Standardisation 8

Readme.org

@@ -174,6 +174,8 @@ scripts migrés vers cette convention sont actuellement :
 - =page_splitter.py=, =cutleft.py=, =plotting.py= et
   =splitting_int.py= ;
 - =gemini_for_labels.py= ;
+- =correction.py=, =submit_batches.py=, =batch_status.py= et
+  =fetch_batched_results.py= ;
 - =annotating.py=, =annotating_with_checks.py= et
   =annotating_by_label.py= ;
 - =reading_annotations.py= et =reading_grouped_annotations.py= ;
@@ -343,6 +345,13 @@ Optional : Set proxy with ~export HTTPS_PROXY="http://10.0.0.1:3128"~
 
     Fais les requêtes de correction à Gemini.
 
+    =correction.py= peut être relancé sans supprimer son état. Les
+    fichiers =correction.json= et =correction_progress.json= sont mis à
+    jour atomiquement. Avec =--overwrite=, leur version précédente reste
+    en place jusqu'à la première écriture réussie de la nouvelle
+    exécution. =--reset= est la seule option qui supprime explicitement
+    cet état et restaure les fichiers =*_old.pdf=.
+
     L'argument =limit= limite le nombre de requêtes à Gemini Pro
     (chères), pour une version low cost, passer =--limit 0=, toutes
     les requêtes seront sur Gemini Flash.
@@ -357,6 +366,16 @@ Optional : Set proxy with ~export HTTPS_PROXY="http://10.0.0.1:3128"~
     + =python batch_status.py=
     + =python fetch_batched_results.py Interro=
     + =python correction.py Interro --deal-with-batched=
+
+    Les quatre commandes de ce flux suivent la convention des scripts
+    standardisés. Les fichiers de requêtes et le résultat JSONL combiné
+    sont publiés atomiquement : une interruption ne laisse pas de fichier
+    final partiellement écrit. =submit_batches.py= conserve aussi les
+    identifiants distants dans =batch_jobs.json= ; la récupération les
+    utilise en priorité et garde la recherche par nom pour les anciens
+    batchs. =batch_status.py --download JOB --output
+    resultat.jsonl= permet aussi de télécharger atomiquement le résultat
+    d'un job particulier.
  3. =python post-correction.py Interro=
 
     - Essaye de corriger des erreurs d'encodage/d'accents dans

batch_status.py

@@ -1,88 +1,93 @@
-import os
+from __future__ import annotations
-import sys
+
 import argparse
+from collections.abc import Sequence
+from pathlib import Path
+
 from google import genai
 
-if "GEMINI_API_KEY" not in os.environ:
+import config
-    sys.exit("Error: GEMINI_API_KEY environment variable not set.")
+from copienator import (
-
+    CliError,
-client = genai.Client()
+    ExitCode,
-
+    atomic_write_bytes,
-def list_jobs():
+    execute,
-    print("Fetching recent batch jobs...\n")
+    standard_parser,
-    try:
+)
-        batch_jobs = client.batches.list()
-        jobs_found = False
-
-        for job in batch_jobs:
-            jobs_found = True
-            state = job.state.name if hasattr(job.state, 'name') else job.state
-
-            print("-" * 60)
-            print(f"Job Name:      {job.name}")
-
-            if hasattr(job, 'display_name') and job.display_name:
-                print(f"Display Name:  {job.display_name}")
-
-            print(f"State:         {state}")
-
-            if state == 'JOB_STATE_FAILED' and hasattr(job, 'error'):
-                print(f"Error:         {job.error}")
-
-            if state == 'JOB_STATE_SUCCEEDED' and hasattr(job, 'dest') and job.dest:
-                if hasattr(job.dest, 'file_name') and job.dest.file_name:
-                    print(f"Output File:   {job.dest.file_name}")
-
-        if not jobs_found:
-            print("No batch jobs found.")
-        else:
-            print("-" * 60)
-            print("\nTo download a completed job, run:")
-            print("python batch_status.py --download batches/<YOUR_BATCH_ID>")
-
-    except Exception as e:
-        sys.exit(f"An error occurred while listing jobs: {e}")
 
 
-def download_job(job_name):
+def _client():
-    print(f"Checking status for {job_name}...\n")
+    if not config.API_KEY:
-    try:
+        raise CliError("GEMINI_API_KEY is not configured")
-        job = client.batches.get(name=job_name)
+    return genai.Client(api_key=config.API_KEY)
-        state = job.state.name if hasattr(job.state, 'name') else job.state
 
-        print(f"State: {state}")
 
-        if state != 'JOB_STATE_SUCCEEDED':
+def list_jobs(*, client=None) -> ExitCode:
-            print("Job is not ready yet or has failed.")
+    client = client or _client()
-            if state == 'JOB_STATE_FAILED' and hasattr(job, 'error'):
+    print("Fetching recent batch jobs...")
-                print(f"Error: {job.error}")
+    jobs = list(client.batches.list())
-            return
+    for job in jobs:
+        state = job.state.name if hasattr(job.state, "name") else job.state
+        print(f"{job.name}: {state}")
+        if getattr(job, "display_name", None):
+            print(f"  Display name: {job.display_name}")
+        if state == "JOB_STATE_FAILED" and getattr(job, "error", None):
+            print(f"  Error: {job.error}")
+        destination = getattr(job, "dest", None)
+        if state == "JOB_STATE_SUCCEEDED" and getattr(
+            destination, "file_name", None
+        ):
+            print(f"  Output file: {destination.file_name}")
+    if not jobs:
+        print("No batch jobs found.")
+    return ExitCode.SUCCESS
 
-        if hasattr(job, 'dest') and job.dest and hasattr(job.dest, 'file_name') and job.dest.file_name:
-            result_file_name = job.dest.file_name
-            print(f"Downloading results from {result_file_name}...")
 
-            file_content_bytes = client.files.download(file=result_file_name)
+def download_job(
-            output_path = f"results_{job_name.replace('/', '_')}.jsonl"
+    job_name: str,
+    *,
+    output: Path | None = None,
+    client=None,
+) -> ExitCode:
+    client = client or _client()
+    job = client.batches.get(name=job_name)
+    state = job.state.name if hasattr(job.state, "name") else job.state
+    print(f"State: {state}")
+    if state != "JOB_STATE_SUCCEEDED":
+        if state == "JOB_STATE_FAILED" and getattr(job, "error", None):
+            print(f"Error: {job.error}")
+        return ExitCode.PARTIAL
+    destination = getattr(job, "dest", None)
+    file_name = getattr(destination, "file_name", None)
+    if not file_name:
+        print("Job succeeded but no output file was found.")
+        return ExitCode.PARTIAL
+    payload = client.files.download(file=file_name)
+    output_path = output or Path(f"results_{job_name.replace('/', '_')}.jsonl")
+    atomic_write_bytes(output_path, payload)
+    print(f"Saved batch results to {output_path}")
+    return ExitCode.SUCCESS
 
