Copy removal

copienator/commands/clean.py

@@ -153,7 +153,8 @@ def build_cleanup_plan(workspace: EvaluationWorkspace) -> CleanupPlan:
 
     files, directories = _workspace_entries(workspace.root)
 
-    kept_files = set(processed_copies)
+    kept_files = {path for path in files if path.is_relative_to(workspace.root / "BackUp")}
+    kept_files.update(processed_copies)
     kept_files.add(workspace.correction_file)
     lock_file = workspace.root / ".copienator-finalisation.lock"
     if lock_file in files:

copienator/copy_removal.py

@@ -0,0 +1,195 @@
+"""Reversible removal of one stable copy ID and its working artifacts."""
+from __future__ import annotations
+
+import io
+import json
+import os
+import re
+import shutil
+import uuid
+from dataclasses import dataclass
+from datetime import datetime
+from pathlib import Path
+
+from PIL import Image, ImageDraw
+
+from copienator import atomic_write_bytes, read_json
+from copienator.evaluation_lock import EvaluationLock
+
+
+@dataclass(frozen=True)
+class RemovalPlan:
+    root: Path
+    copy: str
+    removed: tuple[Path, ...]
+    updated: dict[Path, bytes]
+
+
+def available_copies(root: Path) -> list[str]:
+    names = set()
+    for folder in (root, root / "Copies", root / "Copies Originales"):
+        names.update(p.stem for p in folder.glob("Copie*.pdf")
+                     if re.fullmatch(r"Copie[0-9]+", p.stem))
+    return sorted(names, key=lambda name: int(name[5:]))
+
+
+def _json_bytes(data) -> bytes:
+    return (json.dumps(data, ensure_ascii=False, indent=2) + "\n").encode()
+
+
+def removal_plan(root: Path, copy: str) -> RemovalPlan:
+    root = root.resolve()
+    if not re.fullmatch(r"Copie[0-9]+", copy) or copy not in available_copies(root):
+        raise ValueError("Choisissez une copie existante.")
+    ident = copy[5:]
+    def same_id(value):
+        return isinstance(value, (str, int)) and str(value).removeprefix("Copie").isdecimal() and int(str(value).removeprefix("Copie")) == int(ident)
+    def owned(name):
+        match = re.match(r"^Copie([0-9]+)(?:$|[_. -])", name)
+        return bool(match and same_id(match[1]))
+
+    removed = set()
+    files = []
+    # Never traverse symlinks or the backups themselves.
+    for current, dirs, names in os.walk(root, followlinks=False):
+        directory = Path(current)
+        dirs[:] = [name for name in dirs if name not in {"BackUp", ".git"}
+                   and not (directory / name).is_symlink()]
+        for name in list(dirs):
+            returned = re.search(r" \(([0-9]+)\)$", name) if directory.name == "A Rendre" else None
+            if owned(name) or (returned and same_id(returned[1])):
+                removed.add(directory / name)
+                dirs.remove(name)
+        for name in names:
+            path = directory / name
+            if path.is_symlink():
+                continue
+            if owned(name):
+                removed.add(path)
+            else:
+                files.append(path)
+
+    updated = {}
+    dropped = object()
+    def prune(value, *, rows=False, numeric_keys=False):
+        if isinstance(value, dict):
+            if any(same_id(value.get(key)) for key in ("id", "copy_id", "copy", "copie")):
+                return dropped
+            result = {}
+            for key, item in value.items():
+                if owned(key) or (numeric_keys and same_id(key)):
+                    continue
+                new = prune(item, rows=rows, numeric_keys=numeric_keys)
+                if new is not dropped:
+                    result[key] = new
+            return result
+        if isinstance(value, list):
+            if value and ((rows and same_id(value[0])) or
+                          (isinstance(value[0], str) and value[0] == copy)):
+                return dropped
+            return [new for item in value
+                    if (new := prune(item, rows=rows, numeric_keys=numeric_keys)) is not dropped]
+        return value
+
+    for path in files:
+        relative = path.relative_to(root)
+        # Historical logs and GUI history are audit records, not active inputs.
+        if "logs" in relative.parts or path.name == ".copienator-gui.json":
+            continue
+        if path.name in {"manual_resolutions.txt", "annotation_modifications.txt"}:
+            text = path.read_text()
+            pattern = re.compile(r"\bCopie0*" + str(int(ident)) + r"\b")
+            new_text = "".join(line for line in text.splitlines(keepends=True) if not pattern.search(line))
+            if new_text != text:
+                updated[path] = new_text.encode()
+        if path.suffix == ".json":
+            data = read_json(path)  # Fail before moving anything if an input is malformed.
