Add vrom_hub/cli.py

vrom_hub/cli.py

@@ -0,0 +1,232 @@
+#!/usr/bin/env python3
+"""
+vROM Hub CLI — build and publish vROM packages from the command line.
+
+Usage:
+    # Build from markdown files
+    python -m vrom_hub.cli build my-docs \
+        --name "My Docs" \
+        --description "My project documentation" \
+        --files docs/*.md \
+        --output ./vrom_output
+
+    # Build from URLs
+    python -m vrom_hub.cli build my-docs \
+        --name "My Docs" \
+        --urls https://example.com/docs/page1 https://example.com/docs/page2
+
+    # Build + upload to registry
+    python -m vrom_hub.cli submit my-docs \
+        --name "My Docs" \
+        --description "My project docs" \
+        --files docs/*.md \
+        --tags my-project api
+
+    # List registry
+    python -m vrom_hub.cli list
+
+    # Info about a specific vROM
+    python -m vrom_hub.cli info my-docs
+"""
+
+import argparse
+import glob
+import json
+import logging
+import sys
+from pathlib import Path
+
+
+def cmd_build(args):
+    """Build a vROM locally (no upload)."""
+    from vrom_hub import VromHubBackend
+
+    hub = VromHubBackend(
+        registry_repo=args.registry,
+        max_chunk_tokens=args.max_tokens,
+    )
+
+    pages = _collect_pages(args)
+    if not pages:
+        print("Error: No pages found. Provide --files or --urls.", file=sys.stderr)
+        sys.exit(1)
+
+    result = hub.build_vrom(
+        pages=pages,
+        vrom_id=args.vrom_id,
+        version=args.version or "1.0.0",
+        description=args.description or "",
+        output_dir=args.output or f"./vrom-{args.vrom_id}",
+    )
+
+    print(json.dumps({
+        "vrom_id": result["vrom_id"],
+        "output_dir": result["output_dir"],
+        "stats": result["stats"],
+    }, indent=2))
+
+
+def cmd_submit(args):
+    """Build and upload a vROM to the registry."""
+    from vrom_hub import VromHubBackend
+
+    hub = VromHubBackend(
+        registry_repo=args.registry,
+        max_chunk_tokens=args.max_tokens,
+    )
+
+    pages = _collect_pages(args)
+    if not pages:
+        print("Error: No pages found. Provide --files or --urls.", file=sys.stderr)
+        sys.exit(1)
+
+    result = hub.submit_project(
+        vrom_id=args.vrom_id,
+        name=args.name or args.vrom_id,
+        description=args.description or "",
+        version=args.version or "1.0.0",
+        pages=pages,
+        tags=args.tags,
+        official=args.official,
+        output_dir=args.output,
+        upload=True,
+    )
+
+    print(json.dumps({
+        "vrom_id": result["vrom_id"],
+        "hub_url": result.get("hub_url"),
+        "stats": result["stats"],
+    }, indent=2))
+
+
+def cmd_list(args):
+    """List all vROMs in the registry."""
+    from vrom_hub import VromHubBackend
+
+    hub = VromHubBackend(registry_repo=args.registry)
+    vroms = hub.list_vroms()
+
+    if not vroms:
+        print("No vROMs found in registry.")
+        return
+
+    for v in vroms:
+        tags = ", ".join(v.get("tags", []))
+        print(f"  {v['id']:30s}  v{v['version']:8s}  {v['vectors']:5d} vectors  {v.get('size_mb', '?'):>5}MB  [{tags}]")
+        print(f"    {v['description']}")
+        print()
+
+
+def cmd_info(args):
+    """Show details for a specific vROM."""
+    from vrom_hub import VromHubBackend
+
+    hub = VromHubBackend(registry_repo=args.registry)
+    vroms = hub.list_vroms()
+
+    for v in vroms:
+        if v["id"] == args.vrom_id:
+            print(json.dumps(v, indent=2))
+            return
+
+    print(f"vROM '{args.vrom_id}' not found in registry.", file=sys.stderr)
+    sys.exit(1)
+
+
+def _collect_pages(args) -> list[dict]:
+    """Collect documentation pages from CLI arguments."""
