refone/tiny-rag

How refone/tiny-rag works

Shop Assistant Agent ingests PDF and Markdown documents into a searchable knowledge base, then answers user queries using vectorization, three-way retrieval, and generation. The documented import graph routes PDFs through MinerU, processes Markdown images, splits and recognizes content, then embeds and persists it. Shared configuration, model, storage, and tracing modules support these stages. The query, API, and web implementations are not present in the sampled source; query stages are shown only at the level documented in the README.

Main entry points: main_graph.py Import workflow · node_md_img.py Image processing · node_entry.py File detection · node_pdf_to_md.py PDF conversion

Components 24

Document ingestion

Import workflow src/processor/import_processor/main_graph.py
Import state src/processor/import_processor/state.py
File detection src/processor/import_processor/nodes/node_entry.py
PDF conversion src/processor/import_processor/nodes/node_pdf_to_md.py
Image processing src/processor/import_processor/nodes/node_md_img.py
Document splitting src/processor/import_processor/nodes/node_document_split.py
Item recognition src/processor/import_processor/nodes/node_item_name_recognize.py
Embedding src/processor/import_processor/nodes/node_bge_embedding.py
Vector persistence src/processor/import_processor/nodes/node_upsert_milvus.py

Query answering

Query vectorization src/processor/query_processor
Three-way retrieval src/processor/query_processor
Answer generation src/processor/query_processor

Shared services

Vision-language model src/common/model/vlm.py
Image object storage src/common/client/minio_client.py
Environment config src/common/config/env_config.py
Application config src/common/config/app_config.py
Document types src/common/enum/doc_type.py
Prompt loading src/common/prompt/load_prompt.py

Runtime infrastructure

Node tracing src/utils/node_utils.py
Task tracking src/utils/task_utils.py
Logging src/utils/logging_utils.py

Other components

Assistant user
MinerU service
Milvus

Connections 27

Import workflow → File detection: dispatches first src/processor/import_processor/main_graph.py
Import workflow → PDF conversion: routes PDFs src/processor/import_processor/main_graph.py
Import workflow → Image processing: routes Markdown src/processor/import_processor/main_graph.py
Import workflow → Document splitting: dispatches after images src/processor/import_processor/main_graph.py
Import workflow → Item recognition: dispatches after split src/processor/import_processor/main_graph.py
Import workflow → Embedding: dispatches after recognition src/processor/import_processor/main_graph.py
Import workflow → Vector persistence: dispatches after embedding src/processor/import_processor/main_graph.py
File detection → Import state: updates src/processor/import_processor/nodes/node_entry.py
PDF conversion → MinerU service: uploads and polls src/processor/import_processor/nodes/node_pdf_to_md.py
PDF conversion → Import state: writes Markdown path src/processor/import_processor/nodes/node_pdf_to_md.py
Image processing → Vision-language model: requests image summaries src/processor/import_processor/nodes/node_md_img.py
Image processing → Prompt loading: loads image prompt src/processor/import_processor/nodes/node_md_img.py
Image processing → Image object storage: uploads images src/processor/import_processor/nodes/node_md_img.py
Image processing → Application config: checks image formats src/processor/import_processor/nodes/node_md_img.py
Image processing → Import state: reads and updates src/processor/import_processor/nodes/node_md_img.py
Vector persistence → Milvus: persists vectors README.md
Assistant user → Query vectorization: submits query README.md
Query vectorization → Three-way retrieval: provides query vectors README.md
Three-way retrieval → Milvus: retrieves matches README.md
Three-way retrieval → Answer generation: provides retrieved context README.md
Answer generation → Assistant user: returns answer README.md
Vision-language model → Environment config: reads model settings src/common/model/vlm.py
Image object storage → Environment config: reads storage settings src/common/client/minio_client.py
File detection → Document types: detects file type src/processor/import_processor/nodes/node_entry.py
Node tracing → Task tracking: records node status src/utils/node_utils.py
Node tracing → Logging: logs node activity src/utils/node_utils.py
Logging → Environment config: reads logging settings src/utils/logging_utils.py

Drawn by GitDiagram on September 30, 2026 from the repository's file tree, README and source files. Read as Markdown ·