prompt_template/scripts/validate.py

#!/usr/bin/env python3
"""
Prompt Template Validator

Validiert JSON- und Markdown-Templates gegen definierte Schemas.
"""

import argparse
import json
import os
import re
import sys
from pathlib import Path
from typing import Dict, List, Optional, Tuple


# JSON Schema für JSON-Templates
JSON_SCHEMA = {
    "required": ["name", "template", "variables"],
    "properties": {
        "name": {"type": "string", "minLength": 1},
        "version": {"type": "string", "pattern": r"^\d+\.\d+$"},
        "description": {"type": "string"},
        "role": {"type": "string"},
        "template": {"type": "string", "minLength": 1},
        "variables": {"type": "object"},
        "tags": {"type": "array", "items": {"type": "string"}},
        "language": {"type": "string", "enum": ["de", "en", "fr", "es", "any"]},
    },
}

# Muster für Markdown-Templates
MD_REQUIRED_SECTIONS = ["Template", "Variablen"]


def validate_json_template(filepath: Path) -> Tuple[bool, List[str]]:
    """Validiert ein JSON-Template."""
    errors = []

    try:
        with open(filepath, "r", encoding="utf-8") as f:
            data = json.load(f)
    except json.JSONDecodeError as e:
        return False, [f"❌ JSON Syntax Error: {e}"]
    except Exception as e:
        return False, [f"❌ Datei kann nicht gelesen werden: {e}"]

    # Schemas validieren
    if not isinstance(data, dict):
        return False, ["❌ Root muss ein Object sein"]

    # Required Felder
    for field in JSON_SCHEMA.get("required", []):
        if field not in data:
            errors.append(f"❌ Fehlendes Pflichtfeld: '{field}'")

    # Feld-Typen validieren
    for field, schema in JSON_SCHEMA.get("properties", {}).items():
        if field in data:
            field_type = schema.get("type")
            if field_type == "string" and not isinstance(data[field], str):
                errors.append(f"❌ Feld '{field}' muss ein String sein")
            elif field_type == "object" and not isinstance(data[field], dict):
                errors.append(f"❌ Feld '{field}' muss ein Object sein")
            elif field_type == "array" and not isinstance(data[field], list):
                errors.append(f"❌ Feld '{field}' muss ein Array sein")

            # Pattern validieren
            if "pattern" in schema and isinstance(data[field], str):
                pattern = schema["pattern"]
                # Sicherstellen dass Pattern ^...$ hat für fullmatch
                if not (pattern.startswith("^") and pattern.endswith("$")):
                    fullmatch_pattern = f"^{pattern}$"
                else:
                    fullmatch_pattern = pattern
                if not re.fullmatch(fullmatch_pattern, data[field]):
                    errors.append(
                        f"❌ Feld '{field}' entspricht nicht dem Pattern: {pattern}"
                    )

            # Enum validieren
            if "enum" in schema and isinstance(data[field], str):
                if data[field] not in schema["enum"]:
                    errors.append(
                        f"❌ Feld '{field}' hat ungültigen Wert: '{data[field]}'. Erwartet: {schema['enum']}"
                    )

    # Template prüfen
    if "template" in data:
        template = data["template"]
        if not isinstance(template, str):
            errors.append("❌ Template muss ein String sein")
        else:
            # Variablen im Template prüfen
            if "variables" in data:
                template_vars = set(re.findall(r"\{(\w+)\}", template))
                defined_vars = set(data["variables"].keys())
                undefined_vars = template_vars - defined_vars
                if undefined_vars:
                    errors.append(
                        f"❌ Undefinierte Variablen im Template: {', '.join(undefined_vars)}"
                    )

    # Variablen validieren
    if "variables" in data:
        variables = data["variables"]
        if not isinstance(variables, dict):
            errors.append("❌ Variables muss ein Object sein")
        else:
            for var_name, var_schema in variables.items():
                if not isinstance(var_schema, dict):
                    errors.append(f"❌ Variable '{var_name}' muss ein Object sein")
                    continue
                if "type" not in var_schema:
                    errors.append(f"❌ Variable '{var_name}' benötigt ein 'type' Feld")
                if var_schema.get("type") == "enum":
                    if "values" not in var_schema:
                        errors.append(
                            f"❌ Enum Variable '{var_name}' benötigt 'values' Array"
                        )
                    elif not var_schema.get("values"):
                        errors.append(
                            f"❌ Enum Variable '{var_name}' hat leeres 'values' Array"
                        )

    return len(errors) == 0, errors


def validate_md_template(filepath: Path) -> Tuple[bool, List[str]]:
    """Validiert ein Markdown-Template."""
    errors = []

    try:
        with open(filepath, "r", encoding="utf-8") as f:
            content = f.read()
    except Exception as e:
        return False, [f"❌ Datei kann nicht gelesen werden: {e}"]

    # Mindestlänge
    if len(content.strip()) < 50:
        errors.append("⚠️ Warnung: Template zu kurz (mind. 50 Zeichen)")

