From 11b644c447ae22816e111dcb5557e5c1f33cd52a Mon Sep 17 00:00:00 2001
From: Siddharth VP <siddharthvp@gmail.com>
Date: Sat, 7 Oct 2023 15:28:57 +0530
Subject: [PATCH] Remove invalid database names from dropdown suggestions

This is done by only including dbs against which successful queries have been run in the past.

Also add a Cache-Control header is response so the result is cached in the browser for one hour or till a hard reload is done.

Bug: T289943
---
 quarry/default_config.yaml |  6 ++++++
 quarry/web/api.py          | 21 ++++++++++++++-------
 2 files changed, 20 insertions(+), 7 deletions(-)

diff --git a/quarry/default_config.yaml b/quarry/default_config.yaml
index e74d3ac..a5030f9 100644
--- a/quarry/default_config.yaml
+++ b/quarry/default_config.yaml
@@ -5,6 +5,7 @@ OAUTH_SECRET_TOKEN: "83ed4ac72176286f5c7798459f0a6b16a96853c4"
 SECRET_KEY: 'aaarggghhpleaserescuemeiamstuckinarandomnumbergeneratorfactory'
 DEBUG: True
 DB_HOST: 'db'
+DB_PORT: 3306
 DB_NAME: 'quarry'
 DB_USER: 'quarry'
 DB_PASSWORD: 'quarry'
@@ -31,6 +32,11 @@ QUERY_RESULTS_PER_PAGE: 50
 KILLER_LOG_PATH: 'killer.log'
 MAINTENANCE_MSG: 'This is your local development environment.'
 
+TOOLS_DB_HOST: 'tools.db.svc.wikimedia.cloud'
+TOOLS_DB_PORT: 3306
+TOOLS_DB_USER: ''
+TOOLS_DB_PASSWORD: ''
+
 # https://flask-caching.readthedocs.io/en/latest/#configuring-flask-caching
 CACHE_TYPE: 'RedisCache'
 CACHE_DEFAULT_TIMEOUT: 30
diff --git a/quarry/web/api.py b/quarry/web/api.py
index dd54037..d5425ad 100644
--- a/quarry/web/api.py
+++ b/quarry/web/api.py
@@ -225,21 +225,28 @@ def pref_set(key, value) -> Union[Tuple[str, int], Tuple[Response, int]]:
 @api_blueprint.route("/api/dbs")
 def get_dbs() -> Response:
     known_dbs = (
-        g.conn.session.query(QueryRevision.query_database).distinct().all()
+        g.conn.session.query(QueryRevision.query_database)
+        .join(QueryRun, QueryRevision.id == QueryRun.query_rev_id)
+        .filter(QueryRun.extra_info.notlike('{"error%'))
+        .filter(QueryRevision.query_database.isnot(None))
+        .distinct()
+        .all()
     )
-    return Response(
+    response = Response(
         json.dumps(
             {
                 "dbs": list(
                     set(
-                        db_result[-1].strip().rstrip(";")
+                        db_result[-1].strip()
                         for db_result in known_dbs
-                        # the db data might be NULL, empty strings or spaces+tabs only so this helps a bit to show only
-                        # likely names
-                        if db_result[-1] and db_result[-1].strip()
+                        if db_result[-1]
                     )
                 )
-            }
+            },
+            # Remove spaces to reduce response size
+            separators=(",", ":"),
         ),
         mimetype="application/json",
     )
+    response.headers["Cache-Control"] = "max-age=3600, public"
+    return response