plp-client

Official Python client for PLP (Prompt Library Protocol).

Installation

pip install plp-client

Quick Start

from plp_client import PLPClient, PromptInput

# Initialize client
client = PLPClient("https://prompts.goreal.ai", api_key="your-api-key")

# Get latest version of a prompt
prompt = client.get("marketing/welcome-email")
print(prompt.content)  # "Hello {{name}}..."

# Get specific version
old_prompt = client.get("marketing/welcome-email", "1.0.0")

# Save a new prompt
client.put("support/faq-bot", PromptInput(
    content="You are a helpful FAQ bot. Answer: {{question}}",
    meta={"version": "1.0.0", "author": "yoad"}
))

# Delete a prompt
client.delete("deprecated/old-prompt")

API Reference

PLPClient

Constructor

PLPClient(base_url: str, api_key: Optional[str] = None, 
          headers: Optional[Dict[str, str]] = None, timeout: int = 10)

Parameters:

• base_url (str): Base URL of the PLP server
• api_key (str, optional): Optional Bearer token for authentication
• headers (dict, optional): Additional HTTP headers
• timeout (int): Request timeout in seconds (default: 10)

Methods

get(prompt_id, version=None)

Retrieve a prompt by ID and optional version.

def get(prompt_id: str, version: Optional[str] = None) -> PromptEnvelope

• prompt_id (str): Unique prompt identifier (e.g., "marketing/welcome-email")
• version (str, optional): Optional version string (e.g., "1.2.0"). If omitted, returns latest.

Returns: PromptEnvelope

Raises: PLPError if not found (404) or other errors

put(prompt_id, input)

Create or update a prompt (idempotent upsert).

def put(prompt_id: str, input: PromptInput) -> PromptEnvelope

• prompt_id (str): Unique prompt identifier
• input (PromptInput): Object with content and optional meta

Returns: The saved PromptEnvelope

delete(prompt_id)

Delete a prompt and all its versions.

def delete(prompt_id: str) -> None

• prompt_id (str): Unique prompt identifier

Returns: None (204 No Content on success)

Aliases

fetch(prompt_id, version=None)  # Alias for get()
save(prompt_id, input)          # Alias for put()

Data Classes

PromptEnvelope

class PromptEnvelope:
    id: str
    content: str
    meta: Dict[str, Any]

PromptInput

class PromptInput:
    content: str
    meta: Dict[str, Any]

Error Handling

from plp_client import PLPClient, PLPError

client = PLPClient("https://prompts.goreal.ai")

try:
    prompt = client.get("missing/prompt")
except PLPError as e:
    print(f"Error: {e}")
    print(f"Status code: {e.status_code}")

Examples

Using with OpenAI

from plp_client import PLPClient
from openai import OpenAI

plp = PLPClient("https://prompts.goreal.ai")
openai = OpenAI()

prompt = plp.get("assistant/code-reviewer")

response = openai.chat.completions.create(
    model="gpt-4",
    messages=[{"role": "user", "content": prompt.content}],
    temperature=prompt.meta.get("model_config", {}).get("temperature", 0.7)
)

print(response.choices[0].message.content)

Using with Anthropic

from plp_client import PLPClient
import anthropic

plp = PLPClient("https://prompts.goreal.ai")
client = anthropic.Anthropic()

prompt = plp.get("creative/story-generator")

message = client.messages.create(
    model="claude-3-opus-20240229",
    max_tokens=prompt.meta.get("model_config", {}).get("max_tokens", 1024),
    messages=[{"role": "user", "content": prompt.content}]
)

print(message.content)

Context Manager

from plp_client import PLPClient

with PLPClient("https://prompts.goreal.ai") as client:
    prompt = client.get("my/prompt")
    print(prompt.content)
# Session is automatically closed

Development

# Install dependencies
pip install -e ".[dev]"

# Run tests
pytest

# Run tests with coverage
pytest --cov=plp_client

# Format code
black src/ tests/

# Type checking
mypy src/

License

MIT © GoReal.AI

Contributing

See CONTRIBUTING.md

Learn More

• PLP Specification
• OpenAPI Documentation
• GitHub Repository