personas/build.py

#!/usr/bin/env python3
"""Build script: Generate .yaml, .json, .prompt.md from persona .md files.

Supports config.yaml for dynamic variable injection and user-specific customization.
New users: copy config.example.yaml → config.yaml and customize.
"""

import json
import re
import sys
from pathlib import Path

try:
    import yaml
except ImportError:
    print("PyYAML required: pip install pyyaml")
    sys.exit(1)


def load_config(root: Path) -> dict:
    """Load config.yaml if it exists, otherwise return empty config."""
    config_path = root / "config.yaml"
    if config_path.exists():
        config = yaml.safe_load(config_path.read_text(encoding="utf-8")) or {}
        print(f"Config loaded: {config_path}")
        return config

    example_path = root / "config.example.yaml"
    if example_path.exists():
        print("WARN: No config.yaml found. Using defaults. Copy config.example.yaml → config.yaml to customize.")
    return {}


def flatten_config(config: dict, prefix: str = "") -> dict:
    """Flatten nested config dict for template substitution.

    Example: {"user": {"name": "Salva"}} → {"user.name": "Salva"}
    """
    flat = {}
    for key, value in config.items():
        full_key = f"{prefix}{key}" if not prefix else f"{prefix}.{key}"
        if isinstance(value, dict):
            flat.update(flatten_config(value, full_key))
        elif isinstance(value, list):
            flat[full_key] = value
            flat[f"{full_key}.count"] = len(value)
            flat[f"{full_key}.csv"] = ", ".join(str(v) for v in value if not isinstance(v, dict))
        else:
            flat[full_key] = value
    return flat


def inject_config(content: str, flat_config: dict) -> str:
    """Replace {{config.key}} placeholders with config values."""
    def replacer(match):
        key = match.group(1).strip()
        value = flat_config.get(key, match.group(0))  # keep original if not found
        if isinstance(value, list):
            return ", ".join(str(v) for v in value if not isinstance(v, dict))
        if isinstance(value, bool):
            return "enabled" if value else "disabled"
        return str(value)

    return re.sub(r"\{\{(.+?)\}\}", replacer, content)


def check_conditionals(content: str, flat_config: dict) -> str:
    """Process {{#if key}}...{{/if}} and {{#unless key}}...{{/unless}} blocks."""
    # Handle {{#if key}}content{{/if}}
    def if_replacer(match):
        key = match.group(1).strip()
        body = match.group(2)
        value = flat_config.get(key)
        if value and value not in (False, 0, "", "false", "none", "disabled", None, []):
            return body
        return ""

    content = re.sub(r"\{\{#if (.+?)\}\}(.*?)\{\{/if\}\}", if_replacer, content, flags=re.DOTALL)

    # Handle {{#unless key}}content{{/unless}}
    def unless_replacer(match):
        key = match.group(1).strip()
        body = match.group(2)
        value = flat_config.get(key)
        if not value or value in (False, 0, "", "false", "none", "disabled", None, []):
            return body
        return ""

    content = re.sub(r"\{\{#unless (.+?)\}\}(.*?)\{\{/unless\}\}", unless_replacer, content, flags=re.DOTALL)

    return content


def parse_persona_md(filepath: Path, flat_config: dict) -> dict:
    """Parse a persona markdown file into structured data."""
    content = filepath.read_text(encoding="utf-8")

    # Apply config injection
    if flat_config:
        content = check_conditionals(content, flat_config)
        content = inject_config(content, flat_config)

    # Extract YAML frontmatter
    fm_match = re.match(r"^---\n(.*?)\n---\n(.*)$", content, re.DOTALL)
    if not fm_match:
        print(f"  WARN: No frontmatter in {filepath}")
        return {}

    frontmatter = yaml.safe_load(fm_match.group(1))
    body = fm_match.group(2).strip()

    # Extract sections from body
    sections = {}
    current_section = None
    current_content = []

    for line in body.split("\n"):
        if line.startswith("## "):
            if current_section:
                sections[current_section] = "\n".join(current_content).strip()
            current_section = line[3:].strip().lower().replace(" ", "_").replace("&", "and")
            current_content = []
        else:
            current_content.append(line)

    if current_section:
        sections[current_section] = "\n".join(current_content).strip()

    return {
        "metadata": frontmatter,
        "sections": sections,
        "raw_body": body,
    }


def build_persona(persona_dir: Path, output_dir: Path, flat_config: dict, config: dict):
    """Build all variants for a persona directory."""
    md_files = sorted(persona_dir.glob("*.md"))
    if not md_files:
        return 0

    persona_name = persona_dir.name
    out_path = output_dir / persona_name
    out_path.mkdir(parents=True, exist_ok=True)

    # Load _meta.yaml if exists
    meta_file = persona_dir / "_meta.yaml"
    meta = {}
    if meta_file.exists():
        meta_content = meta_file.read_text(encoding="utf-8")
        if flat_config:
            meta_content = inject_config(meta_content, flat_config)
        meta = yaml.safe_load(meta_content) or {}

