low-code-react-app-b/scripts/component_migration_orchestrator.py

#!/usr/bin/env python3
"""
Batch component migration orchestrator using openai-agents.

Usage:
  python scripts/component_migration_orchestrator.py
"""
from __future__ import annotations

import json
import os
import sys
import textwrap
import difflib
import time
import random
import re
from concurrent.futures import ThreadPoolExecutor, as_completed
from dataclasses import dataclass
from pathlib import Path
from typing import Any, Dict, Iterable, List

try:
    from agents import Agent, Runner  # type: ignore
except Exception as exc:  # pragma: no cover - runtime dependency check
    raise SystemExit(
        "openai-agents is required. Install with: pip install openai-agents"
    ) from exc


ROOT = Path(__file__).resolve().parents[1]
DEFAULT_MODEL = os.getenv("CODEX_MODEL") or os.getenv("OPENAI_MODEL") or "gpt-4o-mini"
API_CALL_DELAY_SECONDS = 2.0
COMPONENT_DIRS = [
    ROOT / "src" / "components" / "atoms",
    ROOT / "src" / "components" / "molecules",
    ROOT / "src" / "components" / "organisms",
]


@dataclass(frozen=True)
class ComponentTarget:
    name: str
    path: Path
    category: str


PROMPT_TEMPLATE = """\
You are converting a React TSX component to the JSON-driven system.
Follow this workflow strictly:
1) Identify if stateful (uses useState/useEffect/other hooks).
2) If stateful, extract logic into a custom hook file in src/hooks/use-<component>.ts.
3) Create JSON definition under src/components/json-definitions/<component>.json.
4) Create interface in src/lib/json-ui/interfaces/<component>.ts.
5) Create config page schema under src/config/pages/<category>/<component-name-kebab>.json.
6) Update json-components-registry.json with a new entry for the component if missing.
7) Provide export updates for:
   - src/hooks/index.ts (if hook exists)
   - src/lib/json-ui/hooks-registry.ts (if hook exists)
   - src/lib/json-ui/interfaces/index.ts
   - src/lib/json-ui/json-components.ts
   - src/components/<category>/index.ts
8) Indicate that TSX file should be deleted.

IMPORTANT:
- Return ONLY valid JSON. No Markdown or code fences.
- For json-components.ts, output ONLY the snippet for the component export, matching this style:
  export const ComponentName = createJsonComponent<ComponentNameProps>(componentNameDef)
  OR
  export const ComponentName = createJsonComponentWithHooks<ComponentNameProps>(componentNameDef, {{ ... }})
- Do NOT output any `jsonComponents` object or registry literal.
- Provide `diffs` with unified diff lines (one string per line) for each target file,
  using the provided file contents as the "before" version.
- Include diffs for the config page schema file and json-components-registry.json.

Config page schema format (src/config/pages/<category>/<component-name-kebab>.json):
- Simple JSON-compatible component:
  {{
    "type": "ComponentName",
    "props": {{ ... }}
  }}
- Wrapper-required component (use when hooks/complex logic or not JSON-compatible):
  {{
    "type": "ComponentName",
    "jsonCompatible": false,
    "wrapperRequired": true,
    "load": {{
      "path": "@/components/<category>/ComponentName",
      "export": "ComponentName"
    }},
    "props": {{ ... }},
    "metadata": {{
      "notes": "Contains hooks - needs wrapper"
    }}
  }}

JSON definition schema (src/components/json-definitions/<component>.json):
- Single UIComponent tree using this shape:
  {{
    "id": "component-id",
    "type": "ComponentType",
    "props": {{ ... }},
    "className": "optional classes",
    "style": {{ ... }},
    "bindings": {{
      "propName": {{
        "source": "data.path",
        "path": "optional.path",
        "transform": "optional js expression"
      }}
    }},
    "dataBinding": "data.path" | {{
      "source": "data.path",
      "path": "optional.path",
      "transform": "optional js expression"
    }},
    "events": [{{ "event": "click", "actions": [{{ "id": "...", "type": "custom" }}] }}] |
              {{ "click": {{ "action": "customActionId" }} }},
    "conditional": {{
      "if": "expression",
      "then": {{ ... }} | [{{ ... }}, "..."] | "...",
      "else": {{ ... }} | [{{ ... }}, "..."] | "..."
    }},
    "loop": {{
      "source": "data.path",
      "itemVar": "item",
      "indexVar": "index"
    }},
    "children": "text" | [{{ ... }}, "..."]
  }}
- Use nested UIComponents in children.

