Support for post annotation label reassignement.

copienator/commands/reading_annotations.py

@@ -9,20 +9,26 @@ import numpy as np
 from pdf2image import convert_from_path
 from PIL import Image, ImageChops, ImageDraw, ImageFilter
 
-from copienator.commands import annotating
-from copienator import configuration, utils
 from copienator import (
     EvaluationWorkspace,
     ExitCode,
     atomic_write_json,
+    configuration,
     evaluation_parser,
     execute,
     read_json,
+    utils,
     workspace_from_args,
 )
 from copienator.annotation_actions import apply_checkbox_actions, apply_score_overrides
-from copienator.answer_info import build_answer_info
 from copienator.annotation_data import AnnotationData, load_annotation_data
+from copienator.annotation_modifications import (
+    BASELINE,
+    output_removals,
+    stage_baseline,
+)
+from copienator.answer_info import build_answer_info
+from copienator.commands import annotating
 from copienator.filesystem import staged_files
 from copienator.return_answers import save_return_answer_options
 
@@ -168,6 +174,8 @@ def apply_actions_and_regenerate(
     scores = dict.fromkeys(all_labels, "")
     answer_labels: list[str] = []
     dirty_images: dict[str, Image.Image] = {}
+    header_heights: dict[str, int] = {}
+    filtered_labels: list[str] = []
     concatenated: list[Image.Image] = []
     filtered: list[Image.Image] = []
     incomplete = False
@@ -221,15 +229,18 @@ def apply_actions_and_regenerate(
                 final_image.paste(body, (0, new_header_height), mask=body)
 
         dirty_images[label] = final_image
+        header_heights[label] = new_header_height
         answer_labels.append(label)
         concatenated.append(final_image)
         if float(scores[label]) != 4.0 or result.get("feedback", []):
             filtered.append(final_image)
+            filtered_labels.append(label)
 
+    if incomplete:
+        return ExitCode.PARTIAL
     concat_image = concatenate(concatenated)
     filtered_image = concatenate(filtered)
-    with staged_files(output_dir, remove=("Concat.jpg", "Concat_F.jpg", "Concat_F.pdf",
-                                         "touched.json", "answer_labels.json")) as staging:
+    with staged_files(output_dir, remove=(*output_removals(output_dir), BASELINE)) as staging:
         for label, image in dirty_images.items():
             image.save(staging / f"{label}.jpg")
         if not preserve_score_file:
@@ -241,6 +252,9 @@ def apply_actions_and_regenerate(
             concat_image.save(staging / "Concat.jpg")
         if filtered_image is not None:
             filtered_image.save(staging / "Concat_F.jpg")
+        stage_baseline(workspace.root, "Bnot", output_dir, staging, dirty_images,
+                       header_heights, build_answer_info(scores, labels_data, answer_labels),
+                       filtered_labels, scores)
 
     if preserve_score_file:
         print(f"  Preserved existing score.json in {output_dir}")

