From 269c6bce54de3422c539d70e668da62469bf4425 Mon Sep 17 00:00:00 2001
From: Edward Betts <edward@4angle.com>
Date: Thu, 6 Jun 2024 13:47:22 +0000
Subject: [PATCH 1/2] Need to use BigInteger for lastrevid

---
 matcher/model.py | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/matcher/model.py b/matcher/model.py
index 40d56f5..4f61118 100644
--- a/matcher/model.py
+++ b/matcher/model.py
@@ -107,7 +107,7 @@ class Item(Base):
     aliases = Column(postgresql.JSONB)
     sitelinks = Column(postgresql.JSONB)
     claims = Column(postgresql.JSONB, nullable=False)
-    lastrevid = Column(Integer, nullable=False, unique=True)
+    lastrevid = Column(BigInteger, nullable=False, unique=True)
     locations: Mapped[list["ItemLocation"]] = relationship(
         "ItemLocation", cascade="all, delete-orphan", backref="item"
     )

From 13ecf4526d1f85a00199c89ccbcaf3ccc50450fa Mon Sep 17 00:00:00 2001
From: Edward Betts <edward@4angle.com>
Date: Thu, 6 Jun 2024 13:48:18 +0000
Subject: [PATCH 2/2] Catch MediaWiki database timeout errors and retry

---
 update.py | 20 +++++++++++++++++++-
 1 file changed, 19 insertions(+), 1 deletion(-)

diff --git a/update.py b/update.py
index f3626a7..ee26a3b 100755
--- a/update.py
+++ b/update.py
@@ -6,6 +6,8 @@ import json
 import typing
 from time import sleep
 
+import requests.exceptions
+
 from matcher import model, wikidata, wikidata_api
 from matcher.database import init_db, session
 
@@ -81,7 +83,14 @@ def handle_edit(change: Change) -> None:
         print(f"{ts}: no need to update {qid}")
         return
 
-    entity = wikidata_api.get_entity(qid)
+    for attempt in range(100):
+        try:
+            entity = wikidata_api.get_entity(qid)
+        except requests.exceptions.ConnectionError:
+            print("connection error, retrying.")
+            sleep(10)
+        else:
+            break
     entity_qid = entity.pop("id")
     if entity_qid != qid:
         print(f"{ts}: item {qid} replaced with redirect")
@@ -123,6 +132,15 @@ def update_database() -> None:
         r = wikidata_api.get_recent_changes(rcstart=start, rccontinue=rccontinue)
 
         reply = r.json()
+        if (
+            "error" in reply
+            and reply["error"]["code"] == "internal_api_error_DBQueryTimeoutError"
+        ):
+            print(reply)
+            sleep(10)
+            continue
+        if "query" not in reply:
+            print(reply)
         for change in reply["query"]["recentchanges"]:
             rctype = change["type"]
             timestamp = change["timestamp"]