+    pages = []
+
+    # From files
+    if hasattr(args, 'files') and args.files:
+        for pattern in args.files:
+            for filepath in sorted(glob.glob(pattern)):
+                path = Path(filepath)
+                content = path.read_text(encoding="utf-8")
+                pages.append({
+                    "content": content,
+                    "source_file": str(path),
+                    "url": "",
+                    "title": None,
+                })
+
+    # From URLs (not fetched here — the backend's fetcher handles it)
+    if hasattr(args, 'urls') and args.urls:
+        from vrom_hub.fetcher import DocFetcher
+        fetcher = DocFetcher()
+        for url in args.urls:
+            try:
+                doc_page = fetcher.fetch_url(url)
+                pages.append({
+                    "content": doc_page.content,
+                    "source_file": doc_page.source_file,
+                    "url": doc_page.url,
+                    "title": doc_page.title,
+                })
+            except Exception as e:
+                print(f"Warning: Failed to fetch {url}: {e}", file=sys.stderr)
+
+    return pages
+
+
+def main():
+    parser = argparse.ArgumentParser(
+        prog="vrom-hub",
+        description="vROM Hub Backend — build and publish vROM packages",
+    )
+    parser.add_argument(
+        "--registry", default="philipp-zettl/vrom-registry",
+        help="HF dataset repo for the vROM registry",
+    )
+    parser.add_argument(
+        "--verbose", "-v", action="store_true",
+        help="Enable verbose logging",
+    )
+
+    sub = parser.add_subparsers(dest="command", required=True)
+
+    # ── build ─────────────────────────────────────────────
+    p_build = sub.add_parser("build", help="Build a vROM locally")
+    p_build.add_argument("vrom_id", help="Unique vROM identifier")
+    p_build.add_argument("--name", help="Human-readable name")
+    p_build.add_argument("--description", "-d", help="Description")
+    p_build.add_argument("--version", default="1.0.0", help="Version (default: 1.0.0)")
+    p_build.add_argument("--files", nargs="+", help="Markdown file glob patterns")
+    p_build.add_argument("--urls", nargs="+", help="URLs to fetch")
+    p_build.add_argument("--output", "-o", help="Output directory")
+    p_build.add_argument("--max-tokens", type=int, default=256, help="Max tokens per chunk")
+    p_build.set_defaults(func=cmd_build)
+
+    # ── submit ────────────────────────────────────────────
+    p_submit = sub.add_parser("submit", help="Build + upload to registry")
+    p_submit.add_argument("vrom_id", help="Unique vROM identifier")
+    p_submit.add_argument("--name", help="Human-readable name")
+    p_submit.add_argument("--description", "-d", help="Description")
+    p_submit.add_argument("--version", default="1.0.0", help="Version (default: 1.0.0)")
+    p_submit.add_argument("--files", nargs="+", help="Markdown file glob patterns")
+    p_submit.add_argument("--urls", nargs="+", help="URLs to fetch")
+    p_submit.add_argument("--output", "-o", help="Output directory (optional)")
+    p_submit.add_argument("--tags", nargs="+", default=[], help="Tags for the vROM")
+    p_submit.add_argument("--official", action="store_true", help="Mark as official")
+    p_submit.add_argument("--max-tokens", type=int, default=256, help="Max tokens per chunk")
+    p_submit.set_defaults(func=cmd_submit)
+
+    # ── list ──────────────────────────────────────────────
+    p_list = sub.add_parser("list", help="List all vROMs in registry")
+    p_list.set_defaults(func=cmd_list)
+
+    # ── info ──────────────────────────────────────────────
+    p_info = sub.add_parser("info", help="Show details for a vROM")
+    p_info.add_argument("vrom_id", help="vROM identifier")
+    p_info.set_defaults(func=cmd_info)
+
+    args = parser.parse_args()
+
+    level = logging.DEBUG if args.verbose else logging.INFO
+    logging.basicConfig(level=level, format="%(name)s | %(message)s")
+
+    args.func(args)
+
+
+if __name__ == "__main__":
+    main()