graph-dependency-analyzer 0.2.0__py3-none-any.whl

src/cli.py

@@ -0,0 +1,512 @@
+"""
+graph-explorer CLI — Knowledge Graph Dependency Analyzer
+
+Usage:
+    graph-explorer init          Setup project (guides through .env, Docker, Ollama)
+    graph-explorer start         Start the API server
+    graph-explorer index <repos> Index repositories into the graph
+    graph-explorer link          Connect cross-repo dependencies
+    graph-explorer status        Show indexing status
+    graph-explorer open          Open the graph in browser
+"""
+from __future__ import annotations
+
+import os
+import shutil
+import subprocess
+import sys
+import time
+from pathlib import Path
+
+import click
+import httpx
+
+# ── Constants ──────────────────────────────────────────────────────────────
+
+API_URL = "http://localhost:8000"
+PACKAGE_ROOT = Path(__file__).parent.parent  # project root when installed
+
+_BANNER = """
+╔══════════════════════════════════════════════════════════╗
+║        Knowledge Graph — Dependency Analyzer             ║
+║        graph-explorer CLI                                ║
+╚══════════════════════════════════════════════════════════╝
+"""
+
+
+# ── Helpers ────────────────────────────────────────────────────────────────
+
+def _api_running() -> bool:
+    try:
+        r = httpx.get(f"{API_URL}/health", timeout=3)
+        return r.status_code == 200
+    except Exception:
+        return False
+
+
+def _docker_running() -> bool:
+    try:
+        result = subprocess.run(
+            ["docker", "info"], capture_output=True, timeout=5
+        )
+        return result.returncode == 0
+    except Exception:
+        return False
+
+
+def _ollama_running() -> bool:
+    try:
+        r = httpx.get("http://localhost:11434/api/tags", timeout=3)
+        return r.status_code == 200
+    except Exception:
+        return False
+
+
+def _ollama_model_installed(model: str) -> bool:
+    try:
+        r = httpx.get("http://localhost:11434/api/tags", timeout=3)
+        if r.status_code == 200:
+            models = [m["name"] for m in r.json().get("models", [])]
+            return any(model in m for m in models)
+    except Exception:
+        pass
+    return False
+
+
+def _find_project_root() -> Path:
+    """Find the project root — works both in dev and when pip-installed."""
+    # When pip installed, look for docker-compose.yml in CWD or parent
+    cwd = Path.cwd()
+    for p in [cwd, *cwd.parents]:
+        if (p / "docker-compose.yml").exists() and (p / ".env").exists():
+            return p
+    # Fallback: package root
+    return PACKAGE_ROOT
+
+
+def _check_env_file(project_root: Path) -> bool:
+    return (project_root / ".env").exists()
+
+
+# ── CLI Group ──────────────────────────────────────────────────────────────
+
+@click.group()
+def cli():
+    """Knowledge Graph Dependency Analyzer — explore cross-repo dependencies."""
+    pass
+
+
+# ── init ──────────────────────────────────────────────────────────────────
+
+@cli.command()
+@click.option("--skip-docker", is_flag=True, help="Skip Docker setup check")
+@click.option("--skip-ollama", is_flag=True, help="Skip Ollama setup check")
+def init(skip_docker: bool, skip_ollama: bool):
+    """
+    First-time setup wizard.
+
+    Guides through:\n
+      1. Creating .env from template\n
+      2. Starting Docker infrastructure (Neo4j + Qdrant)\n
+      3. Setting up Ollama (optional, for local embeddings)\n
+      4. Installing Claude Code skills
+    """
+    click.echo(_BANNER)
+    click.echo("Welcome! Let's get you set up.\n")
+
+    project_root = _find_project_root()
+    env_file = project_root / ".env"
+    env_example = project_root / ".env.example"
+
+    # ── Step 1: .env ───────────────────────────────────────────────────
+    click.echo("📋 Step 1: Environment configuration")
+
+    if env_file.exists():
+        click.echo("  ✅ .env already exists")
+    else:
+        if env_example.exists():
+            shutil.copy(env_example, env_file)
+            click.echo("  ✅ Created .env from .env.example")
+        else:
+            click.echo("  ⚠️  .env.example not found — creating minimal .env")
+            env_file.write_text(_MINIMAL_ENV)
+
+        click.echo("\n  ⚠️  Please edit .env before continuing:")
+        click.echo(f"     {env_file}")
+        click.echo("\n  Required settings:")
+        click.echo("     REPOSITORIES_PATH=/path/to/your/repos")
+        click.echo("     EMBEDDING_PROVIDER=ollama  (or 'flow' if using CI&T Flow)")
+        click.echo()
+        if not click.confirm("  Have you configured .env?", default=False):
+            click.echo("\n  Please edit .env and run 'graph-explorer init' again.")
