AR 15 M4
a7f78d71b2
FastAPI + ChromaDB + E5-large + DeepSeek по паттерну work-pcs-dr-cdss, адаптированному под пациентский контекст: - services: embeddings (E5-large с префиксами), vectorstore (коллекция operators_wiki), document_processor (PDF/DOCX/TXT/MD + чанкер с FAQ- паттерном под wiki), llm_client (системный промпт ассистента клиники), rag_pipeline (одиночный вопрос → retrieval → ответ). - routers: /health, /documents (upload, list, chunks, delete), /query. - static/index.html: шапка со статусом, блок базы знаний с раскрытием чанков по клику, блок тест-вопроса с 3-колоночным ответом (чанки со score / собранный промпт / ответ LLM). - Порт 8003 (8001 занят CDSS, 8002 — voicenote). E2E проверен: загрузка wiki_test.md → 2 чанка, вопрос «как записать ребёнка к лору?» → top score 84.8%, корректный ответ DeepSeek. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
78 lines
1.5 KiB
Python
78 lines
1.5 KiB
Python
from pydantic import BaseModel, Field
|
|
|
|
|
|
class DocumentInfo(BaseModel):
|
|
document_id: str
|
|
name: str
|
|
chunks_count: int
|
|
file_type: str
|
|
created_at: str
|
|
metadata: dict = Field(default_factory=dict)
|
|
|
|
|
|
class ChunkPreview(BaseModel):
|
|
index: int
|
|
section: str = ""
|
|
page_number: int = 0
|
|
text_preview: str = ""
|
|
char_length: int = 0
|
|
|
|
|
|
class DocumentUploadResponse(BaseModel):
|
|
document_id: str
|
|
name: str
|
|
chunks_count: int
|
|
status: str = "indexed"
|
|
created_at: str
|
|
chunks_preview: list[ChunkPreview] = Field(default_factory=list)
|
|
|
|
|
|
class DocumentListResponse(BaseModel):
|
|
documents: list[DocumentInfo]
|
|
total: int
|
|
|
|
|
|
class ChunkDetail(BaseModel):
|
|
index: int
|
|
section: str = ""
|
|
page_number: int = 0
|
|
text: str = ""
|
|
char_length: int = 0
|
|
|
|
|
|
class DocumentChunksResponse(BaseModel):
|
|
document_id: str
|
|
name: str
|
|
file_type: str
|
|
chunks_count: int
|
|
chunks: list[ChunkDetail] = Field(default_factory=list)
|
|
|
|
|
|
class DocumentDeleteResponse(BaseModel):
|
|
ok: bool = True
|
|
deleted_chunks: int
|
|
|
|
|
|
class SourceInfo(BaseModel):
|
|
document_id: str
|
|
document_name: str
|
|
chunk_text: str
|
|
section: str = ""
|
|
page: int = 0
|
|
relevance_score: float = 0.0
|
|
|
|
|
|
class QueryResponse(BaseModel):
|
|
answer: str
|
|
sources: list[SourceInfo]
|
|
model_used: str
|
|
assembled_prompt: str = ""
|
|
|
|
|
|
class HealthResponse(BaseModel):
|
|
status: str = "ok"
|
|
chromadb: str
|
|
embedding_model: str
|
|
documents_count: int
|
|
chunks_count: int
|