-            with open(output_path, "wb") as f:
-                f.write(file_content_bytes)
 
-            print(f"Success! Saved to {output_path}")
+def build_parser() -> argparse.ArgumentParser:
-            print(f"You can now feed this to your correction script using: --deal-with-batched {output_path}")
+    parser = standard_parser("List or download Gemini correction batch jobs")
-        else:
+    parser.add_argument("--download", metavar="JOB_NAME")
-            print("Job succeeded but no output file was found.")
+    parser.add_argument("--output", type=Path, help="Downloaded JSONL destination")
+    return parser
+
+
+def main(argv: Sequence[str] | None = None) -> int:
+    parser = build_parser()
+
+    def handle(args: argparse.Namespace) -> ExitCode:
+        if args.output is not None and not args.download:
+            raise CliError("--output requires --download", ExitCode.INVALID_ARGUMENTS)
+        if args.download:
+            return download_job(args.download, output=args.output)
+        return list_jobs()
+
+    return execute(parser, argv, handle)
 
-    except Exception as e:
-        sys.exit(f"An error occurred while fetching the job: {e}")
 
 if __name__ == "__main__":
-    parser = argparse.ArgumentParser(description="Manage Gemini Batch Jobs")
+    raise SystemExit(main())
-    parser.add_argument("--download", type=str, metavar="JOB_NAME",
-                        help="Download the results for a specific batch job (e.g. batches/123456)")
-    args = parser.parse_args()
-
-    if args.download:
-        download_job(args.download)
-    else:
-        list_jobs()

copienator/__init__.py

@@ -14,6 +14,7 @@ from .cli import (
 from .json_io import (
     JsonLockTimeout,
     atomic_update_json,
+    atomic_write_bytes,
     atomic_write_json,
     atomic_write_text,
     read_json,
@@ -32,6 +33,7 @@ __all__ = [
     "WorkspaceNotFoundError",
     "WorkspaceValidationError",
     "atomic_update_json",
+    "atomic_write_bytes",
     "atomic_write_json",
     "atomic_write_text",
     "evaluation_parser",

copienator/json_io.py

@@ -78,6 +78,10 @@ def atomic_write_text(
     _atomic_write(Path(path), text.encode(encoding))
 
 
+def atomic_write_bytes(path: str | Path, payload: bytes) -> None:
+    _atomic_write(Path(path), payload)
+
+
 def atomic_write_json(
     path: str | Path,
     value: JsonValue,

copienator/workspace.py

@@ -101,6 +101,14 @@ class EvaluationWorkspace:
     def correction_progress_file(self) -> Path:
         return self.root / "correction_progress.json"
 
+    @property
+    def batch_jobs_file(self) -> Path:
+        return self.root / "batch_jobs.json"
+
+    @property
+    def batched_correction_result_file(self) -> Path:
+        return self.root / "batched_correction_result.jsonl"
+
     @property
     def manual_resolutions_file(self) -> Path:
         return self.root / "manual_resolutions.txt"

correction.py

@@ -1,91 +1,59 @@
-import sys
+from __future__ import annotations
-import os
+
-import time
-from pathlib import Path
 import argparse
-import prompting
-import signal
-from google import genai
 import base64
-import shlex
-import json
-import threading
 import concurrent.futures
+import json
+import os
+import shlex
+import shutil
+import sys
+import threading
+import time
+from collections.abc import Sequence
+from pathlib import Path
 
-from copienator import atomic_write_json
+from google import genai
-
-if len(sys.argv) < 2:
-    sys.exit("Usage: python script.py 'InterroTest/Ex 2/Group_1.jpg' OR <InputDir> OR 'file1' 'file2'")
-
-# Parse Arguments
-parser = argparse.ArgumentParser()
-parser.add_argument("paths", nargs="+", help="List of images or directories")
-parser.add_argument("--overwrite", action="store_true",
-                    help="Force redo requests even if output exists")
-parser.add_argument("--limit", type=int, help="limit calls to gemini rpo integer")
-parser.add_argument("--refaire", action="store_true",
-                    help="Redo specific copies/labels defined in refaire.json")
-parser.add_argument("--batch", action="store_true",
-                    help="Generate a JSONL file of requests to send to the Gemini Batch API")
-parser.add_argument("--batch-from", type=str, metavar="LABEL",
-                    help="Do live requests before LABEL, and batch requests from LABEL onwards")
-parser.add_argument("--deal-with-batched", action="store_true",
-                    help="Process a JSONL file containing completed batch results")
-parser.add_argument("--reset", action="store_true",
-                    help="Remove correction.json, revert _old.pdf, delete _new.pdf, then exit")
-args, _ = parser.parse_known_args()
-
-tasks = [] # List of tuples: (filepath_str, label_str)
-results = {}
-
-
-for path_str in args.paths:
-    arg_path = Path(path_str)
-
-    if not arg_path.exists():
-        print(f"Warning: {path_str} not found. Skipping.")
-        continue
-
-    if arg_path.is_file() and arg_path.suffix.lower() == ".jpg":
-        # Handle individual file
-        # Note: assumes structure InterroTest/Ex 2/Group_1.jpg
-        label = arg_path.parent.name
-        INPUT_DIR = arg_path.parent.parent.parent
-        COPIES_DIR = INPUT_DIR / "Copies"
-        GROUPS_DIR = INPUT_DIR / "Par label"
-        tasks.append((str(arg_path), label))
-        if label not in results:
-            results[label] = []
-
-    elif arg_path.is_dir():
-        INPUT_DIR = arg_path
-        COPIES_DIR = INPUT_DIR / "Copies"
-        GROUPS_DIR = INPUT_DIR / "Par label"
-        # Handle directory (original behavior)
-        for sub in GROUPS_DIR.iterdir():
-            if sub.is_dir():
-                label = sub.name
-                if label not in results:
-                    results[label] = []
-                for img in sub.glob("*.jpg"):
-                    tasks.append((str(img), label))
 
+import config
+import grouping
+import prompting
+from copienator import (
+    CliError,
+    EvaluationWorkspace,
+    ExitCode,
+    atomic_write_json,
+    atomic_write_text,
+    execute,
+    read_json,
+    target_parser,
+    workspace_from_target,
+)
+from utils import enonce_total, read_all_labels
 
 NB_THREADS = 12
 
 # PROXY_URL = "http://192.168.241.1:3128"
-PROXY_URL = None
-
-if PROXY_URL:
-    os.environ["http_proxy"] = PROXY_URL
-    os.environ["https_proxy"] = PROXY_URL
-
-import config    
-
 MODEL_ID_pro = config.MODEL_PRO_ID
 MODEL_ID_flash = config.MODEL_FLASH_ID
 api_key = config.API_KEY
 
