features/text_classifier/controller.py

import asyncio
import logging
import os
from io import BytesIO

from fastapi import Depends, HTTPException, UploadFile, status
from fastapi.security import HTTPAuthorizationCredentials, HTTPBearer

from .inferencer import analyze_text_with_sentences, classify_text
from .preprocess import parse_docx, parse_pdf, parse_txt

security = HTTPBearer()


def build_bias_summary(ai_likelihood: float) -> dict[str, object]:
    """Convert an AI likelihood score into a human-readable bias summary."""
    if ai_likelihood > 50:
        overall_bias = "AI"
        bias_statement = f"The text is biased toward AI-generated writing ({ai_likelihood}% AI likelihood)."
    elif ai_likelihood < 50:
        overall_bias = "Human"
        bias_statement = f"The text is biased toward human writing ({100 - ai_likelihood}% human likelihood)."
    else:
        overall_bias = "Balanced"
        bias_statement = "The text is balanced between AI and human writing."

    return {
        "overall_bias": overall_bias,
        "bias_statement": bias_statement,
    }


# Verify Bearer token from Authorization header
async def verify_token(credentials: HTTPAuthorizationCredentials = Depends(security)):
    token = credentials.credentials
    expected_token = os.getenv("MY_SECRET_TOKEN")
    if token != expected_token:
        raise HTTPException(
            status_code=status.HTTP_403_FORBIDDEN, detail="Invalid or expired token"
        )
    return token


# Classify plain text input
async def handle_text_analysis(text: str):
    text = text.strip()
    if not text or len(text.split()) < 10:
        raise HTTPException(
            status_code=400, detail="Text must contain at least 10 words"
        )
    if len(text) > 50000:
        raise HTTPException(
            status_code=413, detail="Text must be less than 50,000 characters"
        )

    label, perplexity, ai_likelihood = await asyncio.to_thread(classify_text, text)
    bias_summary = build_bias_summary(ai_likelihood)
    return {
        "result": label,
        "perplexity": round(perplexity, 2),
        "ai_likelihood": ai_likelihood,
        **bias_summary,
    }


# Extract text from uploaded files (.docx, .pdf, .txt)
async def extract_file_contents(file: UploadFile) -> str:
    content = await file.read()
    file_stream = BytesIO(content)

    if (
        file.content_type
        == "application/vnd.openxmlformats-officedocument.wordprocessingml.document"
    ):
        return parse_docx(file_stream)
    elif file.content_type == "application/pdf":
        return parse_pdf(file_stream)
    elif file.content_type == "text/plain":
        return parse_txt(file_stream)
    else:
        raise HTTPException(
            status_code=415,
            detail="Invalid file type. Only .docx, .pdf and .txt are allowed.",
        )


# Classify text from uploaded file
async def handle_file_upload(file: UploadFile):
    try:
        file_contents = await extract_file_contents(file)
        logging.info(f"Extracted text length: {len(file_contents)} characters")
        if len(file_contents) > 50000:
            return {
                "status_code": 413,
                "detail": "Text must be less than 50,000 characters",
            }

        cleaned_text = file_contents.replace("\n", " ").replace("\t", " ").strip()
        if not cleaned_text:
            raise HTTPException(
                status_code=400,
                detail="The uploaded file is empty or only contains whitespace.",
            )
        # print(f"Cleaned text: '{cleaned_text}'")  # Debugging statement
        label, perplexity, ai_likelihood = await asyncio.to_thread(
            classify_text, cleaned_text
        )
        return {
            "content": file_contents,
            "result": label,
            "perplexity": round(perplexity, 2),
            "ai_likelihood": ai_likelihood,
        }
    except Exception as e:
        logging.error(f"Error processing file: {e}")
        raise HTTPException(status_code=500, detail="Error processing the file")


async def handle_sentence_level_analysis(text: str):
    text = text.strip()
    if not text or len(text.split()) < 10:
        raise HTTPException(
            status_code=400, detail="Text must contain at least 10 words"
        )
    if len(text) > 50000:
        raise HTTPException(
            status_code=413, detail="Text must be less than 50,000 characters"
        )

    result = await asyncio.to_thread(analyze_text_with_sentences, text)
    return result


# Analyze each sentence from uploaded file
async def handle_file_sentence(file: UploadFile):
    try:
        file_contents = await extract_file_contents(file)
        if len(file_contents) > 50000:
            # raise HTTPException(status_code=413, detail="Text must be less than 10,000 characters")
            return {
                "status_code": 413,
                "detail": "Text must be less than 50,000 characters",
            }

        cleaned_text = file_contents.replace("\n", " ").replace("\t", " ").strip()
        if not cleaned_text:
            raise HTTPException(
                status_code=400,
                detail="The uploaded file is empty or only contains whitespace.",
            )

        result = await handle_sentence_level_analysis(cleaned_text)
        return {"content": file_contents, **result}
    except HTTPException:
        raise
    except Exception as e:
        logging.error(f"Error processing file: {e}")
        raise HTTPException(status_code=500, detail="Error processing the file")


def classify(text: str):
    return classify_text(text)