from fastapi import FastAPI
from pydantic import BaseModel
from transformers import pipeline

app = FastAPI()
summarizer = pipeline("summarization", model="facebook/bart-large-cnn")

class TextInput(BaseModel):
    text: str

@app.get("/")
async def root():
    return {"message": "Welcome to the Text Summarization API!"}

@app.post("/summarize")
async def summarize_text(input: TextInput):
    # Count approximate number of words (could be improved with tokenizer count in the future)
    word_count = len(input.text.split())
    

    if word_count < 50:
        max_length = max(10, word_count // 2)  
        min_length = max(3, word_count // 4)   
    elif word_count < 200:
        max_length = max(50, word_count // 3)
        min_length = max(15, word_count // 6)
    else:
        max_length = max(100, word_count // 4)
        min_length = max(30, word_count // 8)
    
    # Prevent max_length from being too large (BART has token limits)
    max_length = min(max_length, 1024)
    
    # Generate summary with dynamic parameters
    summary = summarizer(
        input.text, 
        max_length=max_length, 
        min_length=min_length, 
        do_sample=True,
        temperature=0.7,
        num_beams=4
    )
    
    return {
        "summary": summary[0]["summary_text"],
        "parameters_used": {
            "input_word_count": word_count,
            "max_length": max_length,
            "min_length": min_length
        }
    }