+# Runtime globals retained while the processing helpers are migrated incrementally.
+INPUT_DIR = Path()
+COPIES_DIR = Path()
+GROUPS_DIR = Path()
+output_path = Path()
+progress_path = Path()
+tasks: list[tuple] = []
+tasks_to_process: list[tuple] = []
+results: dict = {}
+completed_tasks: list = []
+errors_summary: list = []
+overwrite = False
+limit = None
+client = None
+start_time = 0.0
+
 # --- Thread-safe Logging ---
 log_lock = threading.Lock()
 thread_logs = {}
@@ -113,57 +81,6 @@ def flush_thread_log(tid=None):
                 f.write("\n".join(thread_logs[tid]) + "\n\n")
             thread_logs[tid].clear()
 
-def handle_interrupt(sig, frame):
-    """Flush all partial/unfinished logs if program is interrupted."""
-    print("\nInterrupt received. Flushing partial logs...", file=sys.stderr)
-    for tid in list(thread_logs.keys()):
-        flush_thread_log(tid)
-    sys.exit(1)
-
-
-signal.signal(signal.SIGINT, handle_interrupt)
-signal.signal(signal.SIGTERM, handle_interrupt)
-# ---------------------------
-
-client = genai.Client(api_key=api_key)
-output_path = INPUT_DIR / "correction.json"
-progress_path = INPUT_DIR / "correction_progress.json"
-
-if args.reset:
-    print("--- Running Reset ---")
-    if output_path.exists():
-        output_path.unlink()
-        print(f"Deleted: {output_path}")
-    if progress_path.exists():
-        progress_path.unlink()
-        print(f"Deleted: {progress_path}")
-
-    if COPIES_DIR.exists():
-        for copie_dir in COPIES_DIR.iterdir():
-            if not copie_dir.is_dir():
-                continue
-
-            # Revert _old.pdf files
-            for old_pdf in copie_dir.glob("*_old.pdf"):
-                orig_pdf = copie_dir / old_pdf.name.replace("_old.pdf", ".pdf")
-                if orig_pdf.exists():
-                    orig_pdf.unlink() # Prevent FileExistsError on Windows
-                old_pdf.rename(orig_pdf)
-                print(f"Moved: {copie_dir.name}/{old_pdf.name} -> {orig_pdf.name}")
-
-            # Delete _new.pdf files
-            for new_pdf in copie_dir.glob("*_new.pdf"):
-                new_pdf.unlink()
-                print(f"Deleted: {copie_dir.name}/{new_pdf.name}")
-
-    sys.exit("Reset almost complete. For each deleted `_new`, you should manually delete the group in `Par label`")
-
-start_time = time.time()
-overwrite = args.overwrite
-limit = args.limit
-completed_tasks = []
-errors_summary = []
-
 # --- Lock for thread-safe file writing ---
 io_lock = threading.Lock()
 pro_lock = threading.Lock()
@@ -171,21 +88,119 @@ pro_count = 0
 flash_count = 0
 pro_quota_exhausted = False
 
-if overwrite:
-    if output_path.exists():
-        output_path.unlink()
-    if progress_path.exists():
-        progress_path.unlink()
-else:
-    if progress_path.exists():
-        with open(progress_path, "r", encoding="utf-8") as f:
-            completed_tasks = json.load(f)
-    if output_path.exists():
-        with open(output_path, "r", encoding="utf-8") as f:
-            results = json.load(f)
 
-completed_set = set((str(f), l) for f, l in completed_tasks)
+def discover_tasks(
-tasks_to_process = [t for t in tasks if (str(t[0]), t[1]) not in completed_set]
+    workspace: EvaluationWorkspace,
+    targets: list[Path],
+) -> tuple[list[tuple[str, str]], list[str]]:
+    workspace.require_directories("Copies", "Par label")
+    discovered: list[tuple[str, str]] = []
+    warnings: list[str] = []
+    for target in targets:
+        if target.is_file():
+            if target.suffix.casefold() != ".jpg":
+                raise CliError(
+                    f"Correction target is not a group JPG: {target}",
+                    ExitCode.INVALID_ARGUMENTS,
+                )
+            try:
+                target.relative_to(workspace.groups_dir)
+            except ValueError as exc:
+                raise CliError(
+                    f"Group image is not inside {workspace.groups_dir}: {target}",
+                    ExitCode.INVALID_ARGUMENTS,
+                ) from exc
+            discovered.append((str(target), target.parent.name))
+            continue
+
+        group_directories = sorted(
+            (path for path in workspace.groups_dir.iterdir() if path.is_dir()),
+            key=lambda path: path.name.casefold(),
+        )
+        for group_directory in group_directories:
+            images = sorted(
+                group_directory.glob("*.jpg"), key=lambda path: path.name.casefold()
+            )
+            discovered.extend(
+                (str(image), group_directory.name) for image in images
+            )
+        if not group_directories:
+            warnings.append(f"No label groups found in {workspace.groups_dir}")
+    return list(dict.fromkeys(discovered)), warnings
+
+
+def configure_runtime(
+    workspace: EvaluationWorkspace,
+    discovered_tasks: list[tuple[str, str]],
+    args: argparse.Namespace,
+    *,
+    api_client=None,
+) -> None:
+    global INPUT_DIR, COPIES_DIR, GROUPS_DIR, output_path, progress_path
+    global tasks, tasks_to_process, results, completed_tasks, errors_summary
+    global overwrite, limit, client, start_time
+    global pro_count, flash_count, pro_quota_exhausted
+
+    INPUT_DIR = workspace.root
+    COPIES_DIR = workspace.copies_dir
+    GROUPS_DIR = workspace.groups_dir
+    output_path = workspace.correction_file
+    progress_path = workspace.correction_progress_file
+    tasks = list(discovered_tasks)
+    overwrite = bool(args.overwrite)
+    limit = args.limit
+    start_time = time.time()
+    errors_summary = []
+    completed_tasks = []
+    results = {label: [] for _file, label in tasks}
+    thread_logs.clear()
+    pro_count = 0
+    flash_count = 0
+    pro_quota_exhausted = False
+
+    if not overwrite:
+        if progress_path.is_file():
+            loaded_progress = read_json(progress_path)
+            if not isinstance(loaded_progress, list):
+                raise TypeError("correction_progress.json must contain a JSON array")
+            completed_tasks = loaded_progress
+        if output_path.is_file():
+            loaded_results = read_json(output_path)
+            if not isinstance(loaded_results, dict):
+                raise TypeError("correction.json must contain a JSON object")
+            results = loaded_results
+
+    completed_set = {(str(file_path), label) for file_path, label in completed_tasks}
+    tasks_to_process = [
+        task for task in tasks if (str(task[0]), task[1]) not in completed_set
+    ]
+    client = api_client
+
+
+def reset_workspace(workspace: EvaluationWorkspace) -> None:
+    """Apply the explicitly requested correction reset."""
+    print("--- Running Reset ---")
+    for path in (workspace.correction_file, workspace.correction_progress_file):
+        if path.exists():
+            path.unlink()
+            print(f"Deleted: {path}")
+    if workspace.copies_dir.is_dir():
+        for copy_directory in workspace.copies_dir.iterdir():
+            if not copy_directory.is_dir():
+                continue
+            for old_pdf in copy_directory.glob("*_old.pdf"):
+                original = old_pdf.with_name(old_pdf.name.replace("_old.pdf", ".pdf"))
+                if original.exists():
+                    original.unlink()
+                old_pdf.replace(original)
+                print(f"Moved: {copy_directory.name}/{old_pdf.name} -> {original.name}")
+            for new_pdf in copy_directory.glob("*_new.pdf"):
+                new_pdf.unlink()
+                print(f"Deleted: {copy_directory.name}/{new_pdf.name}")
+    print(
+        "Reset almost complete. Manually remove groups associated with deleted "
+        "_new PDFs from 'Par label'."
+    )
 
