fix(ci): add quality extra + skip tests when datasets missing (#6)

- CI needs datasets package (in quality extra) for test_cli_extended
- Skip test_analyze_from_file tests when datasets not installed

Co-authored-by: liuxiaotong <liuxiaotong@knowlyr.com>

Author: liuxiaotong

.github/workflows/ci.yml

@@ -25,7 +25,7 @@ jobs:
       - name: Install dependencies
         run: |
           python -m pip install --upgrade pip
-          pip install -e ".[dev]"
+          pip install -e ".[dev,quality]"
 
       - name: Lint with ruff
         run: ruff check .

tests/test_local_source.py

@@ -1,12 +1,15 @@
 """Tests for local file source support (CSV, Parquet, JSONL)."""
 
 import csv
+import importlib
 import json
 import os
 import tempfile
 import unittest
 from pathlib import Path
 
+_has_datasets = importlib.util.find_spec("datasets") is not None
+
 from click.testing import CliRunner
 
 from datarecipe.analyzer import DatasetAnalyzer
@@ -294,6 +297,7 @@ def _create_csv(self):
                 writer.writerow({"text": f"This is sample text number {i} for quality analysis testing.", "label": "pos"})
         return path
 
+    @unittest.skipUnless(_has_datasets, "requires 'datasets' package (install via pip install -e '.[quality]')")
     def test_analyze_from_file(self):
         from datarecipe.quality_metrics import QualityAnalyzer
 
@@ -303,6 +307,7 @@ def test_analyze_from_file(self):
 
         self.assertGreater(report.sample_size, 0)
 
+    @unittest.skipUnless(_has_datasets, "requires 'datasets' package (install via pip install -e '.[quality]')")
     def test_analyze_from_file_auto_detect_field(self):
         from datarecipe.quality_metrics import QualityAnalyzer