+            return
+
+    # ── Step 2: Docker ─────────────────────────────────────────────────
+    click.echo("\n🐳 Step 2: Docker infrastructure (Neo4j + Qdrant)")
+
+    if skip_docker:
+        click.echo("  ⏭️  Skipped (--skip-docker)")
+    elif not _docker_running():
+        click.echo("  ❌ Docker is not running.")
+        click.echo("  Please start Docker Desktop and run 'graph-explorer init' again.")
+        click.echo("  Download: https://docs.docker.com/get-docker/")
+        return
+    else:
+        compose_file = project_root / "docker-compose.yml"
+        if not compose_file.exists():
+            click.echo("  ❌ docker-compose.yml not found")
+            return
+
+        click.echo("  🔄 Starting containers...")
+        result = subprocess.run(
+            ["docker", "compose", "up", "-d"],
+            cwd=str(project_root),
+            capture_output=True, text=True
+        )
+        if result.returncode != 0:
+            click.echo(f"  ❌ Docker failed:\n{result.stderr}")
+            return
+
+        click.echo("  ⏳ Waiting for Neo4j and Qdrant to be ready (~15s)...")
+        time.sleep(15)
+        click.echo("  ✅ Containers running")
+        click.echo("     Neo4j browser: http://localhost:7474")
+        click.echo("     Qdrant:        http://localhost:6333")
+
+    # ── Step 3: Ollama ─────────────────────────────────────────────────
+    click.echo("\n🦙 Step 3: Ollama (local AI embeddings — no rate limits)")
+
+    if skip_ollama:
+        click.echo("  ⏭️  Skipped (--skip-ollama)")
+    else:
+        # Check if .env uses ollama
+        env_content = env_file.read_text()
+        uses_ollama = "EMBEDDING_PROVIDER=ollama" in env_content
+
+        if not uses_ollama:
+            click.echo("  ℹ️  EMBEDDING_PROVIDER is not 'ollama' in .env — skipping")
+            click.echo("     (Using Flow/OpenAI for embeddings)")
+        elif not shutil.which("ollama"):
+            click.echo("  ❌ Ollama is not installed.")
+            click.echo("\n  Install options:")
+            click.echo("     Mac:   brew install ollama")
+            click.echo("     Other: https://ollama.com/download")
+            click.echo("\n  After installing:")
+            click.echo("     ollama serve &")
+            click.echo("     ollama pull nomic-embed-text")
+            click.echo("\n  Then run 'graph-explorer init' again.")
+            return
+        elif not _ollama_running():
+            click.echo("  ⚠️  Ollama is installed but not running.")
+            click.echo("  Start it with: ollama serve")
+            if click.confirm("  Start Ollama now?", default=True):
+                subprocess.Popen(["ollama", "serve"],
+                                 stdout=subprocess.DEVNULL,
+                                 stderr=subprocess.DEVNULL)
+                time.sleep(3)
+                click.echo("  ✅ Ollama started")
+            else:
+                click.echo("  ⚠️  Ollama not running — embeddings will fail")
+        else:
+            click.echo("  ✅ Ollama is running")
+
+        if _ollama_running():
+            model = "nomic-embed-text"
+            if _ollama_model_installed(model):
+                click.echo(f"  ✅ Model '{model}' already installed")
+            else:
+                click.echo(f"  📥 Pulling model '{model}' (~274MB)...")