copienator/commands/reading_grouped_annotations.py

@@ -9,11 +9,11 @@ from typing import Any
 
 from PIL import Image, ImageDraw
 
-from copienator import configuration
 from copienator import (
     EvaluationWorkspace,
     ExitCode,
     atomic_write_json,
+    configuration,
     evaluation_parser,
     execute,
     read_json,
@@ -21,8 +21,15 @@ from copienator import (
     workspace_from_args,
 )
 from copienator.annotation_actions import apply_checkbox_actions, apply_score_overrides
-from copienator.answer_info import build_answer_info
 from copienator.annotation_data import AnnotationData, RefaireList, load_annotation_data
+from copienator.annotation_modifications import (
+    BASELINE,
+    baseline,
+    load_image,
+    output_removals,
+    stage_baseline,
+)
+from copienator.answer_info import build_answer_info
 from copienator.commands import annotating
 from copienator.commands.reading_annotations import (
     concatenate,
@@ -247,6 +254,9 @@ def apply_actions_and_regenerate_grouped(
     touched = dict.fromkeys(all_labels, False)
     answer_labels: list[str] = []
     dirty_images: dict[str, Image.Image] = {}
+    header_heights: dict[str, int] = {}
+    filtered_labels: list[str] = []
+    previous_baseline = baseline(output_dir) if (output_dir / BASELINE).is_file() else None
     concat_images: list[Image.Image] = []
     filtered_groups: list[list[Image.Image]] = []
     incomplete = False
@@ -266,9 +276,19 @@ def apply_actions_and_regenerate_grouped(
         if result.get("error") == "empty-answer":
             continue
         saved_image = output_dir / f"{label}.jpg"
-        if selected_labels and label not in selected_labels and saved_image.is_file():
-            with Image.open(saved_image) as saved:
-                final_image = saved.convert("RGB").copy()
+        baseline_entry = (previous_baseline or {}).get("images", {}).get(label)
+        if selected_labels and label not in selected_labels and (saved_image.is_file() or baseline_entry):
+            if annotation_dir in {"BGnot", "Bnot"}:
+                if previous_baseline is not None and label in previous_baseline["images"]:
+                    entry = previous_baseline["images"][label]
+                    final_image = load_image(output_dir, entry)
+                    header_heights[label] = entry["header_height"]
+                else:
+                    with Image.open(saved_image) as saved:
+                        final_image = saved.convert("RGB").copy()
+            else:
+                with Image.open(saved_image) as saved:
+                    final_image = saved.convert("RGB").copy()
             if (
                 simple_annotated is not None
                 and label in simple_layout["images"]
@@ -278,10 +298,11 @@ def apply_actions_and_regenerate_grouped(
                 final_image = simple_annotated.crop(
                     (0, hmin, simple_annotated.width, hmax)
                 )
-                dirty_images[label] = final_image
+            dirty_images[label] = final_image
             scores[label] = str(old_scores.get(label, scores[label]))
             concat_images.append(final_image)
             answer_labels.append(label)
+            filtered_labels.append(label)
             # Keep previously reviewed content, including handwriting.
             if annotation_dir == "BGnot":
                 extras = get_extra_pdfs_as_images(
@@ -327,6 +348,7 @@ def apply_actions_and_regenerate_grouped(
 
         # Persist every final block, including unchanged answers, for returns.
         dirty_images[label] = final_image
+        header_heights[label] = new_header_height
         answer_labels.append(label)
         concat_images.append(final_image)
 
@@ -335,6 +357,7 @@ def apply_actions_and_regenerate_grouped(
             feedback.get("to_delete", False) for feedback in feedbacks
         )
         if not perfect or has_notes:
+            filtered_labels.append(label)
             extras = (
                 get_extra_pdfs_as_images(workspace.root, label, annotating, all_labels)
                 if annotation_dir == "BGnot"
@@ -346,8 +369,7 @@ def apply_actions_and_regenerate_grouped(
     concat_image = concatenate(concat_images)
     if incomplete:
         return ExitCode.PARTIAL, "\n".join(logs)
-    with staged_files(output_dir, remove=("Concat.jpg", "Concat_F.pdf", "Concat_F.jpg",
-                                         "touched.json", "answer_labels.json")) as staging:
+    with staged_files(output_dir, remove=(*output_removals(output_dir), BASELINE)) as staging:
         if simple_layout is not None:
             simple_layout["replaced"] = sorted(
                 set(simple_layout["replaced"]) | selected_labels
@@ -370,6 +392,9 @@ def apply_actions_and_regenerate_grouped(
                     [image for group in filtered_groups for image in group]
                 )
                 filtered_image.save(staging / "Concat_F.jpg")
+        stage_baseline(workspace.root, annotation_dir, output_dir, staging, dirty_images,
+                       header_heights, build_answer_info(scores, labels_data, answer_labels, touched),
+                       filtered_labels, scores)
     if preserve_score_file:
         logs.append(f"  Preserved existing score.json in {output_dir}")
     logs.append(f"  Saved regenerated files in {output_dir}")