 def call_gemini_with_retries(model_id, contents, config,
                              fallback_model_id=MODEL_ID_flash):
@@ -267,9 +282,6 @@ def correct_boxes_with_gemini(pid, label, pdf_path, original_feedbacks,
 
     return global_feedbacks + corrected_feedbacks
 
-import shutil
-import grouping
-
 def get_next_group_idx(label):
     """Finds the next available Group index for a given label."""
     target_folder = GROUPS_DIR / label
@@ -278,8 +290,6 @@ def get_next_group_idx(label):
     if not existing: return 0
     return max([int(f.stem.split("_")[1]) for f in existing])
 
-from utils import read_all_labels, enonce_total
-
 def handle_label_errors(pid, label, res, pdf_path):
     """Handles Gemini labeling errors, moves/copies files, and returns new tasks."""
     new_tasks = []
@@ -332,7 +342,7 @@ def handle_label_errors(pid, label, res, pdf_path):
         tprint(f"\tHandling additional-answer for {pid} {label}")
         try:
             add_labels = json.loads(call_gemini_with_retries(MODEL_ID_flash, contents, config))
-        except Exception:
+        except Exception:  # noqa: BLE001 - invalid auxiliary model response
             add_labels = []
 
         keep_error = False
@@ -374,7 +384,7 @@ def handle_label_errors(pid, label, res, pdf_path):
 
 def process_single_task(task_tuple, precomputed_response=None):
     try:
-        global pro_count, flash_count, pro_quota_exhausted
+        global pro_count, flash_count
         file_path = task_tuple[0]
         label = task_tuple[1]
         can_spawn_tasks = task_tuple[2] if len(task_tuple) > 2 else True
@@ -383,8 +393,7 @@ def process_single_task(task_tuple, precomputed_response=None):
         json_path = group_name + '.json'
         new_tasks = []
 
-        with open(json_path, 'r') as f:
+        group_data = read_json(json_path)
-            group_data = json.load(f)
 
         n = len(group_data)
         d_data = {l[0]: (l[1], l[2], l[3]) for l in group_data}
@@ -466,7 +475,7 @@ def process_single_task(task_tuple, precomputed_response=None):
                 for (i,f) in enumerate(res["feedback"]):
                     b = f.get("box_2d")
                     if b:
-                        ymin, xmin, ymax, xmax = b
+                        ymin, _xmin, ymax, xmax = b
                         ymin = ymin * total_height // 1000
                         ymax = ymax * total_height // 1000
 
@@ -493,7 +502,7 @@ def process_single_task(task_tuple, precomputed_response=None):
                         res["feedback"] = correct_boxes_with_gemini(
                             pid, label, pdf_path, res["feedback"],
                             yming, ymaxg, width_r, total_height)
-                    except Exception as e:
+                    except Exception as e:  # noqa: BLE001 - correction fallback
                         tprint(f"\tCorrection failed for Copie {pid}, {group_name} : {e}\n\tRemoving the boxes")
                         # Fallback if the second request fails entirely
                         for (i, f) in enumerate(res["feedback"]):
@@ -516,7 +525,7 @@ def process_single_task(task_tuple, precomputed_response=None):
             tprint(f"Error decoding JSON for {file_path}", file=sys.stderr)
             with io_lock:
                 errors_summary.append(("Error decoding JSON response", file_path))
-        except Exception as e:
+        except Exception as e:  # noqa: BLE001 - per-task processing boundary
             error_msg = f"Exception processing {file_path}: {e}"
             print(error_msg, file=sys.stderr)
             with io_lock:
@@ -602,19 +611,20 @@ def resolve_delayed_moves():
 
     return new_tasks
 
-if __name__ == "__main__":
+def run_configured(args: argparse.Namespace) -> ExitCode:
+    global client, tasks_to_process
+    if client is None:
+        client = genai.Client(api_key=api_key)
     if args.refaire:
         refaire_path = INPUT_DIR / "refaire.json"
         overwritten_path = INPUT_DIR / "overwritten_correction.json"
 
         if refaire_path.exists():
-            with open(refaire_path, "r", encoding="utf-8") as f:
+            refaire_list = read_json(refaire_path)
-                refaire_list = json.load(f)
 
             overwritten_data = []
             if overwritten_path.exists():
-                with open(overwritten_path, "r", encoding="utf-8") as f:
+                overwritten_data = read_json(overwritten_path)
-                    overwritten_data = json.load(f)
 
             dirty_results = False
 
@@ -659,10 +669,12 @@ if __name__ == "__main__":
                     pdf_path = copie_dir / f"{label}.pdf"
                     is_new = False
 
-                    if not pdf_path.exists():
+                    if (
-                        if (copie_dir / f"{label}_new.pdf").exists():
+                        not pdf_path.exists()
-                            pdf_path = copie_dir / f"{label}_new.pdf"
+                        and (copie_dir / f"{label}_new.pdf").exists()
-                            is_new = True
+                    ):
+                        pdf_path = copie_dir / f"{label}_new.pdf"
+                        is_new = True
                         # elif (copie_dir / f"{label}_old.pdf").exists():
                             # pdf_path = copie_dir / f"{label}_old.pdf"
 
@@ -690,7 +702,11 @@ if __name__ == "__main__":
                     input(f"About to batch from: {args.batch_from}. Press Enter to confirm...")
                     break
             if args.batch_from not in all_labels:
-                sys.exit(f"Error: Label '{args.batch_from}' not found. Available labels: {all_labels}")
+                raise CliError(
+                    f"Label '{args.batch_from}' not found. Available labels: "
+                    f"{all_labels}",
+                    ExitCode.INVALID_ARGUMENTS,
+                )
 
             target_idx = all_labels.index(args.batch_from)
             live_tasks = []
@@ -714,58 +730,66 @@ if __name__ == "__main__":
 
