doocus first ver

doocus/output_manager.py

@@ -0,0 +1,80 @@
+"""
+Manage per-document output folders and their sidecar files.
+
+Parallels meetus/output_manager.py but for the document contract:
+
+    docs-output/<YYYYMMDD-NNN-stem>/
+        content.md      textified document — the product
+        meta.json       structured metadata
+        thumb.jpg       optional single preview image
+        assets/         optional embedded images / per-page text (later)
+        manifest.json   run config + outputs inventory + source pointer
+"""
+from pathlib import Path
+from datetime import datetime
+import json
+import logging
+from typing import Dict, Any, Optional
+
+from .naming import resolve_run_dir
+
+logger = logging.getLogger(__name__)
+
+
+class DocOutputManager:
+    """Create and populate a run folder for one source document."""
+
+    def __init__(self, source_path: Path, base_output_dir: str = "docs-output", use_cache: bool = True):
+        self.source_path = Path(source_path)
+        self.base_output_dir = Path(base_output_dir)
+        self.use_cache = use_cache
+        self.output_dir = resolve_run_dir(self.base_output_dir, self.source_path.stem, use_cache)
+        logger.info("Output directory: %s", self.output_dir)
+
+    def get_path(self, filename: str) -> Path:
+        return self.output_dir / filename
+
+    def write_content(self, markdown: str) -> Path:
+        path = self.output_dir / "content.md"
+        path.write_text(markdown, encoding="utf-8")
+        return path
+
+    def write_meta(self, meta: Dict[str, Any]) -> Path:
+        path = self.output_dir / "meta.json"
+        with open(path, "w", encoding="utf-8") as f:
+            json.dump(meta, f, indent=2, ensure_ascii=False)
+        return path
+
+    def write_thumb(self, jpeg_bytes: bytes) -> Path:
+        path = self.output_dir / "thumb.jpg"
+        path.write_bytes(jpeg_bytes)
+        return path
+
+    def assets_dir(self) -> Path:
+        d = self.output_dir / "assets"
+        d.mkdir(exist_ok=True)
+        return d
+
+    def save_manifest(self, config: Dict[str, Any], outputs: Dict[str, Any], source_sha256: str) -> Path:
+        manifest_path = self.output_dir / "manifest.json"
+        manifest = {
+            "source": {
+                "name": self.source_path.name,
+                "path": str(self.source_path.absolute()),
+                "sha256": source_sha256,
+            },
+            "processed_at": datetime.now().isoformat(),
+            "configuration": config,
+            "outputs": outputs,
+        }
+        with open(manifest_path, "w", encoding="utf-8") as f:
+            json.dump(manifest, f, indent=2, ensure_ascii=False)
+        logger.info("Saved manifest: %s", manifest_path)
+        return manifest_path
+
+    def load_manifest(self) -> Optional[Dict[str, Any]]:
+        manifest_path = self.output_dir / "manifest.json"
+        if manifest_path.exists():
+            with open(manifest_path, "r", encoding="utf-8") as f:
+                return json.load(f)
+        return None