From d2cb6384ccffe12a24410376609923872dfe707d Mon Sep 17 00:00:00 2001
From: csae8092 <peter.andorfer@oeaw.ac.at>
Date: Tue, 28 May 2024 16:51:59 +0200
Subject: [PATCH 1/2] deleted nonsense professions, closes #195

---
 issue__195_deletenotusedprofessions.ipynb | 127 ++++++++++++++++++++++
 1 file changed, 127 insertions(+)
 create mode 100644 issue__195_deletenotusedprofessions.ipynb

diff --git a/issue__195_deletenotusedprofessions.ipynb b/issue__195_deletenotusedprofessions.ipynb
new file mode 100644
index 0000000..09a4a23
--- /dev/null
+++ b/issue__195_deletenotusedprofessions.ipynb
@@ -0,0 +1,127 @@
+{
+ "cells": [
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "1e1c119a",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# 2024-05-28 run against production\n",
+    "import pandas as pd"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "d95d08e9",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "ProfessionType.objects.all().count()"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "5543f8e0",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "ProfessionType.objects.filter(person=None).count()"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "5b071e98",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "data = []\n",
+    "for x in ProfessionType.objects.filter(person=None).distinct():\n",
+    "    data.append([x.id, x.name])"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "30beda86",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "import pandas as pd"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "4fcb8186",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "df = pd.DataFrame(data, columns=[\"id\", \"name\"])"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "d7acc58e",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "df.to_csv(\"hansi.csv\", index=False)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "18d88961",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "for x in ProfessionType.objects.filter(person=None).distinct():\n",
+    "    x.delete()"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "6bac8d6e",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "ProfessionType.objects.all().count()"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "17dc6357",
+   "metadata": {},
+   "outputs": [],
+   "source": []
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "Django Shell-Plus",
+   "language": "python",
+   "name": "django_extensions"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.10.12"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 5
+}

From 75f34a916879bb3a9fb1cbe0a088262989a3a4fb Mon Sep 17 00:00:00 2001
From: csae8092 <peter.andorfer@oeaw.ac.at>
Date: Mon, 3 Jun 2024 20:22:19 +0200
Subject: [PATCH 2/2] removes entities with wikidata-ids from no wikidata
 collection

---
 dumper/management/commands/wikidata_minter.py | 8 +++++++-
 1 file changed, 7 insertions(+), 1 deletion(-)

diff --git a/dumper/management/commands/wikidata_minter.py b/dumper/management/commands/wikidata_minter.py
index 795aca9..d2de2fc 100644
--- a/dumper/management/commands/wikidata_minter.py
+++ b/dumper/management/commands/wikidata_minter.py
@@ -16,9 +16,15 @@ class Command(BaseCommand):
     help = "mint WikiData IDs for GND-URIs"
 
     def handle(self, *args, **kwargs):
-        LIMIT = 100
+        LIMIT = 2
         USER_AGENT_PMB = "pmb (https://pmb.acdh.oeaw.ac.at)"
         col, _ = Collection.objects.get_or_create(name="No WikiData-ID found")
+        ents = TempEntityClass.objects.filter(uri__uri__icontains="wikidata").filter(collection=col)
+        print(f"found {ents.count()} entities with wikidata-ids but related to {col}")
+        if ents:
+            print(f"remove relation to {col}")
+            for x in tqdm(ents, total=ents.count()):
+                x.collection.remove(col)
         types = ["d-nb.info", "geonames"]
         for uri_type in types:
             print(f"processing URIS with type: {uri_type}")