File size: 1,295 Bytes
c2fb848 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 |
# summarizer_module/__init__.py
from transformers import AutoTokenizer, AutoModelForCausalLM, pipeline
from device_config import get_device
import torch
device = get_device()
# Use a small local model (e.g., Phi-2)
MODEL_ID = "microsoft/phi-2" # Ensure it's downloaded and cached locally
# Load model and tokenizer
model = AutoModelForCausalLM.from_pretrained(MODEL_ID).to(device)
tokenizer = AutoTokenizer.from_pretrained(MODEL_ID)
summarizer = pipeline("text-generation", model=model, tokenizer=tokenizer)
def summarize_flowchart(flowchart_json):
"""
Given a flowchart JSON with 'start' and 'steps', returns a plain English explanation
formatted as bullets and sub-bullets.
Args:
flowchart_json (dict): Structured representation of flowchart
Returns:
str: Bullet-style natural language summary of the logic
"""
prompt = (
"Turn the following flowchart into a bullet-point explanation in plain English.\n"
"Use bullets for steps and sub-bullets for branches.\n"
"\n"
f"Flowchart JSON:\n{flowchart_json}\n"
"\nExplanation:"
)
result = summarizer(prompt, max_new_tokens=300, do_sample=False)[0]["generated_text"]
explanation = result.split("Explanation:")[-1].strip()
return explanation
|