refactor: excel parse
This commit is contained in:
Blizzard
@@ -0,0 +1,233 @@
|
||||
"""SQLAlchemy models matching the Go project's database schema."""
|
||||
|
||||
from datetime import datetime
|
||||
from sqlalchemy import Column, String, Integer, Boolean, Text, DateTime, create_engine
|
||||
from sqlalchemy.orm import DeclarativeBase, Session
|
||||
|
||||
|
||||
class Base(DeclarativeBase):
|
||||
pass
|
||||
|
||||
|
||||
# ── Global DB Models ──
|
||||
|
||||
|
||||
class LLMProvider(Base):
|
||||
__tablename__ = "llm_providers"
|
||||
|
||||
id = Column(String, primary_key=True)
|
||||
name = Column(String, default="")
|
||||
provider = Column(String, default="") # Ollama, DeepSeek, OpenAI, Qwen
|
||||
base_url = Column(String, default="")
|
||||
api_key = Column(String, default="")
|
||||
model_id = Column(String, default="")
|
||||
enabled = Column(Boolean, default=True)
|
||||
created_at = Column(DateTime, default=datetime.utcnow)
|
||||
updated_at = Column(DateTime, default=datetime.utcnow, onupdate=datetime.utcnow)
|
||||
|
||||
|
||||
class VectorDBConfig(Base):
|
||||
__tablename__ = "vector_db_configs"
|
||||
|
||||
id = Column(Integer, primary_key=True, default=1)
|
||||
endpoint = Column(String, default="")
|
||||
api_key = Column(String, default="")
|
||||
status = Column(String, default="disconnected")
|
||||
|
||||
|
||||
class EmbeddingModelConfig(Base):
|
||||
__tablename__ = "embedding_model_configs"
|
||||
|
||||
id = Column(Integer, primary_key=True, default=1)
|
||||
provider = Column(String, default="OpenAI")
|
||||
base_url = Column(String, default="")
|
||||
api_key = Column(String, default="")
|
||||
model_id = Column(String, default="")
|
||||
enabled = Column(Boolean, default=False)
|
||||
dimensions = Column(Integer, default=1024)
|
||||
|
||||
|
||||
class Project(Base):
|
||||
__tablename__ = "projects"
|
||||
|
||||
id = Column(String, primary_key=True)
|
||||
name = Column(String, default="")
|
||||
path = Column(String, default="")
|
||||
created_at = Column(DateTime, default=datetime.utcnow)
|
||||
updated_at = Column(DateTime, default=datetime.utcnow, onupdate=datetime.utcnow)
|
||||
|
||||
|
||||
# ── Project-scoped DB Models ──
|
||||
|
||||
|
||||
class SourceFile(Base):
|
||||
__tablename__ = "source_files"
|
||||
|
||||
id = Column(String, primary_key=True)
|
||||
project_id = Column(String, index=True)
|
||||
name = Column(String, default="")
|
||||
type = Column(String, default="")
|
||||
category = Column(String, default="")
|
||||
file_path = Column(String, default="")
|
||||
size = Column(String, default="")
|
||||
parsed_content = Column(Text, default="")
|
||||
vector_status = Column(String, default="pending")
|
||||
created_at = Column(DateTime, default=datetime.utcnow)
|
||||
updated_at = Column(DateTime, default=datetime.utcnow, onupdate=datetime.utcnow)
|
||||
|
||||
|
||||
class ChatMessage(Base):
|
||||
__tablename__ = "chat_messages"
|
||||
|
||||
id = Column(Integer, primary_key=True, autoincrement=True)
|
||||
project_id = Column(String, index=True)
|
||||
role = Column(String, default="")
|
||||
content = Column(Text, default="")
|
||||
sources = Column(Text, default="")
|
||||
citations = Column(Text, default="")
|
||||
created_at = Column(DateTime, default=datetime.utcnow)
|
||||
|
||||
|
||||
class TemplateChapter(Base):
|
||||
__tablename__ = "template_chapters"
|
||||
|
||||
id = Column(String, primary_key=True)
|
||||
project_id = Column(String, index=True)
|
||||
template_name = Column(String, default="")
|
||||
title = Column(String, default="")
|
||||
status = Column(String, default="idle")
|
||||
progress = Column(Integer, default=0)
|
||||
content = Column(Text, default="")
|
||||
sort_order = Column(Integer, default=0)
|
||||
created_at = Column(DateTime, default=datetime.utcnow)
|
||||
updated_at = Column(DateTime, default=datetime.utcnow, onupdate=datetime.utcnow)
|
||||
