feat: add conversation persistence and context management layer

- SQLAlchemy async database layer (SQLite dev, PostgreSQL prod)
- Conversation and Message models with UUID primary keys
- Token counting utilities using litellm
- Context summarization at 80% token threshold
- REST API endpoints for multi-turn conversations
- 19 conversation tests, 6 token tests (176 total passing)

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>

webber-api/tests/test_tokens.py

@@ -0,0 +1,90 @@
+"""
+Tests for token counting utilities.
+"""
+import pytest
+
+from src.shared.tokens import count_tokens, count_message_tokens, estimate_tokens
+
+
+class TestTokenCounting:
+    """Tests for token counting functions."""
+
+    def test_estimate_tokens_basic(self):
+        """Test basic token estimation."""
+        text = "Hello world"
+        tokens = estimate_tokens(text)
+        # ~4 chars per token
+        assert tokens == len(text) // 4
+
+    def test_estimate_tokens_empty(self):
+        """Test estimation with empty string."""
+        assert estimate_tokens("") == 0
+
+    def test_estimate_tokens_long_text(self):
+        """Test estimation with longer text."""
+        text = "a" * 400
+        tokens = estimate_tokens(text)
+        assert tokens == 100
+
+    def test_count_tokens_basic(self):
+        """Test actual token counting."""
+        text = "Hello, how are you today?"
+        tokens = count_tokens(text)
+        # Should return reasonable token count
+        assert tokens > 0
+        assert tokens < len(text)  # Should be fewer tokens than characters
+
+    def test_count_tokens_empty(self):
+        """Test counting empty string."""
+        tokens = count_tokens("")
+        assert tokens == 0
+
+    def test_count_message_tokens_single(self):
+        """Test counting tokens in single message."""
+        messages = [{"role": "user", "content": "Hello"}]
+        tokens = count_message_tokens(messages)
+        assert tokens > 0
+
+    def test_count_message_tokens_multiple(self):
+        """Test counting tokens in multiple messages."""
+        messages = [
+            {"role": "user", "content": "Hello, how are you?"},
+            {"role": "assistant", "content": "I'm doing well, thank you!"},
+        ]
+        tokens = count_message_tokens(messages)
+        # Should be more than single message
+        single_tokens = count_message_tokens([messages[0]])
+        assert tokens > single_tokens
+
+    def test_count_message_tokens_empty_list(self):
+        """Test counting empty message list."""
+        tokens = count_message_tokens([])
+        # litellm may return small overhead even for empty list
+        assert tokens < 10
+
+
+class TestTokenCountingAccuracy:
+    """Tests for token counting accuracy."""
+
+    def test_code_tokens_reasonable(self):
+        """Test that code is tokenized reasonably."""
+        code = """
+def hello_world():
+    print("Hello, World!")
+    return True
+"""
+        tokens = count_tokens(code)
+        # Code should have reasonable token count
+        assert 10 < tokens < 100
+
+    def test_special_characters(self):
+        """Test tokenization of special characters."""
+        text = "Hello! @#$%^&*() World?"
+        tokens = count_tokens(text)
+        assert tokens > 0
+
+    def test_unicode_text(self):
+        """Test tokenization of unicode text."""
+        text = "Hello 世界 🌍"
+        tokens = count_tokens(text)
+        assert tokens > 0