Add scripts to produce reports on the merge queue flake rate

Determining the flake rate helps us decide how much effort to put into
improving it.

The merge queue flake rate is only half the effective flake rate, since
changes have to pass presubmit before they are allowed onto the merge
queue.

Bug: b/470157423

Author: Databean

tools/merge_queue_stats/jq_report.sh

@@ -0,0 +1,48 @@
+#!/bin/bash
+
+set -e
+
+err() {
+  echo "[$(date +'%Y-%m-%dT%H:%M:%S%z')]: $*" >&2
+}
+
+# https://jqlang.org/manual/
+
+JQ_TIMELINE_EVENTS=$(cat <<'END'
+[.[].data.repository.pullRequests.nodes]
+  | flatten
+  | sort_by(.url|split("/")|last|tonumber)
+END
+)
+
+jq -s "${JQ_TIMELINE_EVENTS}" prs_*.json > timeline_events.json
+
+JQ_COUNT_EVENTS=$(cat <<'END'
+map({
+  url: .url,
+  events: (
+    .timelineItems.nodes
+      | group_by(.__typename)
+      | map({(.[0].__typename): .|length})
+      | add)
+})
+END
+)
+jq "${JQ_COUNT_EVENTS}" timeline_events.json > count_events.json
+
+JQ_ONLY_MERGE_QUEUE=$(cat <<'END'
+.[]
+  | select(.events|has("RemovedFromMergeQueueEvent"))
+  | select(.events|(has("MergedEvent")))
+END
+)
+jq "${JQ_ONLY_MERGE_QUEUE}" count_events.json > only_merge_queue.json
+
+JQ_OUTPUT=$(cat <<'END'
+[.url,.events.RemovedFromMergeQueueEvent]
+  | @csv
+END
+)
+jq "${JQ_OUTPUT}" only_merge_queue.json > output.csv
+
+err "Wrote output to output.csv"

tools/merge_queue_stats/pull_request_timeline_events.sh

@@ -0,0 +1,47 @@
+#!/bin/bash
+
+set -e
+
+err() {
+  echo "[$(date +'%Y-%m-%dT%H:%M:%S%z')]: $*" >&2
+}
+
+# - https://docs.github.com/en/graphql
+# - https://graphql.org/learn/queries/
+# - https://docs.github.com/en/graphql/guides/using-pagination-in-the-graphql-api
+
+page_token=\"\"
+
+for (( page_num = 1;; page_num++ )) do
+  err "Requesting page $page_num"
+  query_str=$(cat <<EOF
+    query {
+      repository(owner:"google", name:"android-cuttlefish") {
+        pullRequests(last: 100, before:${page_token}) {
+          pageInfo {
+            endCursor
+            startCursor
+            hasNextPage
+            hasPreviousPage
+          }
+          nodes {
+            url
+            timelineItems(first: 100) {
+              nodes {
+                __typename
+              }
+            }
+          }
+        }
+      }
+    }
+EOF
+  )
+  gh api graphql -f query="${query_str}" | jq > prs_$page_num.json
+
+  page_info=.data.repository.pullRequests.pageInfo
+  if [[ $(jq $page_info.hasPreviousPage prs_$page_num.json) != "true" ]]; then
+    exit 0;
+  fi
+  page_token=$(jq $page_info.startCursor prs_$page_num.json)
+done