Registry entry format (json-components-registry.json) for the component if missing:
{{
  "type": "ComponentName",
  "name": "ComponentName",
  "category": "layout|input|display|navigation|feedback|data|custom",
  "canHaveChildren": true|false,
  "description": "Short description",
  "status": "supported",
  "source": "atoms|molecules|organisms|ui|custom",
  "jsonCompatible": true|false,
  "wrapperRequired": true|false,
  "load": {{
    "path": "@/components/<category>/ComponentName",
    "export": "ComponentName"
  }},
  "metadata": {{
    "conversionDate": "YYYY-MM-DD",
    "autoGenerated": true,
    "notes": "Optional notes"
  }}
}}
Omit optional fields when not applicable.

Return ONLY valid JSON with this shape:
{{
  "componentName": "...",
  "category": "atoms|molecules|organisms",
  "isStateful": true|false,
  "hook": {{
    "name": "useComponentName",
    "filePath": "src/hooks/use-component-name.ts",
    "source": "...typescript..."
  }} | null,
  "jsonDefinition": {{
    "filePath": "src/components/json-definitions/component-name.json",
    "source": {{ ...json... }}
  }},
  "interface": {{
    "filePath": "src/lib/json-ui/interfaces/component-name.ts",
    "source": "...typescript..."
  }},
  "configPageSchema": {{
    "filePath": "src/config/pages/<category>/component-name.json",
    "source": {{ ...json... }}
  }},
  "jsonComponentExport": {{
    "filePath": "src/lib/json-ui/json-components.ts",
    "source": "...typescript snippet..."
  }},
  "exports": {{
    "hooksIndex": "...typescript snippet or null...",
    "hooksRegistry": "...typescript snippet or null...",
    "interfacesIndex": "...typescript snippet...",
    "componentsIndex": "...typescript snippet..."
  }},
  "diffs": [
    {{
      "path": "src/lib/json-ui/json-components.ts",
      "diffLines": ["@@ ...", "+...", "-..."]
    }}
  ],
  "deleteTsx": true
}}

Component category: {category}
Component path: {path}

Existing file contents for diffing:
{existing_files}

TSX source:
{tsx}
"""


CONFLICT_PROMPT_TEMPLATE = """\
You are resolving a unified diff that failed to apply cleanly.
Return ONLY valid JSON with this shape:
{{
  "path": "{path}",
  "resolvedContent": "...full file content..."
}}

Rules:
- Use the diff to update the original content.
- Preserve unrelated content.
- Do NOT return Markdown or code fences.

Original content:
{original}

Diff lines:
{diff_lines}
"""

CONFLICT_PROMPT_TEMPLATE_STRICT = """\
You must return ONLY valid JSON with BOTH keys: "path" and "resolvedContent".
"resolvedContent" must be the full file content string (no Markdown, no code fences).
Do not omit it. Do not return arrays.

JSON shape:
{{ "path": "{path}", "resolvedContent": "...full file content..." }}

Original content:
{original}

Diff lines:
{diff_lines}
"""

JSON_REPAIR_TEMPLATE = """\
Fix the following output so it is valid JSON only. Return ONLY the corrected JSON.
Do not add commentary or code fences. Preserve all keys and values.