+                subprocess.run(["ollama", "pull", model])
+                click.echo(f"  ✅ Model '{model}' ready")
+
+    # ── Step 4: Claude skills ──────────────────────────────────────────
+    click.echo("\n🤖 Step 4: Claude Code skills")
+    install_script = project_root / "install-graph-skills.sh"
+    if install_script.exists():
+        skills_dir = Path.home() / ".claude" / "skills" / "graph-index"
+        if skills_dir.exists():
+            click.echo("  ✅ Skills already installed globally")
+        else:
+            if click.confirm("  Install /graph-index, /graph-link, /graph-impact skills globally?", default=True):
+                result = subprocess.run(
+                    ["bash", str(install_script)],
+                    capture_output=True, text=True
+                )
+                if result.returncode == 0:
+                    click.echo("  ✅ Skills installed — use /graph-index in Claude Code")
+                else:
+                    click.echo(f"  ⚠️  {result.stderr[:100]}")
+    else:
+        click.echo("  ⏭️  install-graph-skills.sh not found — skipping")
+
+    # ── Done ───────────────────────────────────────────────────────────
+    click.echo("\n" + "─" * 58)
+    click.echo("✅ Setup complete! Next steps:")
+    click.echo()
+    click.echo("  1. Start the API:     graph-explorer start")
+    click.echo("  2. Index your repos:  graph-explorer index <repo1> <repo2>")
+    click.echo("  3. Connect links:     graph-explorer link")
+    click.echo("  4. Open the graph:    graph-explorer open")
+    click.echo()
+
+
+# ── start ──────────────────────────────────────────────────────────────────
+
+@cli.command()
+@click.option("--port", default=8000, show_default=True, help="API port")
+@click.option("--host", default="0.0.0.0", show_default=True, help="API host")
+@click.option("--no-reload", is_flag=True, help="Disable auto-reload")
+def start(port: int, host: str, no_reload: bool):
+    """Start the Knowledge Graph API server."""
+    click.echo(_BANNER)
+
+    if _api_running():
+        click.echo(f"✅ API already running at http://localhost:{port}")
+        click.echo(f"   Docs: http://localhost:{port}/docs")
+        return
+
+    # Verify Docker infra
+    if not _api_running():
+        click.echo("🔄 Checking infrastructure...")
+
+    reload_flag = [] if no_reload else ["--reload"]
+    cmd = [
+        sys.executable, "-m", "uvicorn",
+        "src.presentation.api.rest.main:app",
+        "--host", host,
+        "--port", str(port),
+        *reload_flag,
+    ]
+
+    click.echo(f"🚀 Starting API on http://{host}:{port}")
+    click.echo(f"   Docs:  http://localhost:{port}/docs")
+    click.echo(f"   Graph: http://localhost:{port}/api/analysis/repository-relationships/graph.html")
+    click.echo("\n   Press Ctrl+C to stop\n")
+
+    project_root = _find_project_root()
+    os.chdir(str(project_root))
+    os.execvp(sys.executable, cmd)
+
+
+# ── index ──────────────────────────────────────────────────────────────────
+
+@cli.command()
+@click.argument("repos", nargs=-1, required=True)
+@click.option("--parallel", "-p", default=None, type=int,
+              help="Parallel limit (auto-detected if not set)")
+@click.option("--api-url", default=API_URL, show_default=True)
+def index(repos: tuple, parallel: int | None, api_url: str):
+    """
+    Index repositories into the Knowledge Graph.
