"""Saiten MCP — Data Models (Pydantic)."""
from __future__ import annotations
from datetime import datetime, timezone
from typing import Any
from pydantic import BaseModel, Field
# ---------------------------------------------------------------------------
# Rubric models
# ---------------------------------------------------------------------------
class ScoringCriteria(BaseModel):
"""A single scoring criterion."""
name: str
weight: float = Field(ge=0.0, le=1.0)
description: str
scoring_guide: dict[str, str]
class Rubric(BaseModel):
"""Track-specific scoring rubric."""
track: str
track_display_name: str
criteria: list[ScoringCriteria]
notes: str = ""
# ---------------------------------------------------------------------------
# Submission models
# ---------------------------------------------------------------------------
class Submission(BaseModel):
"""Submission summary for Issue listing."""
issue_number: int
title: str
track: str
project_name: str
repo_url: str | None = None
created_at: str
has_demo: bool = False
class SubmissionDetail(BaseModel):
"""Detailed submission info (scoring target)."""
issue_number: int
title: str
track: str
project_name: str
description: str
repo_url: str | None = None
readme_content: str | None = None
technologies: list[str] = Field(default_factory=list)
technical_highlights: str = ""
has_demo: bool = False
demo_description: str = ""
submission_checklist: dict[str, Any] = Field(default_factory=dict)
team_members: str | None = None
setup_summary: str = ""
# ---------------------------------------------------------------------------
# Score models
# ---------------------------------------------------------------------------
class CriteriaScore(BaseModel):
"""Score for an individual criterion."""
name: str
score: int = Field(ge=1, le=10)
weight: float = Field(ge=0.0, le=1.0)
class SubmissionScore(BaseModel):
"""Scoring result for a single submission.
Enhanced with evidence-anchored scoring fields:
- evidence: per-criterion evidence citations
- confidence: scorer's confidence level
- red_flags_detected: red flag signals found
- bonus_signals_detected: bonus signals found
"""
issue_number: int
project_name: str
track: str
criteria_scores: dict[str, int]
weighted_total: float = Field(ge=0.0, le=100.0)
evidence: dict[str, str] = Field(
default_factory=dict,
description="Per-criterion evidence citations from submission content",
)
confidence: str = Field(
default="medium",
description="Scorer confidence: 'high', 'medium', or 'low'",
)
red_flags_detected: list[str] = Field(
default_factory=list,
description="Red flag signals detected during scoring",
)
bonus_signals_detected: list[str] = Field(
default_factory=list,
description="Bonus signals detected during scoring",
)
strengths: list[str]
improvements: list[str]
summary: str
scored_at: datetime = Field(default_factory=lambda: datetime.now(timezone.utc))
class ScoreMetadata(BaseModel):
"""Metadata for scores.json."""
last_updated: str
version: str = "1.0"
total_submissions: int = 0
scored_count: int = 0
class ScoreStore(BaseModel):
"""Top-level scores.json structure."""
metadata: ScoreMetadata
scores: list[SubmissionScore] = Field(default_factory=list)
