File size: 2,626 Bytes
3370983 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 |
# Prompt Management System
This module provides a centralized way to manage prompts using **PromptLayer** with a local filesystem fallback. It allows you to version prompts, manage environments (dev/staging/prod), and easily switch between local development and cloud-managed prompts.
## 🚀 Usage
Import the `get_prompt` function to load prompts anywhere in your application.
```python
from src.backend.prompts import get_prompt
```
### 1. Load from PromptLayer (Default)
By default, if `PROMPTLAYER_API_KEY` is set in your environment, it will fetch the prompt from PromptLayer using the configured environment label (default: `production`).
```python
# Fetches 'DB_Executor' tagged with current environment (e.g., 'production')
system_prompt = get_prompt("DB_Executor")
```
### 2. Load Latest Version (Ignore Environment)
Useful for testing or when you want to ensure you have the absolute latest saved version from PromptLayer, ignoring any 'prod' or 'dev' tags.
```python
# Fetches the absolute latest version of the template
system_prompt = get_prompt("DB_Executor", latest_version=True)
```
### 3. Force Load from Local File
You can force loading from a local file, which is useful for local development without an internet connection or for testing new prompts before pushing to PromptLayer.
```python
# Loads from src/prompts/templates/db_executor/v1.txt
# (Assuming 'v1.txt' is the file name in that directory, or provide full path)
system_prompt = get_prompt("db_executor/v1", local_prompt_path="src/prompts/templates")
```
If you don't provide a `local_prompt_path` but also don't have a `PROMPTLAYER_API_KEY` set, it defaults to looking in `src/prompts/templates`.
## 📂 Directory Structure
Store your local prompt backups in `src/prompts/templates/`.
```
src/prompts/
├── __init__.py # Exposes get_prompt
├── prompt_layer.py # Core logic
├── templates/ # Local prompt storage
│ ├── db_executor/
│ │ └── v1.txt
│ ├── supervisor/
│ │ └── v1.txt
│ └── ...
└── info.md # This file
```
## ⚙️ Configuration
- **`PROMPTLAYER_API_KEY`**: Set this env var to enable PromptLayer.
- **`PROMPT_ENVIRONMENT`**: Set to `dev`, `staging`, or `production` (default) to control which tagged version is loaded.
## 🔍 Debugging
The system prints clear logs to stdout so you know where your prompt came from:
- `📋 Loaded prompt '...' from PromptLayer (env=production)`
- `📋 Loaded prompt '...' from PromptLayer (latest version)`
- `📄 Loaded prompt '...' from local file: ...`
|