    # Apply config overrides for address
    addresses = config.get("persona_defaults", {}).get("custom_addresses", {})
    if persona_name in addresses:
        meta["address_to"] = addresses[persona_name]

    count = 0
    for md_file in md_files:
        if md_file.name.startswith("_"):
            continue

        variant = md_file.stem
        parsed = parse_persona_md(md_file, flat_config)
        if not parsed:
            continue

        # Build output object
        output = {**meta, **parsed["metadata"], "variant": variant, "sections": parsed["sections"]}

        # Inject config metadata
        if config:
            output["_config"] = {
                "user": config.get("user", {}).get("name", "unknown"),
                "tools": {k: v for k, v in config.get("infrastructure", {}).get("tools", {}).items() if v is True},
                "frameworks": {k: v for k, v in config.get("frameworks", {}).items() if v is True},
                "regional_focus": config.get("regional_focus", {}),
            }

        # Write YAML
        yaml_out = out_path / f"{variant}.yaml"
        yaml_out.write_text(
            yaml.dump(output, allow_unicode=True, default_flow_style=False, sort_keys=False),
            encoding="utf-8",
        )

        # Write JSON
        json_out = out_path / f"{variant}.json"
        json_out.write_text(json.dumps(output, ensure_ascii=False, indent=2), encoding="utf-8")

        # Write plain system prompt (just the body, no config metadata)
        prompt_out = out_path / f"{variant}.prompt.md"
        prompt_out.write_text(parsed["raw_body"], encoding="utf-8")

        count += 1
        print(f"  Built: {persona_name}/{variant} -> .yaml .json .prompt.md")

    return count


def build_catalog(personas_dir: Path, output_dir: Path, config: dict):
    """Generate CATALOG.md from all personas."""
    addresses = config.get("persona_defaults", {}).get("custom_addresses", {})
    catalog_lines = [
        "# Persona Catalog\n",
        f"_Auto-generated by build.py | User: {config.get('user', {}).get('name', 'default')}_\n",
    ]

    for persona_dir in sorted(personas_dir.iterdir()):
        if not persona_dir.is_dir() or persona_dir.name.startswith((".", "_")):
            continue

        meta_file = persona_dir / "_meta.yaml"
        if not meta_file.exists():
            continue

        meta = yaml.safe_load(meta_file.read_text(encoding="utf-8")) or {}
        codename = meta.get("codename", persona_dir.name)
        address = addresses.get(persona_dir.name, meta.get("address_to", "N/A"))
        variants = [f.stem for f in sorted(persona_dir.glob("*.md")) if not f.name.startswith("_")]

        catalog_lines.append(f"## {codename} — {meta.get('role', 'Unknown')}")
        catalog_lines.append(f"- **Domain:** {meta.get('domain', 'N/A')}")
        catalog_lines.append(f"- **Hitap:** {address}")
        catalog_lines.append(f"- **Variants:** {', '.join(variants)}")
        catalog_lines.append("")

    catalog_path = personas_dir / "CATALOG.md"
    catalog_path.write_text("\n".join(catalog_lines), encoding="utf-8")
    print(f"  Catalog: {catalog_path}")


def print_summary(config: dict, total_personas: int, total_variants: int):
    """Print build summary with config status."""
    print("\n" + "=" * 50)
    print(f"BUILD COMPLETE")
    print(f"  Personas: {total_personas}")
    print(f"  Variants: {total_variants}")
    print(f"  Output:   generated/")

    if config:
        user = config.get("user", {}).get("name", "?")
        tools_on = sum(1 for v in config.get("infrastructure", {}).get("tools", {}).values() if v is True)
        frameworks_on = sum(1 for v in config.get("frameworks", {}).values() if v is True)
        regions = config.get("regional_focus", {}).get("primary", [])
        print(f"\n  Config:   {user}")
        print(f"  Tools:    {tools_on} enabled")
        print(f"  Frameworks: {frameworks_on} enabled")
        if regions:
            print(f"  Regions:  {', '.join(regions)}")
    else:
        print("\n  Config:   none (using defaults)")
        print("  Tip:      Copy config.example.yaml → config.yaml to customize")
    print("=" * 50)


def main():
    root = Path(__file__).parent
    personas_dir = root / "personas"

    if not personas_dir.exists():
        print("No personas/ directory found.")
        sys.exit(1)

    output_dir = root / "generated"

    # Load config
    config = load_config(root)
    flat_config = flatten_config(config) if config else {}

    # Find all persona directories
    persona_dirs = [
        d for d in sorted(personas_dir.iterdir()) if d.is_dir() and not d.name.startswith((".", "_"))
    ]

    if not persona_dirs:
        print("No persona directories found.")
        sys.exit(1)

    output_dir.mkdir(parents=True, exist_ok=True)
    print(f"Building {len(persona_dirs)} personas -> {output_dir}\n")

    total_variants = 0
    for pdir in persona_dirs:
        total_variants += build_persona(pdir, output_dir, flat_config, config)

    build_catalog(personas_dir, output_dir, config)
    print_summary(config, len(persona_dirs), total_variants)


if __name__ == "__main__":
    main()