Rudimentary support for --refaire

reading_grouped_annotations.py

@@ -5,6 +5,7 @@ import collections
 import concurrent.futures
 from pathlib import Path
 from PIL import Image
+import threading
 
 import annotating
 
@@ -179,7 +180,7 @@ import argparse
 if __name__ == "__main__":
     parser = argparse.ArgumentParser(description="Read grouped annotations and compile PDFs")
     parser.add_argument("input_path", help="Directory path")
-    parser.add_argument("--with-refaire", action="store_true", help="Merge refaire annotations from Bnot")
+    parser.add_argument("--refaire", action="store_true", help="Merge refaire annotations from Bnot")
     args = parser.parse_args()
 
     root_dir = sys.argv[1]
@@ -194,26 +195,52 @@ if __name__ == "__main__":
     except FileNotFoundError:
         all_labels = []
 
+    refaire_dict = {}
+    if args.refaire:
+        refaire_path = os.path.join(root_dir, "refaire.json")
+        if os.path.exists(refaire_path):
+            with open(refaire_path, "r", encoding="utf-8") as f:
+                refaire_list = json.load(f)
+            for c_name, labels in refaire_list:
+                sid = c_name.replace("Copie", "")
+                refaire_dict[sid] = labels
+        else:
+            print(f"Warning: --refaire flag used, but {refaire_path} not found.")
+
+
     # Load original data
-    original_data = annotating.make_dictionary(root_dir)
+    if args.refaire and refaire_list:
+        original_data = annotating.make_dictionary(root_dir,
+                                                   refaire=True,
+                                                   refaire_list=refaire_list)
+    else:
+        original_data = annotating.make_dictionary(root_dir)
 
     lock = threading.Lock()
     actions_by_student = collections.defaultdict(list)
     notes_by_student = collections.defaultdict(dict)
 
 
-    def process_bgnot_entry(entry):
+    def process_bgnot_entry(entry, only_ids=None):
         gdir = os.path.join(bgnot_dir, entry)
         if not os.path.isdir(gdir) or entry.startswith("Copie"):
             return
+        bnote_path = os.path.join(gdir, "bnote.json")
+        with open(bnote_path, "r") as f:
+            bnote_data = json.load(f)
+
+        if only_ids:
+            id_found = False
+            for d in bnote_data["images"]:
+                if d["id"] in only_ids:
+                    id_found = True
+            if not id_found:
+                return
 
         actions, notes_img = detect_checks_and_notes(gdir)
-        bnote_path = os.path.join(gdir, "bnote.json")
         if not os.path.exists(bnote_path) or notes_img is None:
             return
 
-        with open(bnote_path, "r") as f:
-            bnote_data = json.load(f)
 
         with lock:
             for act in actions:
@@ -230,13 +257,16 @@ if __name__ == "__main__":
 
 
     def process_refaire_entry(sid, r_labels):
-        s_bnot_dir = os.path.join(root_dir, "Bnot", f"Copie{sid}")
+        s_bnot_dir = os.path.join(root_dir, "BRnot", f"Copie{sid}")
         if not os.path.exists(s_bnot_dir): return
-        if not r_labels: r_labels = list(original_data.get(sid, {}).keys())
+        if not r_labels:
+            r_labels = list(original_data.get(sid, {}).keys())
 
         with lock:
-            actions_by_student[sid] = [a for a in actions_by_student[sid] if a.get('label') not in r_labels]
-            for lbl in r_labels: notes_by_student[sid].pop(lbl, None)
+            actions_by_student[sid] = [a for a in actions_by_student[sid]
+                                       if a.get('label') not in r_labels]
+            for lbl in r_labels:
+                notes_by_student[sid].pop(lbl, None)
 
         b_actions, b_notes_img = detect_checks_and_notes(s_bnot_dir)
         b_bnote_path = os.path.join(s_bnot_dir, "bnote.json")
@@ -259,24 +289,20 @@ if __name__ == "__main__":
 
 
     # --- 0. Read refaire.json if requested ---
-    refaire_dict = {}
-    if args.with_refaire:
-        refaire_path = os.path.join(root_dir, "refaire.json")
-        if os.path.exists(refaire_path):
-            with open(refaire_path, "r", encoding="utf-8") as f:
-                refaire_list = json.load(f)
-            for c_name, labels in refaire_list:
-                sid = c_name.replace("Copie", "")
-                refaire_dict[sid] = labels
-        else:
-            print(f"Warning: --with-refaire flag used, but {refaire_path} not found.")
 
-    # Part 1 : lecture des bgnot
-    with concurrent.futures.ThreadPoolExecutor(max_workers=4) as executor:
-        executor.map(process_bgnot_entry, os.listdir(bgnot_dir))
+    if refaire_dict:
+        only_ids = [ids for ids in refaire_dict]
+    else:
+        only_ids = None
 
-    # Part 1.5: Refaire
-    if args.with_refaire and refaire_dict:
+
+    # Lecture des bgnot
+    with concurrent.futures.ThreadPoolExecutor(max_workers=6) as executor:
+        executor.map(lambda x: process_bgnot_entry(x, only_ids=only_ids),
+                     os.listdir(bgnot_dir))
+
+    # Refaire
+    if args.refaire and refaire_dict:
         for sid, labels in refaire_dict.items():
             process_refaire_entry(sid, labels)
 
@@ -296,7 +322,11 @@ if __name__ == "__main__":
     # --- 2. Process each student concurrently using 4 threads ---
     sids = sorted(original_data.keys(), key=natural_key)
     with concurrent.futures.ThreadPoolExecutor(max_workers=4) as executor:
-        futures = {executor.submit(process_student, sid): sid for sid in sids}
+        if refaire_dict:
+            futures = {executor.submit(process_student, sid): sid for sid in refaire_dict}
+        else:
+            futures = {executor.submit(process_student, sid): sid for sid in sids}
+
         for future in concurrent.futures.as_completed(futures):
             output = future.result()
             if output: