Add test first part

tests/integration/test_app.py

@@ -0,0 +1,35 @@
+import pytest
+from fastapi.testclient import TestClient
+from src.app import app
+
+client = TestClient(app)
+
+def test_home_page():
+    response = client.get("/")
+    assert response.status_code == 200
+    assert "Benvenuto" in response.text
+
+def test_predict_endpoint_get():
+    response = client.get("/predict")
+    assert response.status_code == 200
+    assert "Testa il Modello" in response.text
+
+def test_predict_endpoint_post():
+    response = client.post("/predict", data={"text": "I love this!"})
+    assert response.status_code == 200
+    assert any(label in response.text for label in ["positive", "neutral", "negative"])
+
+@pytest.mark.asyncio
+def test_random_tweet_page():
+    response = client.get("/random_tweet")
+    assert response.status_code == 200
+    assert "Sentiment" in response.text
+
+@pytest.mark.asyncio
+def test_random_youtube_page():
+    response = client.get("/random_youtube_comment")
+    assert response.status_code == 200
+    assert "Sentiment" in response.text
+
+
+

tests/unit/test_data.py

@@ -0,0 +1,40 @@
+# tests/test_data.py
+import os
+import subprocess
+from datasets import load_from_disk
+
+TWEET_PROCESSED_PATH = "data/processed/tweet_eval_tokenized"
+YT_PROCESSED_PATH = "data/processed/youtube_comments"
+
+def run_data_preparation(dataset_name):
+    """Esegue lo script di data preparation per il dataset richiesto."""
+    print(f"⚙️  Avvio data_preparation.py per il dataset: {dataset_name}")
+    subprocess.run(
+        ["python", "src/data_preparation.py", "--dataset", dataset_name],
+        check=True
+    )
+
+def test_tweet_eval_dataset_exists_or_create():
+    """Controlla o crea il dataset Tweet Eval preprocessato."""
+    if not os.path.exists(TWEET_PROCESSED_PATH):
+        run_data_preparation("tweet_eval")
+    assert os.path.exists(TWEET_PROCESSED_PATH), "Tweet Eval non disponibile dopo la preparazione"
+
+def test_youtube_dataset_exists_or_create():
+    """Controlla o crea il dataset YouTube preprocessato."""
+    if not os.path.exists(YT_PROCESSED_PATH):
+        run_data_preparation("youtube")
+    assert os.path.exists(YT_PROCESSED_PATH), "YouTube dataset non disponibile dopo la preparazione"
+
+def test_tweet_eval_structure():
+    """Verifica che il dataset Tweet Eval abbia la struttura corretta."""
+    ds = load_from_disk(TWEET_PROCESSED_PATH)
+    assert "text" in ds["test"].features, "Campo 'text' mancante in Tweet Eval"
+    assert "label" in ds["test"].features, "Campo 'label' mancante in Tweet Eval"
+
+def test_youtube_structure():
+    """Verifica che il dataset YouTube abbia la struttura corretta."""
+    ds = load_from_disk(YT_PROCESSED_PATH)
+    assert "CommentText" in ds["train"].features or "CommentText" in ds["train"].features, \
+        "Campo testuale mancante in YouTube dataset"
+    assert "Sentiment" in ds["train"].features, "Campo 'label' mancante in YouTube dataset"

tests/unit/test_model.py

@@ -0,0 +1,26 @@
+from transformers import AutoTokenizer, AutoModelForSequenceClassification
+import torch
+
+MODEL_NAME = "cardiffnlp/twitter-roberta-base-sentiment-latest"
+LABELS = ["negative", "neutral", "positive"]
+
+tokenizer = AutoTokenizer.from_pretrained(MODEL_NAME)
+model = AutoModelForSequenceClassification.from_pretrained(MODEL_NAME)
+
+def test_model_loads():
+    assert model is not None
+    assert tokenizer is not None
+
+def test_model_prediction_shape():
+    text = "I love this product!"
+    inputs = tokenizer(text, return_tensors="pt")
+    with torch.no_grad():
+        outputs = model(**inputs)
+    assert outputs.logits.shape[-1] == len(LABELS)
+
+def test_sentiment_confidence():
+    text = "I hate this"
+    inputs = tokenizer(text, return_tensors="pt")
+    with torch.no_grad():
+        probs = torch.nn.functional.softmax(model(**inputs).logits, dim=-1)
+    assert torch.isclose(probs.sum(), torch.tensor(1.0), atol=1e-3)