Yellhorn MCP

"""Unit tests for comment formatting utilities.""" from datetime import datetime, timezone import pytest from yellhorn_mcp.models.metadata_models import CompletionMetadata, SubmissionMetadata from yellhorn_mcp.utils.comment_utils import ( extract_urls, format_completion_comment, format_submission_comment, ) class TestFormatSubmissionComment: """Test cases for format_submission_comment function.""" def test_basic_submission_comment(self): """Test formatting a basic submission comment.""" metadata = SubmissionMetadata( status="Generating workplan...", model_name="gemini-2.5-pro", search_grounding_enabled=True, yellhorn_version="0.5.0", submitted_urls=None, codebase_reasoning_mode="full", timestamp=datetime(2025, 1, 6, 12, 0, 0, tzinfo=timezone.utc), ) result = format_submission_comment(metadata) assert "## 🚀 Generating workplan..." in result assert "**Model**: `gemini-2.5-pro`" in result assert "**Search Grounding**: ✅ Enabled" in result assert "**Codebase Reasoning**: `full`" in result assert "**Yellhorn Version**: `0.5.0`" in result assert "**Submitted**: 2025-01-06 12:00:00 UTC" in result assert "_This issue will be updated once generation is complete._" in result def test_submission_comment_with_urls(self): """Test formatting a submission comment with URLs.""" metadata = SubmissionMetadata( status="Generating judgement...", model_name="gpt-4o", search_grounding_enabled=False, yellhorn_version="0.5.0", submitted_urls=["https://example.com/api-docs", "https://github.com/user/repo"], codebase_reasoning_mode="lsp", timestamp=datetime(2025, 1, 6, 12, 0, 0, tzinfo=timezone.utc), ) result = format_submission_comment(metadata) assert "## 🚀 Generating judgement..." in result assert "**Model**: `gpt-4o`" in result assert "**Search Grounding**: ❌ Disabled" in result assert "**Referenced URLs**:" in result assert "- https://example.com/api-docs" in result assert "- https://github.com/user/repo" in result class TestFormatCompletionComment: """Test cases for format_completion_comment function.""" def test_successful_completion_comment(self): """Test formatting a successful completion comment with all fields.""" metadata = CompletionMetadata( model_name="gemini-1.5-pro-latest", status="✅ Workplan generated successfully", generation_time_seconds=42.5, input_tokens=10000, output_tokens=2000, total_tokens=12000, estimated_cost=0.1234, model_version_used="gemini-2.5-pro", system_fingerprint=None, search_results_used=5, finish_reason="stop", safety_ratings=[ {"category": "HARM_CATEGORY_DANGEROUS_CONTENT", "probability": "NEGLIGIBLE"} ], context_size_chars=50000, warnings=None, timestamp=datetime(2025, 1, 6, 12, 0, 42, tzinfo=timezone.utc), ) result = format_completion_comment(metadata) assert "## ✅ Workplan generated successfully" in result assert "### Generation Details" in result assert "**Time**: 42.5 seconds" in result assert "**Completed**: 2025-01-06 12:00:42 UTC" in result assert ( "**Model Used**: `gemini-2.5-pro`" in result ) # Uses model_version_used when available assert "### Token Usage" in result assert "**Input Tokens**: 10,000" in result assert "**Output Tokens**: 2,000" in result assert "**Total Tokens**: 12,000" in result assert "**Estimated Cost**: $0.1234" in result assert "**Search Results Used**: 5" in result assert "**Context Size**: 50,000 characters" in result assert "**Finish Reason**: `stop`" in result assert "### Safety Ratings" in result assert "- **HARM_CATEGORY_DANGEROUS_CONTENT**: NEGLIGIBLE" in result def test_failed_completion_comment(self): """Test formatting a failed completion comment.""" metadata = CompletionMetadata( model_name="gemini-1.5-pro-latest", status="⚠️ Workplan generation failed", generation_time_seconds=10.2, input_tokens=None, output_tokens=None, total_tokens=None, estimated_cost=None, model_version_used=None, system_fingerprint=None, search_results_used=None, finish_reason="error", safety_ratings=None, context_size_chars=None, warnings=["API rate limit exceeded", "Please