karate-graph 0.1.1

MCP tool for analyzing Karate Framework projects and generating dependency graphs

Karate Graph (MCP-Powered)

A Model Context Protocol (MCP) tool for AI agents to analyze Karate Framework projects, build dependency graphs, and generate interactive reports.

Why use this with AI?

• Impact analysis with dependency paths and source lines
• Multi-project graph exploration
• Search APIs, workflows, pages, Java/JS usages
• Failure hotspot and flaky-risk prioritization
• Reusable helper discovery before writing new code

Quick Start

1. Install

pip install karate-graph

Install from TestPyPI (pre-release validation):

pip install --index-url https://test.pypi.org/simple/ --extra-index-url https://pypi.org/simple/ karate-graph==0.1.0

Verify installation:

karate-graph-mcp --help
python -c "import karate_graph_analyzer; print('ok')"

For local development:

git clone <repository-url>
cd karate-graph
pip install -r requirements-dev.txt

2. Configure MCP client

{
  "mcpServers": {
    "karate-graph": {
      "command": "karate-graph-mcp",
      "args": [],
      "env": {
        "PYTHONPATH": "C:/path/to/repo/src",
        "PYTHONIOENCODING": "utf-8"
      }
    }
  }
}

3. Ask your AI to:

• Register and analyze project
• Show impact if a component changes
• Find reusable helpers before creating new ones

Available AI Tools

• register_project
• analyze_project
• bulk_analyze
• impact_analysis
• search_api
• search_workflow
• search_test_case
• search_java_usage
• search_js_usage
• search_error_pattern
• search_reusable_function
• change_impact_preview
• test_selection_suggestion
• feature_intent_index
• variable_data_flow_trace
• assertion_map
• call_read_deep_context
• ai_feature_context_pack
• feature_behavior_map
• scenario_similarity_map
• feature_reuse_advisor
• db_query_index
• search_db_usage
• db_data_flow_trace
• db_assertion_map
• db_impact_preview
• top_hotspots
• prioritize_fix_queue
• flaky_risk

Reuse And Smart Retest

1) Reuse search

search_reusable_function(project_name, query, language, limit) now returns:

• tags
• aliases
• usage_examples
• stability_score

This helps AI pick existing helpers with better confidence.

2) Change impact preview

change_impact_preview(project_name, changed_paths, limit) maps changed files/components to impacted test cases via dependency paths.

3) Test selection suggestion

test_selection_suggestion(project_name, changed_paths, limit) suggests a compact high-signal rerun set.

Current strategy:

priority = trigger_count * 10 - min_depth

Feature Understanding For AI

These tools help AI understand Karate .feature files before editing or debugging:

• feature_intent_index(project_name, query, limit)
  • summarizes scenario intent, step roles, API signals, data files, assertions, and call/read usage
• variable_data_flow_trace(project_name, feature_path, scenario_tag, scenario_name, node_id, limit)
  • traces def/set variables from source expressions to usage lines
• assertion_map(project_name, query, limit)
  • indexes status, match, and assert checks across feature files
• call_read_deep_context(project_name, feature_path, scenario_tag, scenario_name, node_id, max_depth, limit)
  • expands nested call read(...) chains, including target feature/scenario context
• ai_feature_context_pack(project_name, feature_path, scenario_tag, scenario_name, node_id, max_call_depth, limit)
  • returns an AI-ready pack with intent, variable flow, assertions, call/read chain, and graph context
• feature_behavior_map(project_name, feature_path, scenario_tag, scenario_name, node_id, limit)
  • groups scenario behavior into preconditions, actions, expectations, plus data inputs and status expectations
• scenario_similarity_map(project_name, query, limit, top_k)
  • finds similar scenarios by intent-keyword overlap to improve reuse and AI suggestion quality
• feature_reuse_advisor(project_name, min_group_size, min_flow_length, limit, include_low_signal)
  • finds duplicate steps and repeated flows, indexes their locations, and returns AI-safe refactor plans

DB Understanding For AI

These tools help AI understand query usage, variable flow, and DB-related impact:

• db_query_index(project_name, query, limit, include_components, link_status)
  • indexes DB query nodes and DB executor/components with operation/table/database/host/dialect/provider/risk/usage/link status
• search_db_usage(project_name, query, limit, link_status)
  • searches DB usage by table, operation, host, file path, or query keywords
• db_data_flow_trace(project_name, feature_path, scenario_tag, scenario_name, node_id, limit)
  • traces DB-related variables, DB call steps, and DB assertions inside selected scenarios
• db_assertion_map(project_name, query, limit)
  • indexes DB-related assertion steps and links them to DB variables/query signatures
• db_impact_preview(project_name, changed_entities, limit)
  • previews impacted test cases from changed DB entities (tables/schemas/hosts/DB feature paths)

Dialect and DB-type detection is included in DB outputs:

• SQL dialects: PostgreSQL, MySQL, MariaDB, Oracle, SQL Server, SQLite, DB2, H2, Redshift, Snowflake, ClickHouse, and generic SQL
• NoSQL/cache/search/graph stores: MongoDB, Redis, DynamoDB, Cassandra/CQL, Elasticsearch/OpenSearch, Neo4j/Cypher
• Returned fields include db_type, dialect, provider, dialect_confidence, dialect_signals, entity_type, and entity_name
• Returned DB index rows also include link_status: linked, orphan, component, or demo
• Use link_status="default" to focus on impact-first rows (linked + orphan) while keeping component/demo context searchable when needed
• Detection uses Strategy + Registry + Value Object so new DB providers can be added as isolated strategies.

Visual Reports

Interactive HTML graph output is generated under each project output/ folder.

• Test cases are shown as @TEST-ID - Scenario name when Jira/test-case tags exist.
• Dashboard search supports TEST-ID, @TEST-ID, scenario name, component name, and feature path.

Publishing

Build and validate locally first:

python -m pip install build twine
python -m build
python -m twine check dist/*

Publish to TestPyPI before PyPI:

python -m twine upload --repository testpypi dist/*
python -m twine upload dist/*

Use an API token for upload. The package ships report assets (style.css, script.js) via package data.

Project Structure

karate-graph/
|-- src/karate_graph_analyzer/
|   |-- mcp_server.py
|   |-- mcp_interface/
|   |-- parser/
|   |-- graph/
|   `-- visualization/
|-- output/
`-- tests/

License

MIT