analyze_cache_misses.clj

(ns mage.escherize.analyze-cache-misses
  (:require
   [clojure.edn :as edn]
   [clojure.java.io :as io]
   [clojure.string :as str]
   [mage.util :as util]))

(defn read-log [log-file]
  (map second (edn/read-string (str "[" (slurp log-file) "]"))))

(defn most-common [n coll]
  (->> coll
       (frequencies)
       (sort-by val >)
       (take n)
       (mapv reverse)))

(defn most-common-cache-misses [log-file]
  (->> log-file
       read-log
       (map :schema-key)
       (most-common 3)))

;;    Yo, this indicates that we are MISSING the cache for this schema.
(most-common-cache-misses "app-startup-master.edn")
;; => [(115 "#object[clojure.core$simple_keyword_QMARK_ 0x4d3f8662 \"clojure.core$simple_keyword_QMARK_@4d3f8662\"]")
;;     (89 ":metabase.lib.schema.expression/base-type")
;;     (89 "[:= :-]")]


(most-common-cache-misses "mlv2-test-master.edn")
;; => [(115 "#object[clojure.core$simple_keyword_QMARK_ 0x2e38e93a \"clojure.core$simple_keyword_QMARK_@2e38e93a\"]")
;;     (89 ":metabase.lib.schema.expression/base-type")
;;     (89 "[:= :-]")]


;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;; on fixed branch ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;


;; giant schemas, probably race conditions:
(most-common-cache-misses "mlv2-test-fix-branch.edn")
;; => [(13 [:maybe :metabase.lib.schema.drill-thru/drill-thru])
;;     (8 :metabase.lib.drill-thru.object-details/object-detail-drill-out)
;;     (8
;;      [:merge [:merge [:merge .... [:map [:expected-query :map] [:drill-args {:optional true} [:maybe [:sequential :any]]]]])]


;; `(def Filter [:ref ::Filter])` has 32 uses
(most-common-cache-misses "app-startup-fix-branch.edn")
;; => [(3 :metabase.lib.test-util.metadata-providers.mock/mock-metadata)
;;     (2
;;      [:map
;;       [:key :keyword]
;;       [:db-schedule-column :keyword]
;;       [:job-class :metabase.sync.task.sync-databases/class]
;;       [:name :string]])
;;     (2 :metabase.lib.schema.metadata/card)]