Error:
{error}

Output to repair:
{output}
"""


def list_components(roots: Iterable[Path]) -> List[ComponentTarget]:
    targets: List[ComponentTarget] = []
    for root in roots:
        if not root.exists():
            continue
        for path in sorted(root.rglob("*.tsx")):
            name = path.stem
            targets.append(
                ComponentTarget(name=name, path=path, category=root.name)
            )
    return targets


def _strip_code_fences(output: str) -> str:
    trimmed = output.strip()
    if trimmed.startswith("```"):
        lines = trimmed.splitlines()
        if len(lines) >= 3 and lines[0].startswith("```") and lines[-1].startswith("```"):
            return "\n".join(lines[1:-1]).strip()
    return trimmed


def _to_kebab_case(name: str) -> str:
    if not name:
        return ""
    return re.sub(r"([A-Z])", r"-\1", name).lower().lstrip("-")


def _extract_json_payload(output: str) -> str:
    trimmed = _strip_code_fences(output)
    start = trimmed.find("{")
    end = trimmed.rfind("}")
    if start == -1 or end == -1 or end < start:
        return trimmed
    return trimmed[start : end + 1]


def _read_file_for_prompt(path: Path) -> str:
    if not path.exists():
        return ""
    return path.read_text(encoding="utf-8")


def _read_existing_file(out_dir: Path, rel_path: str) -> str:
    out_path = out_dir / rel_path
    if out_path.exists():
        return out_path.read_text(encoding="utf-8")
    return _read_file_for_prompt(ROOT / rel_path)


def _build_agent() -> Agent:
    return Agent(
        name="ComponentAnalyzer",
        instructions=(
            "You are a migration assistant. Given a TSX component, analyze it and "
            "produce a JSON-only response that includes: stateful analysis, "
            "hook details if needed, JSON UI definition, TS interface, "
            "and export updates. Follow the provided workflow."
        ),
        model=DEFAULT_MODEL,
    )


def _build_conflict_agent() -> Agent:
    return Agent(
        name="DiffConflictResolver",
        instructions=(
            "Resolve unified diff conflicts by producing the fully merged file content. "
            "Return JSON only with 'path' and 'resolvedContent'."
        ),
        model=DEFAULT_MODEL,
    )


def _build_json_repair_agent() -> Agent:
    return Agent(
        name="JSONRepair",
        instructions="Fix invalid JSON output. Return ONLY valid JSON, no Markdown.",
        model=DEFAULT_MODEL,
    )


def _is_rate_limited(exc: Exception) -> bool:
    status = getattr(exc, "status_code", None)
    if status == 429:
        return True
    response = getattr(exc, "response", None)
    if response is not None and getattr(response, "status_code", None) == 429:
        return True
    message = str(exc).lower()
    return "rate limit" in message or "too many requests" in message or "429" in message


def _run_with_retries(agent: Agent, prompt: str, label: str) -> Any:
    max_retries = 5
    max_attempts = max_retries + 1
    base_delay = 1.5
    max_delay = 20.0
    attempt = 0
    while True:
        try:
            attempt += 1
            print(
                (
                    f"[info] {label} attempt {attempt}/{max_attempts}: "
                    f"sleeping {API_CALL_DELAY_SECONDS:.1f}s before API call"
                ),
                file=sys.stderr,
            )
            time.sleep(API_CALL_DELAY_SECONDS)
            result = Runner.run_sync(agent, prompt)
            print(
                f"[info] {label} attempt {attempt}/{max_attempts} completed",
                file=sys.stderr,
            )
            return result
        except Exception as exc:
            if not _is_rate_limited(exc):
                print(
                    f"[error] {label} attempt {attempt} failed: {exc}",
                    file=sys.stderr,
                )
                raise
            if attempt >= max_attempts:
                raise
            delay = min(max_delay, base_delay * (2 ** (attempt - 1)))
            delay += random.uniform(0, delay * 0.2)
            print(
                f"[warn] rate limited {label}; retry {attempt}/{max_retries} in {delay:.1f}s",
                file=sys.stderr,
            )
            time.sleep(delay)