+
+    Examples:\n
+        graph-explorer index wdc-packing-list wdc-ui\n
+        graph-explorer index wms-oracle --parallel 1
+    """
+    if not _api_running():
+        click.echo("❌ API is not running. Start it first:")
+        click.echo("   graph-explorer start")
+        raise SystemExit(1)
+
+    # Auto-detect parallel limit
+    large_repos = {"wmsportal", "wms-oracle", "wms-oracle-fork"}
+    if parallel is None:
+        parallel = 1 if any(r in large_repos for r in repos) else 3
+
+    click.echo(f"📦 Indexing {len(repos)} repo(s) with parallel={parallel}...\n")
+
+    payload = {
+        "repositories": list(repos),
+        "parallel": parallel,
+        "background": False,
+    }
+
+    try:
+        with httpx.Client(timeout=3600) as client:
+            r = client.post(f"{api_url}/api/indexing/index/batch", json=payload)
+            r.raise_for_status()
+            data = r.json()
+    except httpx.ConnectError:
+        click.echo("❌ Cannot connect to API. Is it running?")
+        click.echo("   graph-explorer start")
+        raise SystemExit(1)
+    except Exception as e:
+        click.echo(f"❌ Error: {e}")
+        raise SystemExit(1)
+
+    results = data.get("results", [])
+    success = data.get("success_count", 0)
+    total = data.get("total_processed", 0)
+
+    # Table
+    click.echo(f"{'Repository':<40} {'Status':<8} {'Files':>6} {'Nodes':>7} {'Chunks':>7}")
+    click.echo("─" * 72)
+    for r in results:
+        name = (r.get("repository_name") or "")[:38]
+        ok = r.get("status") == "success"
+        status = "✅ ok" if ok else "❌ err"
+        files = r.get("files_processed", 0) or 0
+        nodes = r.get("nodes_created", 0) or 0
+        chunks = r.get("chunks_created", 0) or 0
+        err = (r.get("error_message") or "")[:40]
+        click.echo(f"{name:<40} {status:<8} {files:>6} {nodes:>7} {chunks:>7}  {err}")
+
+    click.echo("─" * 72)
+    click.echo(f"{'TOTAL':<40} {success}/{total} ok\n")
+
+    # Has SQL repos?
+    sql_repos = {"wms-oracle", "wms-oracle-fork"}
+    if any(r in sql_repos for r in repos):
+        click.echo("💡 SQL/Oracle repos indexed → run: graph-explorer link --db-only")
+    else:
+        click.echo("💡 Run 'graph-explorer link' to connect cross-repo dependencies")
+
+    click.echo(f"\n🌐 View: {api_url}/api/analysis/repository-relationships/graph.html")
+
+
+# ── link ──────────────────────────────────────────────────────────────────
+
+@cli.command()
+@click.option("--http-only", is_flag=True, help="Only link HTTP calls")
+@click.option("--db-only", is_flag=True, help="Only resolve DB refs")
+@click.option("--api-url", default=API_URL, show_default=True)
+def link(http_only: bool, db_only: bool, api_url: str):
+    """
+    Connect cross-repository dependencies.
+
+    Runs:\n
+      1. cross-language-link — matches HTTP calls to backend endpoints\n
+      2. resolve-db-refs     — connects Oracle/SQL procedures to tables
+    """
+    if not _api_running():
+        click.echo("❌ API is not running. Start it first: graph-explorer start")
+        raise SystemExit(1)
+
+    with httpx.Client(timeout=120) as client:
+
+        if not db_only:
+            click.echo("🔗 Linking HTTP calls (frontend → backend)...")
+            r = client.post(f"{api_url}/api/analysis/impact/cross-language-link")
+            d = r.json()
+            created = d.get("created", 0)
+            click.echo(f"   ✅ {created} CALLS_BACKEND edges created")
+
+        if not http_only:
+            click.echo("🗄️  Resolving DB cross-references (~30s)...")
+            r = client.post(f"{api_url}/api/analysis/impact/resolve-db-refs",
+                            timeout=120)
+            d = r.json()
+            new_edges = d.get("total_new_edges", 0)
+            total = d.get("total_resolved_edges_in_graph", 0)
+            click.echo(f"   ✅ {new_edges} new DB edges — {total} total resolved")
+
+    click.echo(f"\n🌐 View: {api_url}/api/analysis/repository-relationships/graph.html")
+
+
+# ── status ─────────────────────────────────────────────────────────────────
+
+@cli.command()
+@click.option("--api-url", default=API_URL, show_default=True)
+def status(api_url: str):
+    """Show current indexing status and graph stats."""
