feat(lisp): Implement persistent storage via cl-prevalence

Uses the cl-prevalence system to store tasks on disk. The storage
location is either relative to the working directory in which the
system is started or determined (with priority) by the environment
variable `GEMMA_DATA_DIR`.

gemma.asd

@@ -13,7 +13,10 @@
   :version "0.1.0"
   :author "Vincent Ambo"
   :license "GPLv3"
-  :depends-on (alexandria local-time hunchentoot yason)
+  :depends-on (local-time
+               hunchentoot
+               cl-json
+               cl-prevalence)
   :components ((:module "src"
                 :components
                 ((:file "gemma"))))

src/gemma.lisp

@@ -9,7 +9,6 @@
 
 (defpackage gemma
   (:use :cl
-        :alexandria
         :hunchentoot
         :local-time
         :cl-json))
@@ -23,7 +22,10 @@
 ;; Define task management system
 ;;
 (defclass task ()
-  (;; (Unique) name of the task
+  ((id :reader id
+       :initarg :id)
+
+   ;; (Unique) name of the task
    (name :type symbol
          :initarg :name
          :accessor name-of)
@@ -40,24 +42,38 @@
 
    ;; Last completion time
    (done-at :type local-time:timestamp
+            :initarg :done-at
             :accessor last-done-at)))
 
-(defvar *tasks*
+(defvar *gemma-data-dir*
-  (make-hash-table)
+  (pathname (or (sb-posix:getenv "GEMMA_DATA_DIR")
-  "List of all tasks registered in this Gemma instance.")
+                (sb-posix:getcwd)))
+  "Directory in which to store Gemma data.")
+
+(defvar *p-tasks*
+  (cl-prevalence:make-prevalence-system *gemma-data-dir*)
+  "All tasks registered in this Gemma instance.")
+
+;; Initialise database ID counter
+(or (> (length (cl-prevalence:find-all-objects *p-tasks* 'task)) 0)
+    (cl-prevalence:tx-create-id-counter *p-tasks*))
 
 (defmacro deftask (task-name days &optional description)
-  `(setf (gethash (quote ,task-name) *tasks*)
+  (unless (get-task task-name)
-         (make-instance (quote task)
+    `(progn (cl-prevalence:tx-create-object
-                        :name (quote ,task-name)
+             *p-tasks*
-                        :days ,days
+             'task
-                        :description (or ,description ""))))
+             (quote ((name ,task-name)
+                     (days ,days)
+                     (description ,(or description ""))
+                     (done-at ,(local-time:now)))))
+            (cl-prevalence:snapshot *p-tasks*))))
 
 (defun get-task (name)
-  (gethash name *tasks*))
+  (cl-prevalence:find-object-with-slot *p-tasks* 'task 'name name))
 
 (defun list-tasks ()
-  (alexandria:hash-table-values *tasks*))
+  (cl-prevalence:find-all-objects *p-tasks* 'task))
 
 (defun days-remaining (task)
   "Returns the number of days remaining before the supplied TASK reaches its
@@ -76,8 +92,9 @@ maximum interval."
 
 (defun complete-task (name &optional at)
   "Mark the task with NAME as completed, either now or AT specified time."
-  (setf (last-done-at (get-task name))
+  (cl-prevalence:tx-change-object-slots *p-tasks* 'task
-        (or at (local-time:now))))
+                                        (id (get-task name))
+                                        `((done-at ,(or at (local-time:now))))))
 
 ;;
 ;; Define web API
@@ -139,9 +156,15 @@ maximum interval."
 (defun randomise-completion-times ()
   "Set some random completion timestamps for all tasks"
   (mapcar
-   (lambda (key) (complete-task key (local-time:timestamp- (local-time:now)
+   (lambda (task)
+     (complete-task (name-of task)
+                    (local-time:timestamp- (local-time:now)
                                            (random 14)
                                            :day)))
-   (alexandria:hash-table-keys *tasks*)))
+   (cl-prevalence:find-all-objects *p-tasks* 'task)))
+
+(defun clear-all-tasks ()
+  (mapcar (lambda (task) (cl-prevalence:tx-delete-object *p-tasks* 'task (id task)))
+          (cl-prevalence:find-all-objects *p-tasks* 'task)))
 
 ;; (randomise-completion-times)