Initial upload: shell-speak CLI tool with natural language to shell command conversion

tests/test_pattern_matching.py

@@ -0,0 +1,140 @@
+"""Tests for pattern matching engine."""
+
+import os
+import sys
+import tempfile
+from pathlib import Path
+
+import pytest
+
+sys.path.insert(0, str(Path(__file__).parent.parent))
+
+
+class TestNLP:
+    """Tests for NLP preprocessing functions."""
+
+    def test_normalize_text(self):
+        from shell_speak.nlp import normalize_text
+
+        assert normalize_text("  Hello World  ") == "hello world"
+        assert normalize_text("Hello\tWorld") == "hello world"
+        assert normalize_text("multiple   spaces") == "multiple spaces"
+
+    def test_tokenize(self):
+        from shell_speak.nlp import tokenize
+
+        assert tokenize("hello world") == ["hello", "world"]
+        assert tokenize("List running containers") == ["list", "running", "containers"]
+        assert tokenize("") == []
+
+    def test_extract_keywords(self):
+        from shell_speak.nlp import extract_keywords
+
+        keywords = extract_keywords("list running containers")
+        assert "list" in keywords
+        assert "running" in keywords
+        assert "containers" in keywords
+
+    def test_extract_keywords_removes_stopwords(self):
+        from shell_speak.nlp import extract_keywords
+
+        keywords = extract_keywords("the a is are")
+        assert len(keywords) == 0
+
+    def test_calculate_similarity(self):
+        from shell_speak.nlp import calculate_similarity
+
+        assert calculate_similarity("list containers", "list containers") == 1.0
+        assert calculate_similarity("list containers", "show running containers") > 0.1
+        assert calculate_similarity("list containers", "git commit") == 0.0
+
+
+class TestPatternMatcher:
+    """Tests for pattern matching functionality."""
+
+    @pytest.fixture(autouse=True)
+    def setup(self, tmp_path, sample_docker_yaml):
+        from shell_speak.config import ensure_data_dir
+        from shell_speak.library import get_loader
+
+        docker_file = tmp_path / "docker.yaml"
+        docker_file.write_text(sample_docker_yaml)
+
+        os.environ["SHELL_SPEAK_DATA_DIR"] = str(tmp_path)
+
+        loader = get_loader()
+        loader._loaded = False
+        loader._patterns = []
+
+    def test_match_existing_pattern(self):
+        from shell_speak.matcher import get_matcher
+
+        matcher = get_matcher()
+        match = matcher.match("list running containers")
+
+        assert match is not None
+        assert match.command == "docker ps"
+        assert match.confidence > 0.5
+
+    def test_match_partial_pattern(self):
+        from shell_speak.matcher import get_matcher
+
+        matcher = get_matcher()
+        match = matcher.match("show running containers")
+
+        assert match is not None
+        assert "docker" in match.command.lower()
+
+    def test_match_no_match(self):
+        from shell_speak.matcher import get_matcher
+
+        matcher = get_matcher()
+        match = matcher.match("xyz unknown query that does not exist")
+
+        assert match is None
+
+    def test_match_with_tool_filter(self):
+        from shell_speak.matcher import get_matcher
+
+        matcher = get_matcher()
+        match = matcher.match("list running containers", tool="docker")
+
+        assert match is not None
+        assert match.pattern.tool == "docker"
+
+    def test_template_substitution(self):
+        from shell_speak.matcher import get_matcher
+
+        matcher = get_matcher()
+        match = matcher.match("run a container named test with image nginx")
+
+        assert match is not None
+        assert "docker" in match.command.lower()
+
+
+class TestCommandMatch:
+    """Tests for CommandMatch dataclass."""
+
+    def test_command_match_creation(self):
+        from shell_speak.models import CommandMatch, CommandPattern
+
+        pattern = CommandPattern(
+            name="test_pattern",
+            tool="test",
+            description="Test pattern",
+            patterns=["test query"],
+            template="echo test",
+            explanation="Test explanation",
+        )
+
+        match = CommandMatch(
+            pattern=pattern,
+            confidence=0.9,
+            matched_query="test query",
+            command="echo test",
+            explanation="Test explanation",
+        )
+
+        assert match.pattern.name == "test_pattern"
+        assert match.confidence == 0.9
+        assert match.command == "echo test"