+    click.echo("📊 Knowledge Graph Status\n")
+
+    # API
+    if _api_running():
+        click.echo("  ✅ API running")
+        try:
+            r = httpx.get(f"{api_url}/health", timeout=3)
+            for check in r.json().get("checks", []):
+                icon = "✅" if check.get("status") == "healthy" else "❌"
+                click.echo(f"     {icon} {check['service']}: {check.get('message','')}")
+        except Exception:
+            pass
+    else:
+        click.echo("  ❌ API not running → graph-explorer start")
+
+    # Docker
+    click.echo(f"  {'✅' if _docker_running() else '❌'} Docker")
+
+    # Ollama
+    if _ollama_running():
+        click.echo("  ✅ Ollama running")
+        if _ollama_model_installed("nomic-embed-text"):
+            click.echo("     ✅ nomic-embed-text installed")
+    else:
+        click.echo("  ⬜ Ollama not running (ok if using Flow/OpenAI)")
+
+    # Graph stats
+    if _api_running():
+        try:
+            r = httpx.get(f"{api_url}/api/analysis/repository-relationships", timeout=10)
+            d = r.json()
+            repos = d.get("repositories", [])
+            rels = d.get("relationships", [])
+            meta = d.get("meta", {})
+            click.echo(f"\n  📈 Graph: {len(repos)} repos | {meta.get('total_after_dedup',len(rels))} relationships")
+            if repos:
+                click.echo(f"  Indexed repos: {', '.join(repos[:10])}")
+                if len(repos) > 10:
+                    click.echo(f"    ... and {len(repos)-10} more")
+        except Exception:
+            pass
+
+    click.echo()
+
+
+# ── open ───────────────────────────────────────────────────────────────────
+
+@cli.command("open")
+@click.option("--repo", "-r", default=None, help="Open specific repo graph")
+@click.option("--api-url", default=API_URL, show_default=True)
+def open_graph(repo: str | None, api_url: str):
+    """Open the dependency graph in your browser."""
+    import webbrowser
+
+    if repo:
+        url = f"{api_url}/api/export/graph.html?repository={repo}"
+    else:
+        url = f"{api_url}/api/analysis/repository-relationships/graph.html"
+
+    click.echo(f"🌐 Opening: {url}")
+    webbrowser.open(url)
+
+
+# ── Minimal .env template ──────────────────────────────────────────────────
+
+_MINIMAL_ENV = """\
+# ── Neo4j ────────────────────────────────────────────────
+NEO4J_URI=bolt://localhost:7687
+NEO4J_USER=neo4j
+NEO4J_PASSWORD=medline2024
+
+# ── Qdrant ───────────────────────────────────────────────
+QDRANT_URL=http://localhost:6333
+QDRANT_COLLECTION=code_chunks
+QDRANT_VECTOR_SIZE=768
+
+# ── Repositories ─────────────────────────────────────────
+REPOSITORIES_PATH=/path/to/your/repositories
+
+# ── Embedding (ollama | flow | openai) ───────────────────
+EMBEDDING_PROVIDER=ollama
+OLLAMA_BASE_URL=http://localhost:11434
+OLLAMA_EMBEDDING_MODEL=nomic-embed-text
+OLLAMA_EMBEDDING_DIM=768
+
+# Flow (CI&T) — set EMBEDDING_PROVIDER=flow to use
+# OPENAI_API_KEY=your-flow-token
+# OPENAI_BASE_URL=https://flow.ciandt.com/ai-orchestration-api
+# OPENAI_EMBEDDING_MODEL=text-embedding-3-small
+# QDRANT_VECTOR_SIZE=1536
+"""
+
+
+# ── Entry point ────────────────────────────────────────────────────────────
+
+def main():
+    cli()
+
+
+if __name__ == "__main__":
+    main()

src/config/__init__.py

@@ -0,0 +1,5 @@
+"""Configuration package."""
+from src.config.settings import Settings, get_settings
+from src.config import logging_config
+
+__all__ = ["Settings", "get_settings", "logging_config"]