|
||||
|
||||
class TextChunk(Base):
|
||||
__tablename__ = "text_chunks"
|
||||
|
||||
id = Column(String, primary_key=True)
|
||||
project_id = Column(String, index=True)
|
||||
source_id = Column(String, index=True)
|
||||
content = Column(Text, default="")
|
||||
chunk_idx = Column(Integer, default=0)
|
||||
created_at = Column(DateTime, default=datetime.utcnow)
|
||||
|
||||
|
||||
class DeliveryStandard(Base):
|
||||
__tablename__ = "delivery_standards"
|
||||
|
||||
id = Column(Integer, primary_key=True, autoincrement=True)
|
||||
project_id = Column(String, index=True)
|
||||
file_name = Column(String, default="")
|
||||
content = Column(Text, default="")
|
||||
created_at = Column(DateTime, default=datetime.utcnow)
|
||||
updated_at = Column(DateTime, default=datetime.utcnow, onupdate=datetime.utcnow)
|
||||
|
||||
|
||||
# ── Pydantic schemas for API ──
|
||||
|
||||
from pydantic import BaseModel
|
||||
from typing import Optional
|
||||
|
||||
|
||||
class LLMProviderSchema(BaseModel):
|
||||
id: str = ""
|
||||
name: str = ""
|
||||
provider: str = ""
|
||||
url: str = ""
|
||||
key: str = ""
|
||||
model: str = ""
|
||||
enabled: bool = True
|
||||
|
||||
model_config = {"from_attributes": True}
|
||||
|
||||
@classmethod
|
||||
def from_orm_obj(cls, obj: LLMProvider) -> "LLMProviderSchema":
|
||||
return cls(id=obj.id, name=obj.name, provider=obj.provider,
|
||||
url=obj.base_url, key=obj.api_key, model=obj.model_id,
|
||||
enabled=obj.enabled)
|
||||
|
||||
|
||||
class VectorDBConfigSchema(BaseModel):
|
||||
endpoint: str = ""
|
||||
apiKey: str = ""
|
||||
status: str = "disconnected"
|
||||
|
||||
|
||||
class EmbeddingConfigSchema(BaseModel):
|
||||
provider: str = "OpenAI"
|
||||
url: str = ""
|
||||
key: str = ""
|
||||
model: str = ""
|
||||
enabled: bool = False
|
||||
dimensions: int = 1024
|
||||
|
||||
@classmethod
|
||||
def from_orm_obj(cls, obj: EmbeddingModelConfig) -> "EmbeddingConfigSchema":
|
||||
return cls(provider=obj.provider, url=obj.base_url, key=obj.api_key,
|
||||
model=obj.model_id, enabled=obj.enabled,
|
||||
dimensions=obj.dimensions or 1024)
|
||||
|
||||
|
||||
class ProjectSchema(BaseModel):
|
||||
id: str
|
||||
name: str
|
||||
|
||||
model_config = {"from_attributes": True}
|
||||
|
||||
|
||||
class SourceFileSchema(BaseModel):
|
||||
id: str
|
||||
name: str
|
||||
type: str
|
||||
category: str
|
||||
size: str
|
||||
vectorStatus: str = "pending"
|
||||
|
||||
model_config = {"from_attributes": True}
|
||||
|
||||
@classmethod
|
||||
def from_orm_obj(cls, obj: SourceFile) -> "SourceFileSchema":
|
||||
return cls(id=obj.id, name=obj.name, type=obj.type,
|
||||
category=obj.category, size=obj.size,
|
||||
vectorStatus=obj.vector_status)
|
||||
|
||||
|
||||
class ChatMessageSchema(BaseModel):
|
||||
id: int = 0
|
||||
projectId: str = ""
|
||||
role: str = ""
|
||||
content: str = ""
|
||||
sources: str = ""
|
||||
citations: str = ""
|
||||
|
||||
model_config = {"from_attributes": True}
|
||||
|
||||
@classmethod
|
||||
def from_orm_obj(cls, obj: ChatMessage) -> "ChatMessageSchema":
|
||||
return cls(id=obj.id, projectId=obj.project_id, role=obj.role,
|
||||
content=obj.content, sources=obj.sources or "",
|
||||
citations=obj.citations or "")
|
||||
|
||||
|
||||
class TemplateChapterSchema(BaseModel):
|
||||
id: str
|
||||
title: str
|
||||
status: str = "idle"
|
||||
progress: int = 0
|
||||
content: str = ""
|
||||
sortOrder: int = 0
|
||||
|
||||
model_config = {"from_attributes": True}
|
||||
|
||||
@classmethod
|
||||
def from_orm_obj(cls, obj: TemplateChapter) -> "TemplateChapterSchema":
|
||||
return cls(id=obj.id, title=obj.title, status=obj.status,
|
||||
progress=obj.progress, content=obj.content,
|
||||
sortOrder=obj.sort_order)
|
||||
|
||||
|
||||
class DeliveryStandardSchema(BaseModel):
|
||||
fileName: str = ""
|
||||
content: str = ""
|
||||
Reference in New Issue
Block a user