Various HTML improvements

copienator/commands/giving_names.py

@@ -38,6 +38,7 @@ def build_parser() -> argparse.ArgumentParser:
             "directories, matching folders by their trailing copy ID"
         ),
     )
+    parser.add_argument("--incremental", action="store_true", help="Skip unchanged return folders")
     return parser
 
 
@@ -147,6 +148,7 @@ def update_named_return_answers(
 def prepare_named_returns(
     workspace: EvaluationWorkspace,
     annotation_dir_name: str,
+    *, incremental: bool = False,
 ) -> ExitCode:
     workspace.require_directories("Copies", annotation_dir_name)
     workspace.return_dir.mkdir(parents=True, exist_ok=True)
@@ -198,6 +200,12 @@ def prepare_named_returns(
             assigned_names.add(name)
             destination = existing_returns.get(int(copy_id), workspace.return_dir / f"{safe_name} ({copy_id})")
             output_name = destination.name.rsplit(" (", 1)[0]
+            from copienator.reviewed_cache import ReturnCache
+            cache = ReturnCache(workspace.root, source_folder, destination) if incremental else None
+            if cache and cache.unchanged():
+                print(f"Skipping unchanged return {destination.name}")
+                continue
+            copy_failed = False
             destination.mkdir(parents=True, exist_ok=True)
             try:
                 publish_answer_returns(workspace.root, source_folder, destination)
@@ -238,6 +246,9 @@ def prepare_named_returns(
                         file=sys.stderr,
                     )
                     had_errors = True
+                    copy_failed = True
+            if cache and not copy_failed:
+                cache.remember()
 
     unassigned = expected_names - assigned_names
     if unassigned:
@@ -253,10 +264,11 @@ def run(
     *,
     annotation_dir: str,
     update: bool = False,
+    incremental: bool = False,
 ) -> ExitCode:
     if update:
         return update_named_return_answers(workspace, annotation_dir)
-    return prepare_named_returns(workspace, annotation_dir)
+    return prepare_named_returns(workspace, annotation_dir, incremental=incremental)
 
 
 def main(argv: Sequence[str] | None = None) -> int:
@@ -268,6 +280,7 @@ def main(argv: Sequence[str] | None = None) -> int:
             workspace_from_args(args, repository=Path.cwd()),
             annotation_dir=args.annotation_dir,
             update=args.update,
+            incremental=args.incremental,
         ),
     )
 

copienator/commands/modify_annotations.py

@@ -25,16 +25,22 @@ from copienator.annotation_modifications import (
 from copienator.filesystem import publish_file_sets
 
 
-def run(workspace: EvaluationWorkspace, *, annotation_dir: str = "BGnot") -> ExitCode:
+def run(workspace: EvaluationWorkspace, *, annotation_dir: str = "BGnot", incremental: bool = False) -> ExitCode:
     workspace.require_files("labels", DRAFT)
     workspace.require_directories("Copies", annotation_dir)
     text = (workspace.root / DRAFT).read_text(encoding="utf-8")
     instructions = read_instructions(text, annotation_dir)
-    validate_instructions(workspace.root, annotation_dir, instructions)
     directory = workspace.root / annotation_dir
     applied = directory / APPLIED
     previous = read_instructions(applied.read_text(encoding="utf-8"), annotation_dir) if applied.exists() else []
+    if incremental and previous == instructions:
+        print("Skipping unchanged transfer instructions")
+        return ExitCode.SUCCESS
+    validate_instructions(workspace.root, annotation_dir, instructions)
     copy_ids = sorted({item.copy_id for item in [*previous, *instructions]})
+    if incremental:
+        copy_ids = [cid for cid in copy_ids
+                    if [i for i in previous if i.copy_id == cid] != [i for i in instructions if i.copy_id == cid]]
     # Prepare every copy before publishing anything, including the applied file.
     with TemporaryDirectory(prefix=".annotation-edits-", dir=directory) as temporary:
         staging_root = Path(temporary)
@@ -61,12 +67,13 @@ def run(workspace: EvaluationWorkspace, *, annotation_dir: str = "BGnot") -> Exi
 def build_parser():
     parser = evaluation_parser("Modify reviewed answer bodies using annotation_modifications.txt")
     parser.add_argument("--annotation-dir", choices=MODES, default="BGnot")
+    parser.add_argument("--incremental", action="store_true", help="Skip unchanged transfer instructions")
     return parser
 
 
 def main(argv: Sequence[str] | None = None) -> int:
     return execute(build_parser(), argv, lambda args: run(
-        workspace_from_args(args), annotation_dir=args.annotation_dir))
+        workspace_from_args(args), annotation_dir=args.annotation_dir, incremental=args.incremental))
 
 
 if __name__ == "__main__":

copienator/commands/regenerate_reviewed.py

@@ -5,9 +5,19 @@ from copienator.commands.reading_annotations import apply_actions_and_regenerate
 from copienator.commands.reading_grouped_annotations import apply_actions_and_regenerate_grouped
 
 
-def run(workspace: EvaluationWorkspace, mode: str):
+def run(workspace: EvaluationWorkspace, mode: str, *, incremental: bool = False):
     directories = sorted((p for p in (workspace.root / mode).glob('Copie*')
                           if p.is_dir() and p.name[5:].isdigit() and (p / 'score.json').is_file()))
+    from copienator.reviewed_cache import ReviewedReadCache
+    cache = ReviewedReadCache(workspace, mode, directories) if incremental else None
+    if cache:
+        pending = []
+        for directory in directories:
+            if cache.unchanged(directory.name[5:]):
+                print(f"Skipping unchanged {directory.name}", flush=True)
+            else:
+                pending.append(directory)
+        directories = pending
     # Validate every snapshot before starting to publish any copy.
     for directory in directories:
         if not (directory / FILENAME).is_file():
@@ -29,10 +39,14 @@ def run(workspace: EvaluationWorkspace, mode: str):
             print(log, flush=True)
         if result != ExitCode.SUCCESS:
             status = ExitCode.PARTIAL
+        elif cache:
+            # Rendering can normalize the saved snapshot; fingerprint the published inputs.
+            ReviewedReadCache(workspace, mode, [directory]).remember([cid])
     return status
 
 
 def main(argv=None):
     parser = evaluation_parser('Regenerate from saved reviewed inputs without rescanning handwriting')
     parser.add_argument('--annotation-dir', choices=('Bnot', 'BGnot'), required=True)
-    return execute(parser, argv, lambda args: run(workspace_from_args(args), args.annotation_dir))
+    parser.add_argument("--incremental", action="store_true", help="Skip unchanged copies")
+    return execute(parser, argv, lambda args: run(workspace_from_args(args), args.annotation_dir, incremental=args.incremental))