             count_flash = 0
             count_pro = 0
-
+            flash_lines = []
-            with open(batch_flash_file, "w", encoding="utf-8") as f_flash, \
+            pro_lines = []
-                 open(batch_pro_file, "w", encoding="utf-8") as f_pro:
+            for task in batch_tasks:
-
+                file_path, label = task[0], task[1]
-                for task in batch_tasks:
+                json_path = Path(file_path).with_suffix(".json")
-                    file_path, label = task[0], task[1]
+                group_data = read_json(json_path)
-                    group_name = os.path.splitext(file_path)[0]
+                use_flash = len(group_data) >= 4 or group_data[-1][2] <= 500
-                    json_path = group_name + '.json'
+                b64_img = base64.b64encode(Path(file_path).read_bytes()).decode(
-
+                    "utf-8"
-                    with open(json_path, 'r') as jf:
+                )
-                        group_data = json.load(jf)
+                request = {
-                    use_flash = len(group_data) >= 4 or group_data[-1][2] <= 500
+                    "key": file_path,
-
+                    "request": {
-                    image_data = Path(file_path).read_bytes()
+                        "contents": [
-                    b64_img = base64.b64encode(image_data).decode("utf-8")
+                            {
-
-                    # Format payload matching Gemini Batch API file requirements
-                    req = {
-                        "key": file_path,  # The ID returned in the output file
-                        "request": {
-                            "contents": [{
                                 "role": "user",
                                 "parts": [
-                                    {"inlineData": {"mimeType": "image/jpeg", "data": b64_img}},
+                                    {
-                                    {"text": prompting.make_prompt(INPUT_DIR,label)}
+                                        "inlineData": {
-                                ]
+                                            "mimeType": "image/jpeg",
-                            }],
+                                            "data": b64_img,
-                            "generation_config": {
+                                        }
-                                "temperature": 1.0,
+                                    },
-                                "topP": 0.95,
+                                    {"text": prompting.make_prompt(INPUT_DIR, label)},
-                                "maxOutputTokens": 65535,
+                                ],
-                                "responseMimeType": "application/json",
-                                "responseSchema": prompting.UNROLLED_SCHEMA
                             }
-                        }
+                        ],
-                    }
+                        "generation_config": {
+                            "temperature": 1.0,
+                            "topP": 0.95,
+                            "maxOutputTokens": 65535,
+                            "responseMimeType": "application/json",
+                            "responseSchema": prompting.UNROLLED_SCHEMA,
+                        },
+                    },
+                }
+                line = json.dumps(request)
+                if use_flash:
+                    flash_lines.append(line)
+                    count_flash += 1
+                else:
+                    pro_lines.append(line)
+                    count_pro += 1
+            atomic_write_text(
+                batch_flash_file,
+                "\n".join(flash_lines) + ("\n" if flash_lines else ""),
+            )
+            atomic_write_text(
+                batch_pro_file,
+                "\n".join(pro_lines) + ("\n" if pro_lines else ""),
+            )
 
-                    if use_flash:
+            print("Batch generation complete.")
-                        f_flash.write(json.dumps(req) + "\n")
-                        count_flash += 1
-                    else:
-                        f_pro.write(json.dumps(req) + "\n")
-                        count_pro += 1
-
-            print(f"Batch generation complete.")
             print(f" - {count_flash} requests saved to {batch_flash_file} (for {MODEL_ID_flash})")
             print(f" - {count_pro} requests saved to {batch_pro_file} (for {MODEL_ID_pro})")
             print("Upload these files via the File API and create two separate batch jobs.")
 
         # If there's no live tasks to do, and we aren't doing a batched ingestion, exit right away
         if not tasks_to_process and not args.deal_with_batched:
-            sys.exit(0)
+            return ExitCode.SUCCESS
 
     batched_responses = {}
     if args.deal_with_batched:
@@ -807,8 +831,11 @@ if __name__ == "__main__":
                         if new_generated_tasks:
                             for new_task in new_generated_tasks:
                                 futures[executor.submit(process_single_task, new_task)] = new_task
-                    except Exception as e:
+                    except Exception as e:  # noqa: BLE001 - future boundary
                         print(f"Exception during task execution: {e}", file=sys.stderr)
+                        failed_task = futures[future]
+                        with io_lock:
+                            errors_summary.append((str(e), failed_task[0]))
 
             tasks_to_process = [] # Vider la liste une fois traitée
 
@@ -840,9 +867,12 @@ if __name__ == "__main__":
 
     if unresolved_delayed:
         manual_path = INPUT_DIR / "manual_resolutions.txt"
-        with open(manual_path, "w", encoding="utf-8") as f:
+        atomic_write_text(
-            f.write("###  Use -> x>, -x, ss, sx, xx, xs\n")
+            manual_path,
-            f.write("\n".join(unresolved_delayed) + "\n")
+            "###  Use -> x>, -x, ss, sx, xx, xs\n"
+            + "\n".join(unresolved_delayed)
+            + "\n",
+        )
         print(f"\n[!] Unresolved delayed tasks found! Wrote to {manual_path}.")
         print("    Please edit it manually, then run `python resolve_manual.py <InputDir>`")
 
@@ -855,3 +885,101 @@ if __name__ == "__main__":
             print(err, file=sys.stderr)
             escaped_path = shlex.quote(str(file))
             print(f"Run :    python correction.py {escaped_path}")
+    return ExitCode.PARTIAL if errors_summary else ExitCode.SUCCESS
+
+
+def run(
+    workspace: EvaluationWorkspace,
+    targets: list[Path],
+    args: argparse.Namespace,
+    *,
+    api_client=None,
+) -> ExitCode:
+    if args.reset:
+        workspace.require_directories("Copies")
+        reset_workspace(workspace)
+        return ExitCode.SUCCESS
+    workspace.require_directories("Copies", "Par label")
+    workspace.require_files("labels")
+    if args.refaire:
+        workspace.require_files("refaire.json")
+    discovered, warnings = discover_tasks(workspace, targets)
+    for warning in warnings:
+        print(f"Warning: {warning}")
+    configure_runtime(workspace, discovered, args, api_client=api_client)
+    if not discovered and not args.refaire:
+        return ExitCode.PARTIAL
+    try:
+        status = run_configured(args)
+    finally:
+        for thread_id in list(thread_logs):
+            flush_thread_log(thread_id)
+    if warnings and status == ExitCode.SUCCESS:
+        return ExitCode.PARTIAL
+    return status
+
+
+def build_parser() -> argparse.ArgumentParser:
+    parser = target_parser("Correct grouped answers with Gemini")
+    parser.add_argument(
+        "additional_targets",
+        nargs="*",
+        type=Path,
+        help="Additional group JPG files from the same evaluation",
+    )
+    parser.add_argument("--overwrite", action="store_true", help="Redo requests")
+    parser.add_argument("--limit", type=int, help="Maximum Gemini Pro calls")
+    parser.add_argument(
+        "--refaire",
+        action="store_true",
+        help="Redo copies and labels listed in refaire.json",
+    )
+    parser.add_argument(
+        "--batch",
+        action="store_true",
+        help="Generate Gemini batch request JSONL files",
+    )
+    parser.add_argument(
+        "--batch-from",
+        metavar="LABEL",
+        help="Process earlier labels live and batch from LABEL onward",
+    )
+    parser.add_argument(
+        "--deal-with-batched",
+        action="store_true",
+        help="Consume batched_correction_result.jsonl",
+    )
+    parser.add_argument(
+        "--reset",
+        action="store_true",
+        help="Delete correction state, restore _old PDFs, and delete _new PDFs",
+    )
+    return parser
+
+
+def main(argv: Sequence[str] | None = None) -> int:
+    parser = build_parser()
+
+    def handle(args: argparse.Namespace) -> ExitCode:
+        workspace, target = workspace_from_target(args)
+        targets = [target]
+        for additional in args.additional_targets:
+            resolved = additional.expanduser().resolve()
+            if not resolved.exists():
+                raise CliError(
+                    f"Target does not exist: {resolved}",
+                    ExitCode.INVALID_WORKSPACE,
+                )
+            if EvaluationWorkspace.discover(resolved).root != workspace.root:
+                raise CliError(
+                    "All targets must belong to the same evaluation",
+                    ExitCode.INVALID_ARGUMENTS,
+                )
+            targets.append(resolved)
+        return run(workspace, targets, args)
+
+    return execute(parser, argv, handle)
+
+
+if __name__ == "__main__":
+    raise SystemExit(main())