def run_agent_for_component(
    target: ComponentTarget, out_dir: Path, debug: bool = False
) -> Dict[str, Any]:
    tsx = target.path.read_text(encoding="utf-8")
    config_file_name = f"{_to_kebab_case(target.name)}.json"
    existing_files = {
        "src/lib/json-ui/json-components.ts": _read_existing_file(
            out_dir, "src/lib/json-ui/json-components.ts"
        ),
        "src/lib/json-ui/interfaces/index.ts": _read_existing_file(
            out_dir, "src/lib/json-ui/interfaces/index.ts"
        ),
        "src/hooks/index.ts": _read_existing_file(out_dir, "src/hooks/index.ts"),
        "src/lib/json-ui/hooks-registry.ts": _read_existing_file(
            out_dir, "src/lib/json-ui/hooks-registry.ts"
        ),
        "json-components-registry.json": _read_existing_file(
            out_dir, "json-components-registry.json"
        ),
        f"src/config/pages/{target.category}/{config_file_name}": _read_existing_file(
            out_dir, f"src/config/pages/{target.category}/{config_file_name}"
        ),
        f"src/components/{target.category}/index.ts": _read_existing_file(
            out_dir, f"src/components/{target.category}/index.ts"
        ),
    }
    existing_files_blob = "\n\n".join(
        f"--- {path} ---\n{content}" for path, content in existing_files.items()
    )
    prompt = PROMPT_TEMPLATE.format(
        category=target.category,
        path=target.path,
        tsx=tsx,
        existing_files=existing_files_blob,
    )
    result = _run_with_retries(_build_agent(), prompt, f"analysis:{target.name}")
    output = getattr(result, "final_output", None)
    if output is None:
        output = str(result)
    if not isinstance(output, str) or not output.strip():
        raise ValueError(
            "Agent returned empty output. Check OPENAI_API_KEY and model access."
        )
    output = _extract_json_payload(output)
    if debug:
        preview = textwrap.shorten(output.replace("\n", " "), width=300, placeholder="...")
        print(f"[debug] {target.name} raw output preview: {preview}")
    try:
        data = json.loads(output)
    except json.JSONDecodeError as exc:
        data = _repair_json_output(output, str(exc), target.name, debug)
    return data


def _resolve_diff_with_agent(
    path: str, original: str, diff_lines: List[str], debug: bool
) -> str:
    diff_blob = "\n".join(diff_lines)
    prompt = CONFLICT_PROMPT_TEMPLATE.format(
        path=path,
        original=original,
        diff_lines=diff_blob,
    )
    result = _run_with_retries(_build_conflict_agent(), prompt, f"conflict:{path}")
    output = getattr(result, "final_output", None)
    if output is None:
        output = str(result)
    if not isinstance(output, str) or not output.strip():
        raise ValueError("Conflict resolver returned empty output.")
    output = _extract_json_payload(output)
    if debug:
        preview = textwrap.shorten(output.replace("\n", " "), width=300, placeholder="...")
        print(f"[debug] conflict resolver output preview: {preview}")
    try:
        data = json.loads(output)
    except json.JSONDecodeError as exc:
        snippet = output.strip().replace("\n", " ")[:200]
        raise ValueError(
            f"Conflict resolver output was not valid JSON: {exc}. Output starts with: {snippet!r}"
        ) from exc
    resolved = data.get("resolvedContent")
    if not isinstance(resolved, str) or not resolved.strip():
        strict_prompt = CONFLICT_PROMPT_TEMPLATE_STRICT.format(
            path=path,
            original=original,
            diff_lines=diff_blob,
        )
        retry = _run_with_retries(
            _build_conflict_agent(), strict_prompt, f"conflict-strict:{path}"
        )
        retry_output = getattr(retry, "final_output", None)
        if retry_output is None:
            retry_output = str(retry)
        if not isinstance(retry_output, str) or not retry_output.strip():
            raise ValueError("Conflict resolver strict returned empty output.")
        retry_output = _extract_json_payload(retry_output)
        try:
            data = json.loads(retry_output)
        except json.JSONDecodeError as exc:
            snippet = retry_output.strip().replace("\n", " ")[:200]
            raise ValueError(
                f"Conflict resolver strict output was not valid JSON: {exc}. Output starts with: {snippet!r}"
            ) from exc
        resolved = data.get("resolvedContent")
    if not isinstance(resolved, str) or not resolved.strip():
        raise ValueError("Conflict resolver did not return resolvedContent.")
    return resolved


