iragent.models.InternetAgent Class Reference

Public Member Functions
None	__init__ (self, int chunk_size, str model, str base_url, str api_key, float temperature=0.1, int max_token=512, str provider="openai")
str	start (self, str query, num_result)
list[dict]	fast_start (self, str query, int num_result, int|None max_workers=None)

Public Attributes
	chunk_size = chunk_size
	summerize_agent

Protected Member Functions
	_summarize_page (self, result, str query)

Detailed Description

InternetAgent is a tool for conducting web-based searches, retrieving relevant web pages,
chunking their content, and summarizing them using a specified language model.

Attributes:
    chunk_size (int): Maximum token size for each chunk of webpage content.
    summerize_agent (Agent): An instance of the summarization agent for generating summaries
                             from text chunks based on a system prompt.

Args:
    chunk_size (int): Token limit for text chunking.
    model (str): The name or identifier of the language model to be used.
    base_url (str): Base URL for the API that powers the summarization model.
    api_key (str): API key for authenticating with the model provider.
    temperature (float, optional): Sampling temperature for generation. Defaults to 0.1.
    max_token (int, optional): Maximum number of tokens allowed in the summary output. Defaults to 512.
    provider (str, optional): Name of the model provider (e.g., "openai"). Defaults to "openai".

Methods:
    start(query: str, num_result: int) -> list:
        Executes a web search for the given query, retrieves the content of top results,
        splits them into chunks, summarizes them using the summarization agent,
        and returns a list of dictionaries with URL, title, and summarized content.

Constructor & Destructor Documentation

__init__()

None iragent.models.InternetAgent.__init__	(		self,
		int	chunk_size,
		str	model,
		str	base_url,
		str	api_key,
		float	temperature = 0.1,
		int	max_token = 512,
		str	provider = "openai" )

Member Function Documentation

_summarize_page()

iragent.models.InternetAgent._summarize_page ( self, result, str query )

protected

Fetch one page and summarise it (runs in its own thread).

fast_start()

list[dict] iragent.models.InternetAgent.fast_start	(		self,
		str	query,
		int	num_result,
		int | None	max_workers = None )

A convenience wrapper that searches the Web, fetches the content of each hit,
breaks the text into token‑limited chunks, and asks a language‑model “summarizer”
to extract only the information relevant to the user’s query.

----------
Attributes
----------
chunk_size : int
    Maximum token length for each text chunk before it is passed to the
    summarization model.
summerize_agent : Agent
    A pre‑configured LLM “agent” used to turn a chunk of raw page text
    into a concise, query‑focused summary.

----------
Parameters
----------
chunk_size : int
    Token limit used when splitting page text.
model : str
    Name / identifier of the language model (e.g. ``"gpt-4o-2025-05-13"``).
base_url : str
    Base URL for the model’s API endpoint.
api_key : str
    API key or access token.
temperature : float, optional (default = 0.1)
    Sampling temperature.
max_token : int, optional (default = 512)
    Maximum length of each summary returned by the LLM.
provider : str, optional (default = ``"openai"``)
    Identifies the backend.  Special‑casing is included for ``"ollama"``
    because its local HTTP server dislikes shared clients in a pool of
    threads.

----------
Methods
----------
start(query, num_result)
    Serial implementation – easy to read, useful for debugging.
fast_start(query, num_result, max_workers=None)
    Threaded implementation that parallelises I/O for speed.
_summarize_page(result, query)
    Worker routine run by each thread in ``fast_start``.  Not public.

All methods return a ``list[dict]`` whose items look like::

    {
        "url":     "<page URL>",
        "title":   "<page title>",
        "content": "<summarised text>",
    }

start()

str iragent.models.InternetAgent.start	(		self,
		str	query,
			num_result )

Member Data Documentation

chunk_size

iragent.models.InternetAgent.chunk_size = chunk_size

summerize_agent

iragent.models.InternetAgent.summerize_agent

Initial value:

=  Agent(
            name="Summerize Agent",
            model=model,
            base_url=base_url,
            api_key=api_key,
            system_prompt=SUMMARIZER_PROMPT,
            temprature=temperature,
            max_token=max_token,
            provider=provider,
        )

---

The documentation for this class was generated from the following file:

• iragent/models.py