fetch_batched_results.py

@@ -1,63 +1,83 @@
-import os
+from __future__ import annotations
-import sys
+
 import argparse
-from pathlib import Path
+from collections.abc import Sequence
+
 from google import genai
 
-def main():
+import config
-    parser = argparse.ArgumentParser(description="Download and combine completed batch jobs for a directory.")
+from copienator import (
-    parser.add_argument("root_dir", type=str, help="Directory containing the original batches")
+    CliError,
-    args = parser.parse_args()
+    EvaluationWorkspace,
+    ExitCode,
+    atomic_write_bytes,
+    evaluation_parser,
+    execute,
+    read_json,
+    workspace_from_args,
+)
 
-    target_dir = Path(args.root_dir)
-    dir_name = target_dir.name
-    output_path = target_dir / "batched_correction_result.jsonl"
 
-    if "GEMINI_API_KEY" not in os.environ:
+def run(workspace: EvaluationWorkspace, *, client=None) -> ExitCode:
-        sys.exit("Error: GEMINI_API_KEY environment variable not set.")
+    if client is None:
+        if not config.API_KEY:
+            raise CliError("GEMINI_API_KEY is not configured")
+        client = genai.Client(api_key=config.API_KEY)
+    matching = []
+    if workspace.batch_jobs_file.is_file():
+        manifest = read_json(workspace.batch_jobs_file)
+        jobs = manifest.get("jobs") if isinstance(manifest, dict) else None
+        if not isinstance(jobs, dict):
+            raise CliError(f"Invalid batch manifest: {workspace.batch_jobs_file}")
+        matching = [
+            client.batches.get(name=entry["name"])
+            for entry in jobs.values()
+            if isinstance(entry, dict) and isinstance(entry.get("name"), str)
+        ]
+    else:
+        matching = [
+            job
+            for job in client.batches.list()
+            if workspace.name in str(getattr(job, "display_name", ""))
+        ]
+    if not matching:
+        raise CliError(
+            f"No batch jobs found for evaluation {workspace.name!r}"
+        )
+    for job in matching:
+        state = job.state.name if hasattr(job.state, "name") else job.state
+        print(f"{job.display_name}: {state}")
+        if state != "JOB_STATE_SUCCEEDED":
+            print("Not all matching jobs have succeeded yet.")
+            return ExitCode.PARTIAL
 
-    client = genai.Client()
+    chunks = []
+    incomplete = False
+    for job in matching:
+        destination = getattr(job, "dest", None)
+        file_name = getattr(destination, "file_name", None)
+        if not file_name:
+            print(f"Warning: {job.display_name} has no output file.")
+            incomplete = True
+            continue
+        payload = client.files.download(file=file_name)
+        chunks.append(payload.rstrip(b"\n"))
+    if not chunks:
+        return ExitCode.PARTIAL
+    output_path = workspace.batched_correction_result_file
+    atomic_write_bytes(output_path, b"\n".join(chunks) + b"\n")
+    print(f"Saved combined batch results to {output_path}")
+    return ExitCode.PARTIAL if incomplete else ExitCode.SUCCESS
 
-    print(f"Fetching jobs matching '{dir_name}'...")
-    all_jobs = client.batches.list()
-    matching_jobs = []
 
-    # 1. Find jobs associated with this directory
+def build_parser() -> argparse.ArgumentParser:
-    for job in all_jobs:
+    return evaluation_parser("Download and combine correction batch results")
-        if hasattr(job, 'display_name') and job.display_name and dir_name in job.display_name:
-            matching_jobs.append(job)
 
-    if not matching_jobs:
-        sys.exit(f"No batch jobs found containing '{dir_name}' in their display name.")
 
-    # 2. Check that all matching jobs are complete
+def main(argv: Sequence[str] | None = None) -> int:
-    for job in matching_jobs:
+    parser = build_parser()
-        state = job.state.name if hasattr(job.state, 'name') else job.state
+    return execute(parser, argv, lambda args: run(workspace_from_args(args)))
-        print(f"Found Job: {job.display_name} | State: {state}")
-        if state != 'JOB_STATE_SUCCEEDED':
-            sys.exit(f"Error: Job '{job.display_name}' has not succeeded yet. Try again later.")
 
-    # 3. Download and concatenate
-    print("\nAll jobs succeeded. Downloading results...")
-    combined_data = b""
-
-    for job in matching_jobs:
-        if hasattr(job, 'dest') and job.dest and hasattr(job.dest, 'file_name') and job.dest.file_name:
-            print(f"Downloading output for {job.display_name}...")
-            file_content_bytes = client.files.download(file=job.dest.file_name)
-
-            combined_data += file_content_bytes
-            # Ensure proper line separation between files in JSONL
-            if combined_data and not combined_data.endswith(b'\n'):
-                combined_data += b'\n'
-        else:
-            print(f"Warning: Job {job.display_name} succeeded but has no output file.")
-
-    # 4. Save to destination
-    with open(output_path, "wb") as f:
-        f.write(combined_data)
-
-    print(f"\nSuccess! All results concatenated and saved to:\n{output_path}")
 
 if __name__ == "__main__":
-    main()
+    raise SystemExit(main())

submit_batches.py

@@ -1,77 +1,96 @@
-import os
+from __future__ import annotations
-import sys
+
 import argparse
-from pathlib import Path
+from collections.abc import Sequence
+
 from google import genai
 from google.genai import types
 
-def main():
+import config
-    parser = argparse.ArgumentParser(description="Upload JSONL files and create Gemini Batch jobs.")
+from copienator import (
-    parser.add_argument("root_dir", type=str, help="Root directory containing the batch JSONL files")
+    CliError,
-    args = parser.parse_args()
+    EvaluationWorkspace,
+    ExitCode,
+    atomic_write_json,
+    evaluation_parser,
+    execute,
+    read_json,
+    workspace_from_args,
+)
 
-    root_dir = Path(args.root_dir)
 