    # Titel prüfen
    if not content.startswith("# "):
        errors.append("❌ Fehlender Titel (erwartet: # Titel)")

    # Pflichtabschnitte prüfen
    for section in MD_REQUIRED_SECTIONS:
        if f"## {section}" not in content and f"# {section}" not in content:
            errors.append(f"❌ Fehlender Abschnitt: {section}")

    # Variablen-Tabelle prüfen
    if "Variablen" in content:
        var_section_start = content.find("## Variablen")
        if var_section_start == -1:
            var_section_start = content.find("# Variablen")

        if var_section_start != -1:
            var_section = content[var_section_start : var_section_start + 500]
            if "| Variable |" not in var_section:
                errors.append("❌ Variablen-Tabelle nicht im korrekten Format")

    # Template-Block prüfen
    if "Template" in content:
        template_section = content.find("## Template")
        if template_section == -1:
            template_section = content.find("# Template")
        if template_section != -1:
            template_block_start = content.find("```", template_section)
            if template_block_start != -1:
                close_fence = content.find("```", template_block_start + 3)
                if close_fence != -1:
                    template_content = content[template_block_start + 3:close_fence].strip()
                    if "{" not in template_content or "}" not in template_content:
                        errors.append(
                            "⚠️ Warnung: Keine Variablen (z.B. {var}) im Template gefunden"
                        )
                else:
                    errors.append("❌ Kein schließender Code-Block gefunden")
            else:
                errors.append("❌ Kein Code-Block für Template gefunden")
        else:
            errors.append("❌ Kein Template-Abschnitt gefunden")

    return len([e for e in errors if e.startswith("❌")]) == 0, errors


def validate_template(filepath: Path) -> Tuple[bool, List[str]]:
    """Validiert ein Template basierend auf der Dateiendung."""
    if filepath.suffix.lower() == ".json":
        return validate_json_template(filepath)
    elif filepath.suffix.lower() == ".md":
        return validate_md_template(filepath)
    else:
        return False, [f"❌ Unsupported file type: {filepath.suffix}"]


ALLOWED_DIRS = {"system", "user", "custom", "categories"}


def find_templates(directory: Path) -> List[Path]:
    """Findet alle Template-Dateien in einem Verzeichnis Baum."""
    templates = []
    for root, dirs, files in os.walk(directory):
        rel = os.path.relpath(root, directory)
        if rel != "." and rel.split(os.sep)[0] not in ALLOWED_DIRS:
            dirs.clear()  # Skip nicht-erlaubte Subdirs
            continue
        for file in files:
            if file.endswith((".json", ".md")):
                templates.append(Path(root) / file)
    return templates


def main():
    parser = argparse.ArgumentParser(
        description="Validiert Prompt-Templates",
        formatter_class=argparse.RawDescriptionHelpFormatter,
        epilog="""
Beispiele:
  python validate.py templates/system/code_reviewer.json
  python validate.py --all
  python validate.py --type-json
        """,
    )
    parser.add_argument("path", nargs="?", help="Pfad zum Template oder Verzeichnis")
    parser.add_argument("--all", action="store_true", help="Alle Templates validieren")
    type_group = parser.add_mutually_exclusive_group()
    type_group.add_argument(
        "--type-json", action="store_true", help="Nur JSON-Templates validieren"
    )
    type_group.add_argument(
        "--type-md", action="store_true", help="Nur Markdown-Templates validieren"
    )

    args = parser.parse_args()

    base_dir = Path(__file__).parent.parent

    if args.all:
        # Alle Templates finden
        templates = find_templates(base_dir / "templates")

        if args.type_json:
            templates = [t for t in templates if t.suffix == ".json"]
        if args.type_md:
            templates = [t for t in templates if t.suffix == ".md"]

    elif args.path:
        path = Path(args.path)
        if path.is_dir():
            templates = find_templates(path)
            if args.type_json:
                templates = [t for t in templates if t.suffix == ".json"]
            if args.type_md:
                templates = [t for t in templates if t.suffix == ".md"]
        else:
            templates = [path]
    else:
        print("❌ Bitte Pfad angeben oder --all verwenden")
        print("Beispiel: python validate.py --all")
        sys.exit(1)

    if not templates:
        print("❌ Keine Templates gefunden")
        sys.exit(1)

    # Validierung
    total = len(templates)
    valid = 0
    invalid = 0

    print(f"\n{'=' * 60}")
    print(f"Validiere {total} Template(s)...\n")

    for template_path in sorted(templates):
        is_valid, errors = validate_template(template_path)
        # Make path relative to base_dir, handling both absolute and relative paths
        try:
            rel_path = str(template_path.relative_to(base_dir))
        except ValueError:
            # If template_path is not under base_dir, use absolute path
            rel_path = str(template_path)

        if is_valid:
            print(f"✅ {rel_path}")
            valid += 1
        else:
            print(f"❌ {rel_path}")
            for error in errors:
                print(f"   {error}")
            invalid += 1

    print(f"\n{'=' * 60}")
    print(f"Ergebnis: {valid} ✅ | {invalid} ❌ | {total} Total")

    if invalid > 0:
        sys.exit(2)  # Validierungsfehler
    sys.exit(0)


if __name__ == "__main__":
    main()