app.py

from fastapi import FastAPI, Request
from pydantic import BaseModel
import torch
import torch.nn as nn
from transformers import DistilBertTokenizer, DistilBertModel, AutoModel, AutoTokenizer
from langdetect import detect
from huggingface_hub import snapshot_download
import os


# Device
device = torch.device("cuda" if torch.cuda.is_available() else "cpu")

# Download model repos from HF Hub
english_repo = snapshot_download("koyu008/English_Toxic_Classifier")
hinglish_repo = snapshot_download("koyu008/HInglish_comment_classifier")

# Tokenizers
english_tokenizer = DistilBertTokenizer.from_pretrained("distilbert-base-uncased")
hinglish_tokenizer = AutoTokenizer.from_pretrained("xlm-roberta-base")


# English Model
class ToxicBERT(nn.Module):
    def __init__(self):
        super().__init__()
        self.bert = DistilBertModel.from_pretrained("distilbert-base-uncased")
        self.dropout = nn.Dropout(0.3)
        self.classifier = nn.Linear(self.bert.config.hidden_size, 6)

    def forward(self, input_ids, attention_mask):
        output = self.bert(input_ids=input_ids, attention_mask=attention_mask).last_hidden_state[:, 0]
        return self.classifier(self.dropout(output))


# Hinglish Model
class HinglishToxicClassifier(nn.Module):
    def __init__(self):
        super().__init__()
        self.bert = AutoModel.from_pretrained("xlm-roberta-base")
        hidden_size = self.bert.config.hidden_size
        self.pool = lambda hidden: torch.cat([
            hidden.mean(dim=1),
            hidden.max(dim=1).values
        ], dim=1)
        self.bottleneck = nn.Sequential(
            nn.Linear(2 * hidden_size, hidden_size),
            nn.ReLU(),
            nn.Dropout(0.2)
        )
        self.classifier = nn.Linear(hidden_size, 2)

    def forward(self, input_ids, attention_mask):
        hidden = self.bert(input_ids=input_ids, attention_mask=attention_mask).last_hidden_state
        pooled = self.pool(hidden)
        x = self.bottleneck(pooled)
        return self.classifier(x)


# Instantiate and load models
english_model = ToxicBERT().to(device)
english_model.load_state_dict(torch.load(os.path.join(english_repo, "bert_toxic_classifier.pt"), map_location=device))
english_model.eval()

hinglish_model = HinglishToxicClassifier().to(device)
hinglish_model.load_state_dict(torch.load(os.path.join(hinglish_repo, "best_hinglish_model.pt"), map_location=device))
hinglish_model.eval()

# Labels
english_labels = ['toxic', 'severe toxic', 'obscene', 'threat', 'insult', 'identity hate']
hinglish_labels = ['not toxic', 'toxic']

# FastAPI
app = FastAPI()

from fastapi.middleware.cors import CORSMiddleware

app.add_middleware(
    CORSMiddleware,
    allow_origins=["*"],  # Or restrict to your frontend domain
    allow_credentials=True,
    allow_methods=["*"],
    allow_headers=["*"],
)


class TextIn(BaseModel):
    text: str


@app.post("/api/predict")
@app.post("/api/predict")
def predict(data: TextIn):
    text = data.text
    try:
        lang = detect(text)
    except:
        lang = "unknown"

    if lang == "en":
        tokenizer = english_tokenizer
        model = english_model
        inputs = tokenizer(text, return_tensors="pt", padding=True, truncation=True).to(device)
        with torch.no_grad():
            outputs = model(**inputs)
            probs = torch.sigmoid(outputs).squeeze().cpu().tolist()
        return {"language": "English", "predictions": dict(zip(english_labels, probs))}

    else:
        tokenizer = hinglish_tokenizer
        model = hinglish_model
        inputs = tokenizer(text, return_tensors="pt", padding=True, truncation=True).to(device)
        with torch.no_grad():
            outputs = model(**inputs)
            probs = torch.softmax(outputs, dim=1).squeeze().cpu().tolist()
        return {"language": "Hinglish", "predictions": dict(zip(hinglish_labels, probs))}

@app.get("/")
def root():
    return {"message": "API is running"}