95 lines
2.6 KiB
Python
95 lines
2.6 KiB
Python
"""Document models for indexed documentation."""
|
|
|
|
from dataclasses import dataclass, field
|
|
from datetime import datetime
|
|
from enum import Enum
|
|
from typing import Optional
|
|
|
|
|
|
class SourceType(str, Enum):
|
|
"""Enumeration of supported documentation source types."""
|
|
|
|
OPENAPI = "openapi"
|
|
README = "readme"
|
|
CODE = "code"
|
|
|
|
|
|
@dataclass
|
|
class Document:
|
|
"""Represents an indexed document chunk."""
|
|
|
|
id: str
|
|
content: str
|
|
source_type: SourceType
|
|
title: str
|
|
file_path: str = ""
|
|
metadata: dict = field(default_factory=dict)
|
|
created_at: datetime = field(default_factory=datetime.utcnow)
|
|
|
|
def to_dict(self) -> dict:
|
|
"""Convert document to dictionary for serialization."""
|
|
return {
|
|
"id": self.id,
|
|
"content": self.content,
|
|
"source_type": self.source_type.value,
|
|
"title": self.title,
|
|
"file_path": self.file_path,
|
|
"metadata": self.metadata,
|
|
"created_at": self.created_at.isoformat(),
|
|
}
|
|
|
|
@classmethod
|
|
def from_dict(cls, data: dict) -> "Document":
|
|
"""Create document from dictionary."""
|
|
return cls(
|
|
id=data["id"],
|
|
content=data["content"],
|
|
source_type=SourceType(data["source_type"]),
|
|
title=data["title"],
|
|
file_path=data.get("file_path", ""),
|
|
metadata=data.get("metadata", {}),
|
|
created_at=datetime.fromisoformat(data["created_at"]),
|
|
)
|
|
|
|
|
|
@dataclass
|
|
class SearchResult:
|
|
"""Represents a search result with relevance score."""
|
|
|
|
document: Document
|
|
score: float
|
|
highlights: list[str] = field(default_factory=list)
|
|
|
|
def to_dict(self) -> dict:
|
|
"""Convert search result to dictionary."""
|
|
return {
|
|
"id": self.document.id,
|
|
"content": self.document.content,
|
|
"source_type": self.document.source_type.value,
|
|
"title": self.document.title,
|
|
"file_path": self.document.file_path,
|
|
"score": self.score,
|
|
"highlights": self.highlights,
|
|
}
|
|
|
|
|
|
@dataclass
|
|
class IndexStats:
|
|
"""Statistics about the indexed collection."""
|
|
|
|
total_documents: int = 0
|
|
openapi_count: int = 0
|
|
readme_count: int = 0
|
|
code_count: int = 0
|
|
last_indexed: Optional[datetime] = None
|
|
|
|
def to_dict(self) -> dict:
|
|
"""Convert stats to dictionary."""
|
|
return {
|
|
"total_documents": self.total_documents,
|
|
"openapi_count": self.openapi_count,
|
|
"readme_count": self.readme_count,
|
|
"code_count": self.code_count,
|
|
"last_indexed": self.last_indexed.isoformat() if self.last_indexed else None,
|
|
}
|