tasks/text.py

from fastapi import APIRouter
from datetime import datetime
from datasets import load_dataset
from sklearn.metrics import accuracy_score
import random

from .utils.evaluation import TextEvaluationRequest
from .utils.emissions import tracker, clean_emissions_data, get_space_info


import tensorflow as tf
from huggingface_hub import hf_hub_download
from transformers import ElectraTokenizer


router = APIRouter()

DESCRIPTION = "Electra with balanced dataset"
ROUTE = "/text"



@router.post(ROUTE, tags=["Text Task"], 
             description=DESCRIPTION)
async def evaluate_text(request: TextEvaluationRequest):
    """
    Evaluate text classification for climate disinformation detection.
    
    Current Model: Random Baseline
    - Makes random predictions from the label space (0-7)
    - Used as a baseline for comparison
    """
    # Get space info
    username, space_url = get_space_info()

    # Define the label mapping
    LABEL_MAPPING = {
        "0_not_relevant": 0,
        "1_not_happening": 1,
        "2_not_human": 2,
        "3_not_bad": 3,
        "4_solutions_harmful_unnecessary": 4,
        "5_science_unreliable": 5,
        "6_proponents_biased": 6,
        "7_fossil_fuels_needed": 7
    }

    # Download our pre-trained model from Hugging Face
    model_path = hf_hub_download(repo_id="julianaconsuegra/electra-base-climate-disinformation", filename="tf_model.h5")

    # Load the model
    model = tf.keras.models.load_model(model_path)

    # Load and prepare the dataset
    dataset = load_dataset(request.dataset_name)

    # Convert string labels to integers
    dataset = dataset.map(lambda x: {"label": LABEL_MAPPING[x["label"]]})

    # Split dataset
    train_test = dataset["train"]
    test_dataset = dataset["test"]
    
    # Start tracking emissions
    tracker.start()
    tracker.start_task("inference")

    #--------------------------------------------------------------------------------------------
    # YOUR MODEL INFERENCE CODE HERE
    # Update the code below to replace the random baseline by your model inference within the inference pass where the energy consumption and emissions are tracked.
    #--------------------------------------------------------------------------------------------   
    # Load ELECTRA tokenizer
    tokenizer = ElectraTokenizer.from_pretrained("google/electra-base-discriminator")
    
    # Tokenize test data with same parameters as training
    inputs = tokenizer(
        test_dataset["text"],
        truncation=True,
        padding="max_length",
        return_tensors="tf"
    )
    
    # Run model prediction
    logits = model.predict({
        "input_ids": inputs["input_ids"],
        "attention_mask": inputs["attention_mask"]
    })
    predictions = tf.argmax(logits, axis=1).numpy()
    
    # Get ground truth labels
    true_labels = test_dataset["label"]

    #--------------------------------------------------------------------------------------------
    # YOUR MODEL INFERENCE STOPS HERE
    #--------------------------------------------------------------------------------------------   

    
    # Stop tracking emissions
    emissions_data = tracker.stop_task()
    
    # Calculate accuracy
    accuracy = accuracy_score(true_labels, predictions)
    
    # Prepare results dictionary
    results = {
        "username": username,
        "space_url": space_url,
        "submission_timestamp": datetime.now().isoformat(),
        "model_description": DESCRIPTION,
        "accuracy": float(accuracy),
        "energy_consumed_wh": emissions_data.energy_consumed * 1000,
        "emissions_gco2eq": emissions_data.emissions * 1000,
        "emissions_data": clean_emissions_data(emissions_data),
        "api_route": ROUTE,
        "dataset_config": {
            "dataset_name": request.dataset_name,
            "test_size": request.test_size,
            "test_seed": request.test_seed
        }
    }
    
    return results