app.py

from fastapi import FastAPI
from pydantic import BaseModel
from transformers import pipeline, AutoTokenizer
import nltk
import os
import uvicorn

nltk.download('punkt', quiet=True)

app = FastAPI()

HF_AUTH_TOKEN = os.getenv("HF_TOKEN") 

MODEL_NAME = "VincentMuriuki/legal-summarizer"
summarizer = pipeline("summarization", model=MODEL_NAME, token=HF_AUTH_TOKEN)
tokenizer = AutoTokenizer.from_pretrained(MODEL_NAME, token=HF_AUTH_TOKEN)

class SummarizeInput(BaseModel):
    text: str

class ChunkInput(BaseModel):
    text: str
    max_tokens: int = 512  # Default chunk size

# Summarize endpoint
@app.post("/summarize")
def summarize_text(data: SummarizeInput):
    summary = summarizer(data.text, max_length=150, min_length=30, do_sample=False)
    return {"summary": summary[0]["summary_text"]}

@app.post("/chunk")
def chunk_text(data: ChunkInput):
    sentences = nltk.sent_tokenize(data.text)
    chunks = []
    current_chunk = ""
    current_token_count = 0

    for sentence in sentences:
        token_count = len(tokenizer.tokenize(sentence))
        if current_token_count + token_count > data.max_tokens:
            if current_chunk:
                chunks.append(current_chunk.strip())
            current_chunk = sentence
            current_token_count = token_count
        else:
            current_chunk = f"{current_chunk} {sentence}".strip()
            current_token_count += token_count

    if current_chunk:
        chunks.append(current_chunk.strip())

    return {"chunks": chunks}
if __name__ == "__main__":
    uvicorn.run(app, host="0.0.0.0", port=7860)