agents/contracts/RepoContext.py

from dataclasses import dataclass, field
from enum import Enum
from pathlib import Path
from typing import Optional, Literal, Any, Generic

import pandas as pd
from pydantic import field_validator, BaseModel, Field, ConfigDict, field_serializer

from contracts.CodeLanguages import Language
from contracts.GraphAnalyzer import AnalyzeResult, DependencyGraph


@dataclass
class RepoContext:
    """Result object carrying all info a coding agent needs."""
    remote_url: str
    branch: str
    base_branch: str
    head_commit: str
    is_new_clone: bool
    is_new_branch: bool
    repo_dir: Path = Field(..., description="Absolute path to the cloned repository")
    original_branch: Optional[str] = None
    uncommitted_changes: Optional[str] = None  # git diff (tracked)
    untracked_files: list[str] = field(default_factory=list)
    stash_ref: Optional[str] = None
    env: dict = field(default_factory=dict)


    @field_validator("repo_path")
    @classmethod
    def repo_must_exist(cls, v: Path) -> Path:
        if not v.is_dir():
            raise ValueError(f"Repository path does not exist or is not a directory: {v}")
        return v.resolve()




class ToolRequirement(BaseModel):
    """Describes a required external tool and how to install it."""
    tool_name: str
    command: str = Field(..., description="CLI command that must be on PATH")
    install_guide: str = Field(..., description="Markdown installation instructions")
    docs_url: Optional[str] = None


class AnalysisError(BaseModel):
    """Structured error returned when analysis cannot proceed."""
    error_code: str
    message: str
    missing_tools: list[ToolRequirement] = Field(default_factory=list)
    language: Optional[Language] = None
    entrypoint: Optional[str] = None


class AnalysisInput(BaseModel):
    """Input schema for the analyzer."""
    repo: RepoContext


class AnalysisResult(BaseModel):
    """Top-level result — either a graph or an error."""
    model_config = ConfigDict(arbitrary_types_allowed=True)

    success: bool
    graph: Optional[DependencyGraph] = None
    error: Optional[AnalysisError] = None
    leaf_clusters: Any = None
    res_all: AnalyzeResult | None = None

    @field_serializer("leaf_clusters")
    def _ser_leaf_clusters(self, v: Any, _info):
        if isinstance(v, pd.DataFrame):
            return v.to_dict(orient="records")
        return v  # already JSON-safe