-    if "GEMINI_API_KEY" not in os.environ:
+def run(workspace: EvaluationWorkspace, *, client=None) -> ExitCode:
-        sys.exit("Error: GEMINI_API_KEY environment variable not set.")
+    if client is None:
-
+        if not config.API_KEY:
-    client = genai.Client()
+            raise CliError("GEMINI_API_KEY is not configured")
-
+        client = genai.Client(api_key=config.API_KEY)
-    # Define the batch files and their corresponding models
+    batches = (
-    batches_to_create = [
+        (
-        {
+            "flash",
-            "file_path": root_dir / "batch_requests_flash.jsonl",
+            workspace.root / "batch_requests_flash.jsonl",
-            "model_id": "gemini-3-flash-preview",
+            config.MODEL_FLASH_ID,
-            "display_name": f"flash-correction-{root_dir.name}"
+            f"flash-correction-{workspace.name}",
-        },
+        ),
-        {
+        (
-            "file_path": root_dir / "batch_requests_pro.jsonl",
+            "pro",
-            "model_id": "gemini-3.1-pro-preview",
+            workspace.root / "batch_requests_pro.jsonl",
-            "display_name": f"pro-correction-{root_dir.name}"
+            config.MODEL_PRO_ID,
-        }
+            f"pro-correction-{workspace.name}",
-    ]
+        ),
-
+    )
-    for batch in batches_to_create:
+    manifest = {
-        file_path = batch["file_path"]
+        "version": 1,
-        model_id = batch["model_id"]
+        "evaluation": workspace.name,
-        display_name = batch["display_name"]
+        "jobs": {},
-
+    }
-        # Check if the file exists
+    if workspace.batch_jobs_file.is_file():
-        if not file_path.exists():
+        previous = read_json(workspace.batch_jobs_file)
+        if isinstance(previous, dict) and isinstance(previous.get("jobs"), dict):
+            manifest["jobs"] = previous["jobs"]
+    started = 0
+    for tier, file_path, model_id, display_name in batches:
+        if not file_path.is_file():
             print(f"Skipping {model_id}: {file_path.name} does not exist.")
             continue
-
-        # Check if the file is empty (e.g., if all tasks went to Flash, Pro might be empty)
         if file_path.stat().st_size == 0:
             print(f"Skipping {model_id}: {file_path.name} is empty.")
             continue
-
+        print(f"Uploading {file_path.name} for model {model_id}...")
-        print(f"Processing {file_path.name} for model {model_id}...")
+        uploaded = client.files.upload(
-
-        # 1. Upload the file to the File API
-        print(f"  Uploading file...")
-        uploaded_file = client.files.upload(
             file=str(file_path),
             config=types.UploadFileConfig(
                 display_name=f"{display_name}-input",
-                mime_type='jsonl'
+                mime_type="jsonl",
-            )
+            ),
         )
-        print(f"  Uploaded successfully! File ID: {uploaded_file.name}")
+        job = client.batches.create(
-
-        # 2. Create the batch job
-        print(f"  Starting batch job...")
-        batch_job = client.batches.create(
             model=model_id,
-            src=uploaded_file.name,
+            src=uploaded.name,
-            config={
+            config={"display_name": display_name},
-                'display_name': display_name,
-            },
         )
-        print(f"  Success! Batch Job Name: {batch_job.name}\n")
+        started += 1
+        manifest["jobs"][tier] = {
+            "name": job.name,
+            "display_name": display_name,
+            "model": model_id,
+            "request_file": file_path.name,
+        }
+        atomic_write_json(workspace.batch_jobs_file, manifest)
+        print(f"Started batch job: {job.name}")
+    if not started:
+        print("No non-empty batch request files were found.")
+        return ExitCode.PARTIAL
+    return ExitCode.SUCCESS
+
+
+def build_parser() -> argparse.ArgumentParser:
+    return evaluation_parser("Upload correction JSONL files and start Gemini batches")
+
+
+def main(argv: Sequence[str] | None = None) -> int:
+    parser = build_parser()
+    return execute(parser, argv, lambda args: run(workspace_from_args(args)))
 
-    print("-" * 50)
-    print("All batch jobs have been initiated.")
 
 if __name__ == "__main__":
-    main()
+    raise SystemExit(main())

tests/test_gui_core.py

@@ -2,6 +2,7 @@ from __future__ import annotations
 
 import importlib.util
 import io
+import json
 import os
 import queue
 import sys
@@ -11,6 +12,7 @@ import unittest
 from concurrent.futures import ThreadPoolExecutor
 from contextlib import redirect_stderr
 from pathlib import Path
+from types import SimpleNamespace
 from unittest.mock import Mock, patch
 
 from PIL import Image
@@ -21,6 +23,7 @@ from copienator import (
     WorkspaceNotFoundError,
     WorkspaceValidationError,
     atomic_update_json,
+    atomic_write_bytes,
     atomic_write_json,
     read_json,
     workspace_from_target,
@@ -114,6 +117,12 @@ class WorkspaceTests(unittest.TestCase):
 
 
 class AtomicJsonTests(unittest.TestCase):
+    def test_atomic_binary_round_trip(self) -> None:
+        with tempfile.TemporaryDirectory() as directory:
+            path = Path(directory) / "result.jsonl"
+            atomic_write_bytes(path, b'{"one":1}\n')
+            self.assertEqual(path.read_bytes(), b'{"one":1}\n')
+
     def test_atomic_round_trip_and_unicode(self) -> None:
         with tempfile.TemporaryDirectory() as directory:
             path = Path(directory) / "nested" / "state.json"
@@ -313,6 +322,14 @@ class StandardCliTests(unittest.TestCase):
             "gemini_for_labels": load_script_module(
                 "gemini_for_labels.py", "gemini_for_labels"
             ),
+            "correction": load_script_module("correction.py", "correction"),
+            "submit_batches": load_script_module(
+                "submit_batches.py", "submit_batches"
+            ),
+            "batch_status": load_script_module("batch_status.py", "batch_status"),
+            "fetch_batched_results": load_script_module(
+                "fetch_batched_results.py", "fetch_batched_results"
+            ),
             "copies_tools": load_script_module(
                 "copies_tools.py", "copienator_copies_tools_test"
             ),
@@ -355,6 +372,9 @@ class StandardCliTests(unittest.TestCase):
                 "page_splitter": [missing],
                 "plotting": [missing],
                 "gemini_for_labels": [missing],
+                "correction": [missing],
+                "submit_batches": [missing],
+                "fetch_batched_results": [missing],
             }
             for name, arguments in invocations.items():
                 with self.subTest(script=name), redirect_stderr(io.StringIO()):
@@ -472,6 +492,21 @@ class StandardCliTests(unittest.TestCase):
                 "default",
                 {"target": evaluation, "overwrite": True},
             ),
+            "correction": (
+                "correction",
+                "live",
+                {"target": evaluation, "overwrite": True, "limit": 5},
+            ),
+            "submit_batches": (
+                "submit_batches",
+                "default",
+                {"target": evaluation},
+            ),
+            "fetch_batched_results": (
+                "fetch_batches",
+                "default",
+                {"target": evaluation},
+            ),
         }
         for module_name, (step_id, variant_id, values) in cases.items():
             step = steps[step_id]
@@ -900,6 +935,135 @@ class StandardCliTests(unittest.TestCase):
                 "Ex 1",
             )
 