try again later"], timestamp=datetime(2025, 1, 6, 12, 0, 10, tzinfo=timezone.utc), ) result = format_completion_comment(metadata) assert "## ⚠️ Workplan generation failed" in result assert "**Time**: 10.2 seconds" in result assert "**Finish Reason**: `error`" in result assert "### ⚠️ Warnings" in result assert "- API rate limit exceeded" in result assert "- Please try again later" in result # Ensure token usage section is not included when no tokens assert "### Token Usage" not in result def test_openai_completion_comment(self): """Test formatting an OpenAI completion comment with system fingerprint.""" metadata = CompletionMetadata( model_name="gpt-4o", status="✅ Judgement generated successfully", generation_time_seconds=35.8, input_tokens=8000, output_tokens=1500, total_tokens=9500, estimated_cost=0.0975, model_version_used="gpt-4o-2024-01-01", system_fingerprint="fp_abc123def456", search_results_used=None, finish_reason="stop", safety_ratings=None, context_size_chars=45000, warnings=None, timestamp=datetime(2025, 1, 6, 12, 0, 35, tzinfo=timezone.utc), ) result = format_completion_comment(metadata) assert "**System Fingerprint**: `fp_abc123def456`" in result assert "**Search Results Used**:" not in result # Should not show if None assert "### Safety Ratings" not in result # Should not show if None def test_model_name_fallback(self): """Test that model_name is used when model_version_used is None.""" metadata = CompletionMetadata( model_name="gemini-1.5-pro-latest", status="✅ Workplan generated successfully", generation_time_seconds=30.0, input_tokens=5000, output_tokens=1000, total_tokens=6000, estimated_cost=0.06, model_version_used=None, # No model version available system_fingerprint=None, search_results_used=None, finish_reason="stop", safety_ratings=None, context_size_chars=25000, warnings=None, timestamp=datetime(2025, 1, 6, 12, 0, 30, tzinfo=timezone.utc), ) result = format_completion_comment(metadata) # Should show model_name as fallback assert "**Model Used**: `gemini-1.5-pro-latest`" in result # Should not show "gemini-2.5-pro" or similar assert "gemini-2.5-pro" not in result class TestExtractUrls: """Test cases for extract_urls function.""" def test_extract_single_url(self): """Test extracting a single URL from text.""" text = "Check out the documentation at https://example.com/docs for more info." urls = extract_urls(text) assert urls == ["https://example.com/docs"] def test_extract_multiple_urls(self): """Test extracting multiple URLs from text.""" text = """ Here are some useful resources: - API docs: https://api.example.com/v2/docs - GitHub repo: https://github.com/user/project - Website: http://example.org """ urls = extract_urls(text) assert len(urls) == 3 assert "https://api.example.com/v2/docs" in urls assert "https://github.com/user/project" in urls assert "http://example.org" in urls def test_extract_duplicate_urls(self): """Test that duplicate URLs are deduplicated while preserving order.""" text = """ Visit https://example.com for info. Also check https://github.com/repo and https://example.com again. """ urls = extract_urls(text) assert urls == ["https://example.com", "https://github.com/repo"] def test_no_urls_in_text(self): """Test handling text with no URLs.""" text = "This is just plain text without any URLs." urls = extract_urls(text) assert urls == [] def test_urls_with_special_characters(self): """Test extracting URLs with query parameters and fragments.""" text = """ Search at https://example.com/search?q=python&limit=10 And bookmark https://docs.example.com/guide#section-2 """ urls = extract_urls(text) assert len(urls) == 2 assert "https://example.com/search?q=python&limit=10" in urls assert "https://docs.example.com/guide#section-2" in urls def test_urls_in_parentheses(self): """Test that URLs in parentheses are extracted correctly.""" text = "For more details (see https://example.com/details) contact support." urls = extract_urls(text) assert urls == ["https://example.com/details"]