def _repair_json_output(output: str, error: str, label: str, debug: bool) -> Dict[str, Any]:
    prompt = JSON_REPAIR_TEMPLATE.format(error=error, output=output)
    result = _run_with_retries(_build_json_repair_agent(), prompt, f"repair:{label}")
    fixed_output = getattr(result, "final_output", None)
    if fixed_output is None:
        fixed_output = str(result)
    if not isinstance(fixed_output, str) or not fixed_output.strip():
        raise ValueError("JSON repair returned empty output.")
    fixed_output = _extract_json_payload(fixed_output)
    if debug:
        preview = textwrap.shorten(
            fixed_output.replace("\n", " "), width=300, placeholder="..."
        )
        print(f"[debug] {label} repaired output preview: {preview}")
    return json.loads(fixed_output)


def _write_if_content(path: Path, content: str) -> None:
    if not content.strip():
        return
    path.parent.mkdir(parents=True, exist_ok=True)
    path.write_text(content, encoding="utf-8")


def _merge_snippet(path: Path, content: str) -> None:
    if not content.strip():
        return
    path.parent.mkdir(parents=True, exist_ok=True)
    existing = path.read_text(encoding="utf-8") if path.exists() else ""
    existing_lines = existing.splitlines()
    new_lines = content.strip().splitlines()
    if not new_lines:
        return
    matcher = difflib.SequenceMatcher(None, existing_lines, new_lines)
    match = matcher.find_longest_match(0, len(existing_lines), 0, len(new_lines))
    if match.size == len(new_lines):
        return
    if match.a + match.size == len(existing_lines) and match.b == 0:
        merged_lines = existing_lines + new_lines[match.size:]
    else:
        merged_lines = existing_lines + ([""] if existing_lines else []) + new_lines
    merged_text = "\n".join(merged_lines).rstrip() + "\n"
    if merged_text == existing:
        return
    path.write_text(merged_text, encoding="utf-8")


def _apply_unified_diff(original: str, diff_lines: List[str]) -> str:
    if not diff_lines:
        return original
    lines = original.splitlines()
    out: List[str] = []
    i = 0
    idx = 0
    applied = False

    header_re = re.compile(r"^@@ -(\d+)(?:,(\d+))? \+(\d+)(?:,(\d+))? @@")

    while idx < len(diff_lines):
        line = diff_lines[idx]
        if line.startswith("---") or line.startswith("+++"):
            idx += 1
            continue
        if not line.startswith("@@"):
            idx += 1
            continue
        match = header_re.match(line)
        if not match:
            raise ValueError(f"Invalid diff header: {line!r}")
        old_start = int(match.group(1))
        old_start_index = max(old_start - 1, 0)
        out.extend(lines[i:old_start_index])
        i = old_start_index
        idx += 1
        applied = True
        while idx < len(diff_lines) and not diff_lines[idx].startswith("@@"):
            hunk_line = diff_lines[idx]
            if hunk_line.startswith(" "):
                expected = hunk_line[1:]
                if i >= len(lines) or lines[i] != expected:
                    raise ValueError("Diff context mismatch while applying patch.")
                out.append(lines[i])
                i += 1
            elif hunk_line.startswith("-"):
                expected = hunk_line[1:]
                if i >= len(lines) or lines[i] != expected:
                    raise ValueError("Diff delete mismatch while applying patch.")
                i += 1
            elif hunk_line.startswith("+"):
                out.append(hunk_line[1:])
            elif hunk_line.startswith("\\"):
                pass
            idx += 1