+    def test_correction_overwrite_keeps_previous_state_until_a_commit(self) -> None:
+        module = self.modules["correction"]
+        with tempfile.TemporaryDirectory() as directory:
+            evaluation = Path(directory) / "Exam"
+            evaluation.mkdir()
+            correction = evaluation / "correction.json"
+            progress = evaluation / "correction_progress.json"
+            atomic_write_json(correction, {"Ex 1": [[{"id": "01"}]]})
+            atomic_write_json(progress, [["old.jpg", "Ex 1"]])
+            args = module.build_parser().parse_args(
+                [str(evaluation), "--overwrite"]
+            )
+            module.configure_runtime(
+                EvaluationWorkspace(evaluation),
+                [("new.jpg", "Ex 1")],
+                args,
+                api_client=Mock(),
+            )
+            self.assertEqual(read_json(correction), {"Ex 1": [[{"id": "01"}]]})
+            self.assertEqual(read_json(progress), [["old.jpg", "Ex 1"]])
+            self.assertEqual(module.results, {"Ex 1": []})
+            self.assertEqual(module.tasks_to_process, [("new.jpg", "Ex 1")])
+
+    def test_correction_reset_restores_old_and_deletes_new_files(self) -> None:
+        module = self.modules["correction"]
+        with tempfile.TemporaryDirectory() as directory:
+            evaluation = Path(directory) / "Exam"
+            copy_dir = evaluation / "Copies" / "Copie01"
+            copy_dir.mkdir(parents=True)
+            atomic_write_json(evaluation / "correction.json", {"old": True})
+            atomic_write_json(evaluation / "correction_progress.json", ["old"])
+            (copy_dir / "Ex 1.pdf").write_bytes(b"current")
+            (copy_dir / "Ex 1_old.pdf").write_bytes(b"original")
+            (copy_dir / "Ex 2_new.pdf").write_bytes(b"generated")
+
+            self.assertEqual(module.main([str(evaluation), "--reset"]), 0)
+            self.assertFalse((evaluation / "correction.json").exists())
+            self.assertFalse((evaluation / "correction_progress.json").exists())
+            self.assertEqual((copy_dir / "Ex 1.pdf").read_bytes(), b"original")
+            self.assertFalse((copy_dir / "Ex 1_old.pdf").exists())
+            self.assertFalse((copy_dir / "Ex 2_new.pdf").exists())
+
+    def test_correction_batch_request_files_are_written_atomically(self) -> None:
+        module = self.modules["correction"]
+        with tempfile.TemporaryDirectory() as directory:
+            evaluation = Path(directory) / "Exam"
+            group_dir = evaluation / "Par label" / "Ex 1"
+            (evaluation / "Copies").mkdir(parents=True)
+            group_dir.mkdir(parents=True)
+            (evaluation / "labels").write_text("Ex 1\n", encoding="utf-8")
+            image = group_dir / "Group_1.jpg"
+            image.write_bytes(b"image")
+            atomic_write_json(
+                image.with_suffix(".json"),
+                [["01", 0, 400, 1.0, "Ex 1"]],
+            )
+            args = module.build_parser().parse_args([str(evaluation), "--batch"])
+            module.configure_runtime(
+                EvaluationWorkspace(evaluation),
+                [(str(image), "Ex 1")],
+                args,
+                api_client=Mock(),
+            )
+            with patch.object(module.prompting, "make_prompt", return_value="prompt"):
+                self.assertEqual(module.run_configured(args), 0)
+            lines = (evaluation / "batch_requests_flash.jsonl").read_text().splitlines()
+            self.assertEqual(len(lines), 1)
+            self.assertEqual(json.loads(lines[0])["key"], str(image))
+            self.assertEqual(
+                (evaluation / "batch_requests_pro.jsonl").read_text(), ""
+            )
+
+    def test_batch_helpers_use_mocked_api_and_atomic_combination(self) -> None:
+        submit = self.modules["submit_batches"]
+        fetch = self.modules["fetch_batched_results"]
+        status = self.modules["batch_status"]
+        with tempfile.TemporaryDirectory() as directory:
+            evaluation = Path(directory) / "Exam"
+            evaluation.mkdir()
+            (evaluation / "batch_requests_flash.jsonl").write_text(
+                "{}\n", encoding="utf-8"
+            )
+            client = Mock()
+            client.files.upload.return_value = SimpleNamespace(name="files/input")
+            client.batches.create.return_value = SimpleNamespace(name="batches/1")
+            self.assertEqual(
+                submit.run(EvaluationWorkspace(evaluation), client=client), 0
+            )
+            self.assertEqual(client.batches.create.call_count, 1)
+            manifest = json.loads((evaluation / "batch_jobs.json").read_text())
+            self.assertEqual(manifest["jobs"]["flash"]["name"], "batches/1")
+
+            jobs = [
+                SimpleNamespace(
+                    name="batches/1",
+                    display_name=f"flash-correction-{evaluation.name}",
+                    state=SimpleNamespace(name="JOB_STATE_SUCCEEDED"),
+                    dest=SimpleNamespace(file_name="files/flash-result"),
+                ),
+                SimpleNamespace(
+                    name="batches/2",
+                    display_name=f"pro-correction-{evaluation.name}",
+                    state=SimpleNamespace(name="JOB_STATE_SUCCEEDED"),
+                    dest=SimpleNamespace(file_name="files/pro-result"),
+                ),
+            ]
+            client.batches.get.return_value = jobs[0]
+            client.files.download.side_effect = [b'{"flash":1}\n', b'{"pro":1}']
+            self.assertEqual(
+                fetch.run(EvaluationWorkspace(evaluation), client=client), 0
+            )
+            self.assertEqual(
+                (evaluation / "batched_correction_result.jsonl").read_bytes(),
+                b'{"flash":1}\n',
+            )
+
+            job = SimpleNamespace(
+                state=SimpleNamespace(name="JOB_STATE_SUCCEEDED"),
+                dest=SimpleNamespace(file_name="files/result"),
+            )
+            client.batches.get.return_value = job
+            client.files.download.side_effect = None
+            client.files.download.return_value = b'{"downloaded":true}\n'
+            output = evaluation / "one-result.jsonl"
+            self.assertEqual(
+                status.download_job("batches/1", output=output, client=client), 0
+            )
+            self.assertEqual(output.read_bytes(), b'{"downloaded":true}\n')
+
     def test_post_correction_main_cleans_json_atomically(self) -> None:
         module = self.modules["post_correction"]
         with tempfile.TemporaryDirectory() as directory: