wehub-resource-sync
121 lines
4.2 KiB
Python
121 lines
4.2 KiB
Python
"""
|
|
SpeechGraph Module
|
|
"""
|
|
|
|
from typing import Optional, Type
|
|
|
|
from pydantic import BaseModel
|
|
|
|
from ..models import OpenAITextToSpeech
|
|
from ..nodes import FetchNode, GenerateAnswerNode, ParseNode, TextToSpeechNode
|
|
from ..utils.logging import get_logger
|
|
from ..utils.save_audio_from_bytes import save_audio_from_bytes
|
|
from .abstract_graph import AbstractGraph
|
|
from .base_graph import BaseGraph
|
|
|
|
logger = get_logger(__name__)
|
|
|
|
|
|
class SpeechGraph(AbstractGraph):
|
|
"""
|
|
SpeechyGraph is a scraping pipeline that scrapes the web, provide an answer
|
|
to a given prompt, and generate an audio file.
|
|
|
|
Attributes:
|
|
prompt (str): The prompt for the graph.
|
|
source (str): The source of the graph.
|
|
config (dict): Configuration parameters for the graph.
|
|
schema (BaseModel): The schema for the graph output.
|
|
llm_model: An instance of a language model client, configured for generating answers.
|
|
embedder_model: An instance of an embedding model clienta
|
|
configured for generating embeddings.
|
|
verbose (bool): A flag indicating whether to show print statements during execution.
|
|
headless (bool): A flag indicating whether to run the graph in headless mode.
|
|
model_token (int): The token limit for the language model.
|
|
|
|
Args:
|
|
prompt (str): The prompt for the graph.
|
|
source (str): The source of the graph.
|
|
config (dict): Configuration parameters for the graph.
|
|
schema (BaseModel): The schema for the graph output.
|
|
|
|
Example:
|
|
>>> speech_graph = SpeechGraph(
|
|
... "List me all the attractions in Chioggia and generate an audio summary.",
|
|
... "https://en.wikipedia.org/wiki/Chioggia",
|
|
... {"llm": {"model": "openai/gpt-3.5-turbo"}}
|
|
"""
|
|
|
|
def __init__(
|
|
self,
|
|
prompt: str,
|
|
source: str,
|
|
config: dict,
|
|
schema: Optional[Type[BaseModel]] = None,
|
|
):
|
|
super().__init__(prompt, config, source, schema)
|
|
|
|
self.input_key = "url" if source.startswith("http") else "local_dir"
|
|
|
|
def _create_graph(self) -> BaseGraph:
|
|
"""
|
|
Creates the graph of nodes representing the workflow for web scraping and audio generation.
|
|
|
|
Returns:
|
|
BaseGraph: A graph instance representing the web scraping and audio generation workflow.
|
|
"""
|
|
|
|
fetch_node = FetchNode(input="url | local_dir", output=["doc"])
|
|
|
|
parse_node = ParseNode(
|
|
input="doc",
|
|
output=["parsed_doc"],
|
|
node_config={"chunk_size": self.model_token, "llm_model": self.llm_model},
|
|
)
|
|
|
|
generate_answer_node = GenerateAnswerNode(
|
|
input="user_prompt & (relevant_chunks | parsed_doc | doc)",
|
|
output=["answer"],
|
|
node_config={
|
|
"llm_model": self.llm_model,
|
|
"additional_info": self.config.get("additional_info"),
|
|
"schema": self.schema,
|
|
},
|
|
)
|
|
|
|
text_to_speech_node = TextToSpeechNode(
|
|
input="answer",
|
|
output=["audio"],
|
|
node_config={"tts_model": OpenAITextToSpeech(self.config["tts_model"])},
|
|
)
|
|
|
|
return BaseGraph(
|
|
nodes=[fetch_node, parse_node, generate_answer_node, text_to_speech_node],
|
|
edges=[
|
|
(fetch_node, parse_node),
|
|
(parse_node, generate_answer_node),
|
|
(generate_answer_node, text_to_speech_node),
|
|
],
|
|
entry_point=fetch_node,
|
|
graph_name=self.__class__.__name__,
|
|
)
|
|
|
|
def run(self) -> str:
|
|
"""
|
|
Executes the scraping process and returns the answer to the prompt.
|
|
|
|
Returns:
|
|
str: The answer to the prompt.
|
|
"""
|
|
|
|
inputs = {"user_prompt": self.prompt, self.input_key: self.source}
|
|
self.final_state, self.execution_info = self.graph.execute(inputs)
|
|
|
|
audio = self.final_state.get("audio", None)
|
|
if not audio:
|
|
raise ValueError("No audio generated from the text.")
|
|
save_audio_from_bytes(audio, self.config.get("output_path", "output.mp3"))
|
|
logger.info("Audio saved to %s", self.config.get("output_path", "output.mp3"))
|
|
|
|
return self.final_state.get("answer", "No answer found.")
|