Commands: manage update search vector with maintenance script

2 files changed, 67 insertions, 17 deletions

diff --git a/CHANGES.md b/CHANGES.md
index 00d70235d..434a94855 100644
--- a/CHANGES.md
+++ b/CHANGES.md
@@ -6,6 +6,10 @@ date: 2022-12-05
 Ishtar changelog
 ================
 
+### Features/improvements ###
+- Commands: manage update search vector with maintenance script
+
+
 v4.0.32 - 2022-12-12
 --------------------
 
diff --git a/ishtar_common/management/commands/ishtar_maintenance.py b/ishtar_common/management/commands/ishtar_maintenance.py
index 1eeab179a..c4ad52ab6 100644
--- a/ishtar_common/management/commands/ishtar_maintenance.py
+++ b/ishtar_common/management/commands/ishtar_maintenance.py
@@ -28,6 +28,42 @@ if not os.path.exists(log_path):
     os.mkdir(log_path, mode=0o770)
 
 
+def task_update_search_vectors(quiet=False, log=False):
+    for model in apps.get_models():
+        if not hasattr(model, "update_search_vector") or \
+                not (getattr(model, "BASE_SEARCH_VECTORS", None) or
+                     getattr(model, "INT_SEARCH_VECTORS", None) or
+                     getattr(model, "M2M_SEARCH_VECTORS", None) or
+                     getattr(model, "PARENT_SEARCH_VECTORS", None)):
+            continue
+        store_results = []
+        msg = "-> processing {}: ".format(model._meta.verbose_name)
+        q = model.objects
+        ln = q.count()
+
+        changed_nb = 0
+        for idx, obj_id in enumerate(q.values("pk").all()):
+            obj = model.objects.get(pk=obj_id["pk"])
+            if not quiet:
+                cmsg = "\r{} {}/{}".format(msg, idx + 1, ln)
+                sys.stdout.write(cmsg)
+                sys.stdout.flush()
+            changed = obj.update_search_vector()
+            if changed:
+                if log:
+                    store_results.append(
+                        (
+                            obj.pk,
+                            str(obj),
+                        )
+                    )
+                changed_nb += 1
+
+        log_name = "update-search-vector-" + slugify(model.__name__)
+        csv_cols = ["id", "name"]
+        _end_task(changed_nb, msg, quiet, store_results, log, log_name, csv_cols)
+
+
 def task_check_cached_label(quiet=False, log=False):
     for model in apps.get_models():
         if model.__name__.startswith("Historical"):
@@ -75,21 +111,27 @@ def task_check_cached_label(quiet=False, log=False):
                 obj._no_move = True
                 obj._no_geo_check = True
                 obj.save()
+        log_name = "update-cached-label-" + slugify(model.__name__)
+        csv_cols = ["id", "name", "attribute", "old", "new"]
+        _end_task(changed_nb, msg, quiet, store_results, log, log_name, csv_cols)
+
+
+def _end_task(changed_nb, msg, quiet, store_results, log, log_name, csv_cols):
+    if not quiet:
+        if changed_nb:
+            cmsg = f"\r{msg} {changed_nb} updated" + 20 * " " + "\n"
+        else:
+            cmsg = " " * 80
+        sys.stdout.write(cmsg)
+    if log and changed_nb:
+        filename = f"{get_time().replace(':', '')}-{log_name}.csv"
+        path = os.sep.join([log_path, filename])
+        with open(path, 'w+') as fle:
+            writer = csv.writer(fle)
+            writer.writerow(csv_cols)
+            writer.writerows(store_results)
         if not quiet:
-            if changed_nb:
-                cmsg = f"\r{msg} {changed_nb} updated" + 20 * " " + "\n"
-            else:
-                cmsg = " " * 80
-            sys.stdout.write(cmsg)
-        if log and changed_nb:
-            filename = f"{slugify(model.__name__)}-{get_time().replace(':', '')}.csv"
-            path = os.sep.join([log_path, filename])
-            with open(path, 'w+') as fle:
-                writer = csv.writer(fle)
-                writer.writerow(["id", "name", "attribute", "old", "new"])
-                writer.writerows(store_results)
-            if not quiet:
-                sys.stdout.write(f"log: {path} written.")
+            sys.stdout.write(f"log: {path} written.")
 
 
 def task_main_image(quiet=False, log=False):
@@ -113,7 +155,7 @@ def task_main_image(quiet=False, log=False):
             sys.stdout.write(f"{nb} main image fixed for {model.__name__}\n")
 
 
-def task_missing_parcels(quiet=False):
+def task_missing_parcels(quiet=False, log=False):
     Parcel = apps.get_model("archaeological_operations", "Parcel")
     q = Parcel.objects.filter(context_record__isnull=False, operation=None)
     nb = q.count()
@@ -145,13 +187,17 @@ def get_time():
 
 TASKS = {
     "main_image": {
-        "help": "for items with images and no main image, put the first one created as a main image.",
+        "help": "for items with images and no main image, put the first one created as a main image",
         "action": task_main_image,
     },
     "cached_label": {
-        "help": "regenerate cached label on all tables if necessary",
+        "help": "regenerate cached label on all tables",
         "action": task_check_cached_label,
     },
+    "update_search_vector": {
+        "help": "regenerate search vectors on all tables",
+        "action": task_update_search_vectors,
+    },
     "operation_missing_parcels": {
         "help": "fix lost parcel association on operation from context records.",
         "action": task_missing_parcels,