
+            if response and response.user:
+                return self._response_to_user(response)
+            return None
+        except Exception:
+            return None
+
+    def sign_out(self) -> bool:
+        """Sign out the current user.
+
+        Returns:
+            True if successful
+        """
+        try:
+            self.client.auth.sign_out()
+            return True
+        except Exception:
+            return False
+
+    def _session_to_user(self, session) -> Optional[User]:
+        """Convert Supabase session to User model."""
+        if not session or not session.user:
+            return None
+
+        user_meta = session.user.user_metadata or {}
+
+        return User(
+            id=session.user.id,
+            email=session.user.email,
+            github_handle=user_meta.get("user_name") or user_meta.get("preferred_username"),
+            avatar_url=user_meta.get("avatar_url"),
+            access_token=session.access_token if hasattr(session, "access_token") else None,
+        )
+
+    def _response_to_user(self, response) -> Optional[User]:
+        """Convert Supabase auth response to User model."""
+        if not response or not response.user:
+            return None
+
+        user_meta = response.user.user_metadata or {}
+
+        return User(
+            id=response.user.id,
+            email=response.user.email,
+            github_handle=user_meta.get("user_name") or user_meta.get("preferred_username"),
+            avatar_url=user_meta.get("avatar_url"),
+            access_token=response.session.access_token if response.session else None,
+        )
+
+
+# Singleton instance
+_auth_instance: Optional[AuthProvider] = None
+
+
+def get_auth() -> AuthProvider:
+    """Get the auth provider instance."""
+    global _auth_instance
+    if _auth_instance is None:
+        _auth_instance = AuthProvider()
+    return _auth_instance

emdash_core/db/models.py

@@ -0,0 +1,91 @@
+"""Data models for the database layer."""
+
+from dataclasses import dataclass, field
+from datetime import datetime
+from enum import Enum
+from typing import Optional
+
+
+class FeatureStatus(str, Enum):
+    """Status of a feature in the workflow."""
+
+    TODO = "todo"
+    IN_DESIGN_REVIEW = "in_design_review"
+    IN_PROGRESS = "in_progress"
+    IN_PR = "in_pr"
+    DONE = "done"
+
+
+class PRStatus(str, Enum):
+    """Status of a pull request."""
+
+    OPEN = "open"
+    MERGED = "merged"
+    CLOSED = "closed"
+
+
+@dataclass
+class Project:
+    """A project in the system."""
+
+    id: str
+    name: str
+    repo_url: Optional[str] = None
+    owner_id: Optional[str] = None
+    created_at: Optional[datetime] = None
+    updated_at: Optional[datetime] = None
+
+
+@dataclass
+class TeamMember:
+    """A team member belonging to a project."""
+
+    id: str
+    project_id: str
+    name: str
+    email: Optional[str] = None
+    github_handle: Optional[str] = None
+    role: Optional[str] = None
+    user_id: Optional[str] = None  # Links to auth.users
+    created_at: Optional[datetime] = None
+
+
+@dataclass
+class Feature:
+    """A feature belonging to a project."""
+
+    id: str
+    project_id: str
+    title: str
+    description: Optional[str] = None
+    status: FeatureStatus = FeatureStatus.TODO
+    spec: Optional[str] = None
+    plan: Optional[str] = None
+    tasks: Optional[str] = None
+    created_at: Optional[datetime] = None
+    updated_at: Optional[datetime] = None
+    # Populated by joins
+    assignees: list["TeamMember"] = field(default_factory=list)
+    prs: list["FeaturePR"] = field(default_factory=list)
+
+
+@dataclass
+class FeatureAssignee:
+    """Junction table for feature-assignee many-to-many relationship."""
+
+    feature_id: str
+    team_member_id: str
+    assigned_at: Optional[datetime] = None
+
+
+@dataclass
+class FeaturePR:
+    """A pull request linked to a feature."""
+
+    id: str
+    feature_id: str
+    pr_url: str
+    pr_number: int
+    title: Optional[str] = None
+    status: PRStatus = PRStatus.OPEN
+    created_at: Optional[datetime] = None