template/README.md

Prompt Template

Desription

Prompt Template.

Code Example

Example

from towhee import ops, pipe
import requests

towhee_docs = requests.get('https://raw.githubusercontent.com/towhee-io/towhee/main/README.md').content

temp = """{question}

input:
{context}
"""

sys_message = """Your name is TowheeChat."""

p = (
    pipe.input('question', 'doc', 'history')
    .map('doc', 'doc', lambda x: x[:2000])
    .map(('question', 'doc', 'history'), 'prompt', ops.prompt.template(temp, ['question', 'context'], sys_message))
    .map('prompt', 'answer', ops.LLM.OpenAI())
    .output('answer')
)

an1 = p('What is your name?', [], []).get()[0]
print(an1)

an2 = p('Tell me something about Towhee', towhee_docs, []).get()[0]
print(an2)

an3 = p('How to use it', towhee_docs, [('Tell me something about Towhee', an2)]).get()[0]
print(an3)

Factory Constructor

Create the operator via the following factory method:

ops.prompt.template(temp, keys, sys_msg)

Parameters:

temp: str

​ A template to create a prompt as the last user message.

keys: list

​ A list of keys used in template.

sys_msg: str=None

​ A system message, defaults to None. If None, it will not pass any system message.

Returns: List[Dict]

More Resources

• ChatGPT+ Vector database + prompt-as-code - The CVP Stack - Zilliz blog: Extend the capability of ChatGPT with a Vector database and prompts-as-code
• What is Prompt as Code (Prompt Engineering): Explores what prompt engineering is, how it works in NLP, and best practices for effective prompt engineering.
• An LLM Powered Text to Image Prompt Generation with Milvus - Zilliz blog: An interesting LLM project powered by the Milvus vector database for generating more efficient text-to-image prompts.
• Prompting in LangChain - Zilliz blog: Prompting is one of today's most popular and important tasks in AI app building. Learn how to use LangChain for more complex prompts.
• What is Prompt Chaining: Prompt chaining in NLP uses structured prompts to break a complex task into smaller subtasks. This sequential approach improves coherence and accuracy in LLM outputs.