chore: checkpoint web Blender migration source, tools and docs

Accumulated uncommitted work: GHOST Web backend, WebGL/GPU web layer,
web workflow engine and protocol, tooling, docs and task records.
Large generated evidence, user-testing archives, build snapshots and
imported test assets remain untracked.

Co-Authored-By: Claude Code <noreply@anthropic.com>

tools/web/compare-blend-id-fields.py

@@ -0,0 +1,67 @@
+"""Compare serialized ID fields using Blender's reader; not a Main-equivalence verdict."""
+import argparse
+import collections
+import hashlib
+import json
+import sys
+from pathlib import Path
+
+def read_ids(blend):
+    result = {}
+    for block in blend.blocks:
+        if block.code in (b'DATA', b'DNA1', b'ENDB'): continue
+        struct = blend.structs[block.sdna_index]
+        if not any(f.dna_name.name_only == b'id' and f.dna_type.dna_type_id == b'ID' for f in struct.fields): continue
+        name = block.get((b'id', b'name'))
+        if not isinstance(name, str): raise TypeError('ID name must be decoded text')
+        key = block.code.decode('ascii') + ':' + name
+        if key in result: raise ValueError('Duplicate ID identity needs library disambiguation: ' + key)
+        blend.handle.seek(block.file_offset); result[key] = (struct, blend.handle.read(block.size))
+    return result
+
+def changed_fields(struct, before, after, prefix='', base=0):
+    for field in struct.fields:
+        if not hasattr(field.dna_type, 'fields'): raise TypeError('DNA field type must be a DNAStruct')
+        lo, hi = base + field.dna_offset, base + field.dna_offset + field.dna_size
+        if not (0 <= lo <= hi <= len(before) and hi <= len(after)): raise ValueError('DNA field outside block')
+        if before[lo:hi] == after[lo:hi]: continue
+        name = prefix + field.dna_name.name_full.decode('ascii')
+        if field.dna_type.fields and not field.dna_name.is_pointer and field.dna_name.array_size == 1:
+            yield from changed_fields(field.dna_type, before, after, name + '.', lo)
+        else:
+            yield {'field': name, 'dnaType': field.dna_type.dna_type_id.decode('ascii'), 'bytes': field.dna_size,
+                   'beforeHex': before[lo:hi][:32].hex(), 'afterHex': after[lo:hi][:32].hex(), 'previewTruncated': field.dna_size > 32}
+
+def main():
+    parser = argparse.ArgumentParser(description=__doc__)
+    parser.add_argument('before', type=Path); parser.add_argument('after', type=Path)
+    parser.add_argument('--output', required=True, type=Path); parser.add_argument('--dependency-path', type=Path)
+    args = parser.parse_args()
+    sys.path.insert(0, str(Path(__file__).resolve().parents[2] / 'blender-5.2.0/tools/modules'))
+    if args.dependency_path: sys.path.insert(0, str(args.dependency_path.resolve()))
+    try: import blendfile
+    except ModuleNotFoundError as error: raise SystemExit(f'Reader dependency missing: {error.name}; provide --dependency-path (python3-zstandard).')
+    with blendfile.open_blend(str(args.before)) as before, blendfile.open_blend(str(args.after)) as after:
+        def dna_hash(blend):
+            block = next(block for block in blend.blocks if block.code == b'DNA1')
+            blend.handle.seek(block.file_offset); return hashlib.sha256(blend.handle.read(block.size)).hexdigest()
+        if dna_hash(before) != dna_hash(after): raise ValueError('DNA schemas differ; use snapshots from one candidate')
+        left, right = read_ids(before), read_ids(after)
+        changes = []
+        for key in sorted(left.keys() & right.keys()):
+            struct, old = left[key]; other, new = right[key]
+            if struct.dna_type_id != other.dna_type_id or len(old) != len(new): raise ValueError('ID layout changed: ' + key)
+            fields = list(changed_fields(struct, old, new))
+            if fields: changes.append({'id': key, 'fields': fields})
+        inventory = lambda blend: dict(sorted(collections.Counter(block.code.decode('ascii') + ':' + blend.structs[block.sdna_index].dna_type_id.decode('ascii') for block in blend.blocks if block.code != b'ENDB').items()))
+        result = {'schemaVersion': 1, 'scope': 'Serialized ID fields and block counts; DATA contents are not a semantic equality verdict',
+                  'changedIds': changes, 'addedIds': sorted(right.keys() - left.keys()), 'removedIds': sorted(left.keys() - right.keys()),
+                  'blockCountsBefore': inventory(before), 'blockCountsAfter': inventory(after), 'dnaSha256': dna_hash(before)}
+    for name in ['before', 'after']:
+        file = getattr(args, name)
+        with file.open('rb') as stream: result[name] = {'path': str(file), 'sha256': hashlib.file_digest(stream, 'sha256').hexdigest()}
+    args.output.parent.mkdir(parents=True, exist_ok=True)
+    temporary = args.output.with_name(args.output.name + '.tmp'); temporary.write_text(json.dumps(result, indent=2) + '\n'); temporary.replace(args.output)
+    print(json.dumps({'output': str(args.output), 'changedIds': len(changes), 'addedIds': len(result['addedIds']), 'removedIds': len(result['removedIds'])}))
+
+if __name__ == '__main__': main()