Paweł Domański
170 lines
3.6 KiB
Markdown
170 lines
3.6 KiB
Markdown
---
|
||
type: Note
|
||
---
|
||
# BUdowa folderów
|
||
|
||
Poniżej masz **wierne przepisanie (OCR) struktury z obrazka** w formie tekstowej:
|
||
|
||
***
|
||
|
||
# **GENERATIVE AI PROJECT STRUCTURE**
|
||
|
||
```text
|
||
generative_ai_project/
|
||
├── config/
|
||
│ ├── model_config.yaml
|
||
│ └── logging_config.yaml
|
||
│
|
||
├── data/
|
||
│ ├── cache/
|
||
│ ├── embeddings/
|
||
│ └── vectordb/
|
||
│
|
||
├── src/
|
||
│ ├── core/
|
||
│ │ ├── base_llm.py
|
||
│ │ ├── gpt_client.py
|
||
│ │ ├── claude_client.py
|
||
│ │ ├── local_llm.py
|
||
│ │ └── model_factory.py
|
||
│ │
|
||
│ ├── prompts/
|
||
│ │ ├── templates.py
|
||
│ │ └── chain.py
|
||
│ │
|
||
│ ├── rag/
|
||
│ │ ├── embedder.py
|
||
│ │ ├── retriever.py
|
||
│ │ ├── vector_store.py
|
||
│ │ └── indexer.py
|
||
│ │
|
||
│ ├── processing/
|
||
│ │ ├── chunking.py
|
||
│ │ ├── tokenizer.py
|
||
│ │ └── preprocessor.py
|
||
│ │
|
||
│ └── inference/
|
||
│ ├── inference_engine.py
|
||
│ └── response_parser.py
|
||
│
|
||
├── tests/
|
||
│ ├── unit/
|
||
│ │ ├── test_llm_clients.py
|
||
│ │ └── test_prompts.py
|
||
│ │
|
||
│ └── integration/
|
||
│ ├── test_end_to_end.py
|
||
│ └── test_api_integration.py
|
||
│
|
||
├── scripts/
|
||
│ ├── setup_env.sh
|
||
│ ├── run_tests.sh
|
||
│ ├── build_embeddings.py
|
||
│ └── cleanup.py
|
||
│
|
||
├── .gitignore
|
||
├── Dockerfile
|
||
├── docker-compose.yml
|
||
└── requirements.txt
|
||
```
|
||
|
||
***
|
||
|
||
# **Opis sekcji (z prawej strony obrazka)**
|
||
|
||
### config/ – Project configuration files:
|
||
|
||
- model_config.yaml: Model settings and API keys
|
||
- logging_config.yaml: Logging configuration
|
||
|
||
***
|
||
|
||
### data/ – Stores all project data:
|
||
|
||
- cache/: Cached and processed data
|
||
- embeddings/: Vector embeddings
|
||
- vectordb/: Vector database files
|
||
|
||
***
|
||
|
||
### src/ – Main application code
|
||
|
||
#### core/ – LLM abstraction and integrations:
|
||
|
||
- base_llm.py: Common interface for all LLMs
|
||
- gpt_client.py: OpenAI GPT client
|
||
- claude_client.py: Anthropic Claude client
|
||
- local_llm.py: Local/self-hosted models
|
||
- model_factory.py: Model selection factory
|
||
|
||
***
|
||
|
||
#### prompts/ – Prompt design utilities:
|
||
|
||
- templates.py: Prompt templates
|
||
- chain.py: Multi-step prompt chaining
|
||
|
||
***
|
||
|
||
#### rag/ – Retrieval-Augmented Generation module:
|
||
|
||
- embedder.py: Embedding creation
|
||
- retriever.py: Document search
|
||
- vector_store.py: Vector DB wrapper
|
||
- indexer.py: Document indexing
|
||
|
||
***
|
||
|
||
#### processing/ – Data preparation pipeline:
|
||
|
||
- chunking.py: Text splitting
|
||
- tokenizer.py: Tokenization utilities
|
||
- preprocessor.py: Data cleaning
|
||
|
||
***
|
||
|
||
#### inference/ – Model execution flow:
|
||
|
||
- inference_engine.py: Inference engine
|
||
- response_parser.py: Output formatting
|
||
|
||
***
|
||
|
||
### tests/ – Testing modules
|
||
|
||
#### unit/ – Function-level tests:
|
||
|
||
- test_llm_clients.py
|
||
- test_prompts.py
|
||
|
||
#### integration/ – End-to-end tests:
|
||
|
||
- test_end_to_end.py
|
||
- test_api_integration.py
|
||
|
||
***
|
||
|
||
### scripts/ – Automation scripts for common tasks:
|
||
|
||
- setup_env.sh: Environment setup
|
||
- run_tests.sh: Test execution
|
||
- build_embeddings.py: Embedding generation
|
||
- cleanup.py: Cleanup tasks
|
||
|
||
***
|
||
|
||
### Configuration Files:
|
||
|
||
- .gitignore: Git exclusions
|
||
- Dockerfile: Container configuration
|
||
- docker-compose.yml: Multi-service setup
|
||
- requirements.txt: Dependencies
|
||
|
||
***
|
||
|
||
Jeśli chcesz — mogę:
|
||
|
||
- przerobić to na **org-mode (pod Twój workflow)**
|
||
- zrobić **template repo (ZIP / Git skeleton)**
|
||
- albo rozbić to na **architekturę pod Twoje use-case’y (DBA + AI + automatyzacja)**
|