promptstudio-sdk 1.0.254

A Python SDK for PromptStudio

PromptStudio Python SDK

A Python SDK for interacting with PromptStudio API and AI platforms directly.

Installation

From PyPI

pip install promptstudio-sdk

From Source

git clone https://github.com/your-repo/promptstudio-sdk.git
cd promptstudio-sdk
pip install -e .

Development Setup

1. Create a virtual environment:

python -m venv venv

2. Activate the virtual environment:

# On Windows
venv\Scripts\activate

# On Unix or MacOS
source venv/bin/activate

3. Install dependencies:

pip install -r requirements.txt

---

Usage

Initializing the SDK

from promptstudio_sdk import PromptStudio

client = PromptStudio({
    'api_key': 'YOUR_API_KEY',
    'env': 'prod',  # Use 'prod' for production environment
    'bypass': True,
    'is_logging': True,
    'timeout': 5000  # Timeout in milliseconds
})

Configuration Options

bypass (default: False)

The bypass parameter determines whether to use the local AI provider directly or route requests through PromptStudio's API:

• When bypass=True: Requests go directly to the AI provider, bypassing PromptStudio's API
• When bypass=False: Requests are routed through PromptStudio's API for additional processing and logging

When bypass=True, you can use the role-based message format in user_message, such as:

user_message = [
    {"role": "user", "content": [{"type": "text", "text": "Hi"}]},
    {"role": "assistant", "content": [{"type": "text", "text": "Hello! How can I help you?"}]},
    {"role": "user", "content": [{"type": "text", "text": "What is Python?"}]}
]

Note: This format is only supported when bypass=True. If used with bypass=False, an error will be raised.

When bypass=False, you must use the simpler format:

user_message = [
    {"type": "text", "text": "Hello"}
]

is_session_enabled (default: True)

Controls whether conversation history is maintained across requests.

is_logging (default: True)

Determines whether interactions are logged for analytics.

shot (default: -1)

Controls how many message pairs to include from the beginning of the conversation:

• -1: All previous messages are included
• 0: No previous messages are included
• n > 0: First n pairs (2n messages) are included

timeout (default: 25)

Sets the maximum time (in milliseconds) to wait for a response. If exceeded, raises a TimeoutError.

---

Using shot

# Example: Using shot to include first 2 pairs of messages
response = client.chat_with_prompt(
    prompt_id="your_prompt_id",
    user_message=[{"type": "text", "text": "Hello"}],
    session_id="your_session_id",
    shot=2
)

Memory Control

memory_type (default: "fullMemory")

Controls how past conversation context is included in each request.

Supported values:

• "fullMemory": Maintains the complete conversation history.
• "windowMemory": Keeps a sliding window of recent messages,includes only the latest N messages, controlled via window_size.
• "summarizedMemory":Maintains a summarized version of the conversation history.

window_size (used with "windowMemory")

When memory_type is set to "windowMemory", window_size defines how many past messages to include.

Example:

response = client.chat_with_prompt({
    "prompt_id": "abc123",
    "user_message": [{"type": "text", "text": "Tell me a joke"}],
    "session_id": "session_xyz",
    "memory_type": "windowMemory",
    "window_size": 6
})

---

Chatting with a Prompt

response = client.chat_with_prompt(
    prompt_id="your_prompt_id",
    user_message=[
        {
            "type": "text",
            "text": "Hello, how are you?"
        }
    ],
    memory_type="fullMemory",
    window_size=0,
    session_id="your_session_id",
    variables={},
    is_session_enabled=True,
    shot=2,
)

print(response)

---

Tag Field (Optional Metadata)

You can pass a tag dictionary to include custom metadata such as user identifiers :

tag = {
    "userId": "680b5b825149777520281b5b",
    "userName":"Roshni",
    "Location":"India"
}

• Useful for custom tracking, filtering, or analytics

---

Complete Example

from promptstudio_sdk import PromptStudio

def main():
    client = PromptStudio({
        'api_key': 'YOUR_API_KEY',
        'env': 'test',
        'bypass': True,
        'is_logging': True,
        'timeout': 500 
    })

    try:

        response = client.chat_with_prompt(
            prompt_id="your_prompt_id",
            user_message=[
                {"type": "text", "text": "Hello, how are you?"}
            ],
            memory_type="windowMemory",
            window_size=10,
            session_id="your_session_id",
            variables={},
            is_session_enabled=True,
            shot=2,
        )
        print("Chat response:", response)

    except Exception as e:
        print(f"An error occurred: {e}")

if __name__ == "__main__":
    main()

---

Session Retrieval

response = await client.get_session(session_id="your_session_id")

This returns the complete session state, including historical messages and any metadata.

Prompt Identifier Retrieval

response = await client.get_prompt_identifier(prompt_id="68d430715a2dd51457poie96")

This returns the unique prompt identifier for the given prompt_id.

Prompt Data Retrieval

response = response = await client.get_prompt_data(prompt_id="687a268dc5719ceeoi783216")

This returns the complete prompt details (metadata, content, etc.) for the given prompt_id.

---

Type Hints

from typing import Dict, List, Union, Optional

ImageMessage = Dict[str, Union[str, Dict[str, str]]]
TextMessage = Dict[str, str]
UserMessage = List[Union[ImageMessage, TextMessage]]

Memory = Literal["fullMemory", "windowMemory", "summarizedMemory"]

RequestPayload = Dict[str, Union[UserMessage, Memory, int, str, Dict[str, str], Optional[int]]]

---

Contributing

1. Fork the repository
2. Create a new branch
3. Make your changes
4. Run tests
5. Submit a pull request

License

This SDK is released under the MIT License.