agents

A model for generating a brag document from a list of documents.

Functions:

Name	Description
generate_brag_document	Generate a brag document from a list of text chunks.

generate_brag_document async

generate_brag_document(model_name: KnownModelName, chunks: Iterable[str], language: str = 'english', input_brag_document: str | None = None) -> str

Generate a brag document from a list of text chunks.

This function takes a list of text chunks, each representing a contribution or achievement, and uses an AI agent to generate a comprehensive brag document. The brag document highlights the user's accomplishments, technical skills, and contributions.

Each chunk is a string of text that represents a part of the document. Examples include:

• GitHub Pull Requests
• Blog posts
• Research papers
• Meeting notes
• Project reports

The caller is responsible for providing the chunks in the correct order and formatted in a sensible manner.

Parameters:

Name	Type	Description	Default
model_name	KnownModelName	The name of the AI model to use for generating the brag document.	required
chunks	Iterable[str]	An iterable of strings, where each string is a chunk of text representing a contribution or achievement. This is expected to contain at least one chunk.	required
language	str	The language in which to generate the brag document.	'english'
input_brag_document	str | None	An optional existing brag document to update with new contributions. If provided, the function will update this document. Otherwise, a new document will be generated from scratch.	None

Returns:

Type	Description
str	A string containing the generated brag document.

Source code in src/brag/agents.py

async def generate_brag_document(
    model_name: KnownModelName,
    chunks: Iterable[str],
    language: str = "english",
    input_brag_document: str | None = None,
) -> str:
    """Generate a brag document from a list of text chunks.

    This function takes a list of text chunks, each representing a contribution or achievement,
    and uses an AI agent to generate a comprehensive brag document. The brag document
    highlights the user's accomplishments, technical skills, and contributions.

    Each chunk is a string of text that represents a part of the document. Examples include:

    - GitHub Pull Requests
    - Blog posts
    - Research papers
    - Meeting notes
    - Project reports

    The caller is responsible for providing the chunks in the correct order and formatted in
    a sensible manner.

    Args:
        model_name: The name of the AI model to use for generating the brag document.
        chunks: An iterable of strings, where each string is a chunk of text
            representing a contribution or achievement. This is expected to contain at
            least one chunk.
        language: The language in which to generate the brag document.
        input_brag_document: An optional existing brag document to update with new
            contributions. If provided, the function will update this document.
            Otherwise, a new document will be generated from scratch.

    Returns:
        A string containing the generated brag document.
    """
    if input_brag_document:
        # If an existing brag document is provided, use it as the starting point
        brag_document = input_brag_document
        # When using an existing document, we need to process all input chunks
        # including the first one
        chunks_to_process = chunks
    else:
        # If no existing document is provided, generate a new one from the first chunk
        first_chunk, remaining_chunks = _head_and_tail(chunks)
        # Generate initial summary from the first chunk.
        initial_brag_document_generator_agent = _build_agent_from_system_prompt(
            model_name,
            system_prompt=f"""
                You are an expert in creating compelling brag documents that highlight a person's achievements and skills.
                Your task is to analyze a document and extract key accomplishments, technical skills demonstrated, and contributions made.
                Focus on quantifiable results and impactful contributions.
                Present the information in a concise and engaging manner, suitable for showcasing the individual's value.
                Return only the generated brag document without extra comments or code fences.
                Generate the brag document in {language}.
            """,
        )
        brag_document = await _generate_initial_brag_document(
            initial_brag_document_generator_agent,
            first_chunk,
        )
        # We've already processed the first chunk, so only process the remaining ones
        chunks_to_process = remaining_chunks

    # Iteratively refine the brag document with the chunks
    for chunk in chunks_to_process:
        brag_document_updater_agent = _build_agent_from_system_prompt(
            model_name,
            system_prompt=f"""
                You are an expert in refining existing brag documents by incorporating new information.
                Your task is to integrate new context into an existing brag document, ensuring that the document remains concise, engaging, and highlights the individual's key achievements and skills.
                Focus on seamlessly weaving in new accomplishments, technical skills, and contributions, while maintaining a consistent tone and style.
                Feel free to modify and re-arrange existing content to better reflect the new information.
                Return only the generated brag document without extra comments or code fences.
                Generate the brag document in {language}.
            """,
        )
        brag_document = await _update_brag_document(
            brag_document_updater_agent, brag_document, chunk
        )

    return brag_document