+            is_group = path.name.startswith("Group_") and "Par label" in relative.parts
+            if is_group and isinstance(data, list):
+                targets = [entry for entry in data if isinstance(entry, list) and entry and same_id(entry[0])]
+                if targets:
+                    remaining = [entry for entry in data if entry not in targets]
+                    image_path = path.with_suffix(".jpg")
+                    if not remaining:
+                        removed.add(path)
+                        if image_path.exists():
+                            removed.add(image_path)
+                        continue
+                    # Blank the removed answer without shifting other students' coordinates.
+                    from copienator.commands.grouping import LABEL_HEIGHT, SEPARATOR_HEIGHT
+                    with Image.open(image_path) as source:
+                        image = source.convert("RGB")
+                    draw = ImageDraw.Draw(image)
+                    for entry in targets:
+                        top = int(entry[1]) - LABEL_HEIGHT
+                        draw.rectangle((0, max(0, top - SEPARATOR_HEIGHT), image.width, int(entry[2]) - 1), fill="white")
+                    buffer = io.BytesIO()
+                    image.save(buffer, format="JPEG", quality=95)
+                    updated[image_path] = buffer.getvalue()
+            numeric_keys = path.name in {"custom_remarks.json", "general_remarks.json"}
+            new = prune(data, rows=is_group or path.name == "refaire.json", numeric_keys=numeric_keys)
+            if new is dropped:
+                removed.add(path)
+            elif new != data:
+                updated[path] = _json_bytes(new)
+
+    # Cached contact sheets and batch payloads contain several students. Discard
+    # them as a unit so stale recognition/results cannot reintroduce the copy.
+    for directory in (root / "Cutleft/Names/Grids",):
+        if directory.is_dir() and not directory.is_symlink():
+            removed.add(directory)
+    for path in files:
+        if (path.name.startswith(("batch_", "batched_")) or
+                path.name in {"correction_progress.json", "correction_pending_responses.json",
+                              "histogramme.pdf", "merged_histogramme.pdf"}):
+            removed.add(path)
+    # Collapse children when a whole directory is being removed.
+    removed = {p for p in removed if not any(parent in removed for parent in p.parents)}
+    updated = {p: data for p, data in updated.items()
+               if p not in removed and not any(parent in removed for parent in p.parents)}
+    return RemovalPlan(root, copy, tuple(sorted(removed)), updated)
+
+
+def remove_copy(root: Path, copy: str) -> Path:
+    """Back up every changed file first; roll back all edits on failure."""
+    root = root.resolve()
+    if (root / "BackUp").is_symlink():
+        raise ValueError("BackUp doit être un dossier dans l’évaluation, pas un lien symbolique.")
+    with EvaluationLock(root):
+        plan = removal_plan(root, copy)
+        backup = plan.root / "BackUp" / (datetime.now().strftime("%Y%m%d-%H%M%S") + "-" + copy + "-" + uuid.uuid4().hex[:8])
+        originals = backup / "files"
+        originals.mkdir(parents=True)
+        paths = [*plan.removed, *plan.updated]
+        for path in paths:
+            target = originals / path.relative_to(plan.root)
+            target.parent.mkdir(parents=True, exist_ok=True)
+            if path.is_dir():
+                shutil.copytree(path, target, symlinks=True)
+            else:
+                shutil.copy2(path, target)
+        manifest = {"copy": copy, "removed": [str(p.relative_to(plan.root)) for p in plan.removed],
+                    "updated": [str(p.relative_to(plan.root)) for p in plan.updated], "status": "prepared"}
+        atomic_write_bytes(backup / "manifest.json", _json_bytes(manifest))
+        changed = []
+        try:
+            for path in plan.removed:
+                changed.append(path)
+                if path.is_dir():
+                    shutil.rmtree(path)
+                else:
+                    path.unlink()
+            for path, data in plan.updated.items():
+                changed.append(path)
+                atomic_write_bytes(path, data)
+            manifest["status"] = "complete"
+            atomic_write_bytes(backup / "manifest.json", _json_bytes(manifest))
+        except Exception:
+            for path in reversed(changed):
+                source = originals / path.relative_to(plan.root)
+                if source.is_dir():
+                    shutil.copytree(source, path, dirs_exist_ok=True, symlinks=True)
+                else:
+                    shutil.copy2(source, path)
+            raise
+        return backup