    if not applied:
        raise ValueError("No diff hunks found to apply.")
    out.extend(lines[i:])
    return "\n".join(out) + ("\n" if original.endswith("\n") else "")


def write_output(out_dir: Path, data: Dict[str, Any], target: ComponentTarget) -> None:
    component_name = data.get("componentName") or target.name or "unknown-component"

    report_path = out_dir / "migration-reports" / f"{component_name}.json"
    report_path.parent.mkdir(parents=True, exist_ok=True)
    report_path.write_text(
        json.dumps(data, indent=2, sort_keys=True),
        encoding="utf-8",
    )

    hook = data.get("hook")
    if hook and hook.get("source"):
        hook_path = out_dir / hook["filePath"]
        _write_if_content(hook_path, hook["source"])

    json_def = data.get("jsonDefinition") or {}
    json_def_path = (
        out_dir / "src" / "components" / "json-definitions" / f"{component_name}.json"
    )
    json_def_source = json.dumps(
        json_def.get("source", {}), indent=2, sort_keys=True
    )
    _write_if_content(json_def_path, json_def_source)

    interface = data.get("interface") or {}
    interface_path = out_dir / interface.get(
        "filePath", f"src/lib/json-ui/interfaces/{component_name}.ts"
    )
    _write_if_content(interface_path, interface.get("source", ""))

    diffs = data.get("diffs") or []
    if isinstance(diffs, list) and diffs:
        for diff_entry in diffs:
            path_value = diff_entry.get("path")
            diff_lines = diff_entry.get("diffLines") or []
            if not path_value or not isinstance(diff_lines, list):
                continue
            target_path = (out_dir / path_value).resolve()
            if not target_path.is_relative_to(out_dir.resolve()):
                raise ValueError(f"Diff path escapes output directory: {path_value}")
            existing = target_path.read_text(encoding="utf-8") if target_path.exists() else ""
            try:
                merged = _apply_unified_diff(existing, diff_lines)
            except ValueError as exc:
                print(
                    f"[warn] diff apply failed for {path_value}: {exc}; attempting AI resolution.",
                    file=sys.stderr,
                )
                merged = _resolve_diff_with_agent(
                    path_value, existing, diff_lines, debug=True
                )
            _write_if_content(target_path, merged)
        return

    json_component_export = (data.get("jsonComponentExport") or {}).get("source", "")
    _merge_snippet(
        out_dir / "src/lib/json-ui/json-components.ts",
        json_component_export,
    )

    exports = data.get("exports") or {}
    _merge_snippet(
        out_dir / "src/lib/json-ui/interfaces/index.ts",
        exports.get("interfacesIndex", ""),
    )
    _merge_snippet(
        out_dir / "src/hooks/index.ts", exports.get("hooksIndex", "") or ""
    )
    _merge_snippet(
        out_dir / "src/lib/json-ui/hooks-registry.ts",
        exports.get("hooksRegistry", "") or "",
    )

    components_index_path = out_dir / "src/components" / target.category / "index.ts"
    _merge_snippet(components_index_path, exports.get("componentsIndex", ""))


def main() -> int:
    out_dir = (ROOT / "migration-out").resolve()
    workers = 4

    targets = list_components(COMPONENT_DIRS)

    out_dir.mkdir(parents=True, exist_ok=True)

    failures: List[str] = []
    with ThreadPoolExecutor(max_workers=workers) as executor:
        futures = {
            executor.submit(run_agent_for_component, t, out_dir, True): t for t in targets
        }
        for future in as_completed(futures):
            target = futures[future]
            try:
                data = future.result()
                write_output(out_dir, data, target)
                print(f"ok: {target.name}")
            except Exception as exc:
                failures.append(f"{target.name}: {exc}")
                print(f"error: {target.name}: {exc}", file=sys.stderr)

    if failures:
        print("\nFailures:", file=sys.stderr)
        for failure in failures:
            print(f"- {failure}", file=sys.stderr)
        return 1
    return 0


if __name__ == "__main__":
    raise SystemExit(main())