Initial backup: folder structure manifest, tree summary, scraper scripts, and text metadata

generate_manifest.py

@@ -0,0 +1,41 @@
+import os
+import json
+import time
+
+root_dir = os.path.dirname(os.path.abspath(__file__))
+manifest_path = os.path.join(root_dir, "workspace_manifest.json")
+summary_path = os.path.join(root_dir, "workspace_tree_summary.txt")
+
+folders = []
+counts = {"dirs": 0, "files": 0, "by_ext": {}}
+
+print(f"Scanning {root_dir}...")
+start_time = time.time()
+
+with open(summary_path, "w", encoding="utf-8") as sum_f:
+    for dirpath, dirnames, filenames in os.walk(root_dir):
+        # Exclude git directories and system temp
+        dirnames[:] = [d for d in dirnames if d not in ('.git', '__pycache__')]
+        rel_dir = os.path.relpath(dirpath, root_dir)
+        counts["dirs"] += 1
+        
+        sum_f.write(f"[{rel_dir}] (dirs: {len(dirnames)}, files: {len(filenames)})\n")
+        
+        folders.append({
+            "path": rel_dir if rel_dir != "." else "",
+            "subdirs": list(dirnames),
+            "file_count": len(filenames),
+        })
+
+        for f in filenames:
+            counts["files"] += 1
+            ext = os.path.splitext(f)[1].lower()
+            counts["by_ext"][ext] = counts["by_ext"].get(ext, 0) + 1
+
+elapsed = time.time() - start_time
+print(f"Scan finished in {elapsed:.1f}s. Total dirs: {counts['dirs']}, files: {counts['files']}.")
+
+with open(manifest_path, "w", encoding="utf-8") as f:
+    json.dump({"root": os.path.basename(root_dir), "stats": counts, "directories": folders}, f, indent=2)
+
+print(f"Saved manifest to {manifest_path} and {summary_path}")