tasks/text.py

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

import skops
from skops.hub_utils import download
from skops.io import load

import pandas as pd

from huggingface_hub import hf_hub_download
import joblib

REPO_ID = "kantundpeterpan/frugal-ai-toy"
FILENAME = "tfidf_rf.skops"

# import nltk
# from nltk.tokenize import WordPunctTokenizer
# from nltk.stem import WordNetLemmatizer
# from nltk.corpus import stopwords
# import string
# nltk.download('stopwords')

# stop = set(stopwords.words('english') + list(string.punctuation))

# def tokenize_quote(r):
#     tokens = nltk.word_tokenize(r.lower())
#     cleaned = [word for word in tokens if word not in stop]
#     return cleaned

# def lemmatize_tokens(tokens: list):
#     return [lemmatizer.lemmatize(t) for t in tokens]

# def lemmatize_X(X):
#     return X.quote.apply(tokenize_quote).apply(lemmatize_tokens).apply(lambda x: " ".join(x))

import random

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

router = APIRouter()

DESCRIPTION = "tfidf-rf"
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
    }

    # 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"].train_test_split(test_size=request.test_size, seed=request.test_seed)
    test_dataset = train_test["test"]
    test_df = pd.DataFrame(test_dataset)
    
    # 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.
    #--------------------------------------------------------------------------------------------   

    
    #download model
    download(repo_id = "kantundpeterpan/frugal-ai-toy", dst = "skops_test")
    #get unknwown types
    unknown = skops.io.get_untrusted_types(file = "skops_test/tfidf_rf.skops")
    
    #load model
    model = model = load("skops_test/tfidf_rf.skops", trusted = unknown)
    
    # Make predictions
    true_labels = test_dataset["label"]
    predictions = [
        LABEL_MAPPING[r] for r in model.predict(test_dataset)    
    ]

    #--------------------------------------------------------------------------------------------
    # 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