Skip to content

A RAD database adapter that is generalized to work with key-value databases

Notifications You must be signed in to change notification settings

fulcrologic/fulcro-rad-kvstore

Repository files navigation

fulcro rad kvstore

A Fulcro RAD plugin that provides an interface for building database adaptors to support databases that generally store things as documents with keys.

Warning
ALPHA SOFTWARE. This library is not yet heavily tested and there are reports that it does not yet work well as a Clojurescript-side adapter. Contributions are encouraged. Fulcrologic does not have the resources at present to maintain this particular plugin.

This Fulcro RAD plugin with adaptors is in development and is not yet ready for production.

A recent copy of the RAD Demo source code is included. So this project’s source code makes up the first three steps here:

RAD Demo -> Key Value plugin -> ::key-value/key-store -> ::kv-key-store/store -> Redis | Firebase | MongoDB ...

::kv-key-store/store is the binding for the adaptor supplied by another library.

Take a look inside shadow-cljs.edn at [:builds :main :devtools :preloads]. Your Fulcro Inspect setup may be different to mine. Just reverse the commented-out lines if that’s the case.

In a terminal type yarn then make cljs. Shadow CLJS should lead you through the browser/CLJS part of the installation. For the server side setup a local REPL choosing 'Run with Deps' with the alias dev. Bring the REPL up. (user/refresh) will reload the namespaces (compilation step) and (user/restart) will get the Demo’s dependency graph of Mount components up and running, which includes seeding the database.

Back to the client. Bring up a browser tab at http://localhost:9630/ and start a watch on the :main build. When compilation is complete refresh the browser at http://localhost:3000/. The Demo Application should now be working. Press the "Login" button in the top right corner and then "Login" again on the pre-filled modal Login Dialog.

The following namespace aliases are used in the content of this document:

(ns x
  (:require
     [com.fulcrologic.rad.database-adapters.key-value :as key-value]
     [com.fulcrologic.rad.database-adapters.key-value.key-store :as kv-key-store]
     [com.fulcrologic.rad.database-adapters.key-value-options :as kvo]
     [com.fulcrologic.rad.database-adapters.key-value.write :as kv-write]
     [com.fulcrologic.rad.database-adapters.key-value.pathom :as kv-pathom]
     ;; Don't need an alias to load a ns so the multimethod's defmethod can be found
     [com.fulcrologic.rad.database-adapters.key-value.redis]
     [com.fulcrologic.rad.database-adapters.datomic :as datomic]
     [com.example.components.config :as config]
     [com.example.components.delete-middleware :as delete]
     [com.example.components.save-middleware :as save]
     [com.example.components.parser :as parser]
     [com.fulcrologic.rad.pathom :as pathom]
     [immutant.web :as web]
     ))

The RAD Demo is at com.example.

This library allows Fulcro RAD to read and write to a Konserve-supported key value database. Thus supporting a new database may mean a trivial 'include it' fix here, or contributing to Konserve in the first instance. For example at time of writing MongoDB was not supported by Konserve.

Entities are always stored with:
  • key being a stricter version of eql/ident? (specifically table has to be a /id keyword and the second part a uuid?)

  • value being a map of attributes (an entity!), where the attribute values are either scalars or references. A reference can be either an ident or a vector of idents.

To create a ::key-value/key-store call key-value/start passing in the RAD configuration map. ::kv-key-store/store inside the returned map is where the real Konserve adaptor is located. Unless the :kind of adaptor you are creating is :memory, :filestore or :indexeddb make sure you have required the namespace associated with the adaptor you are requesting, so that the defmethod of the make-adaptor multimethod can be found. As mentioned in a comment in the example :require, there is no need for an alias.

{::kv-key-store/keys [store]} (key-value/start config/config)

::key-value/key-store can be destructured:

{::kv-key-store/keys [store table->rows table->ident-rows ident->entity write-entity ids->entities]}

The easiest way to get started is to use these functions directly. All bar one of them are for retrieving entities. For example ident→entity returns what is stored at an ident, and table→rows returns all the data of a table. Here is how you would get all the active accounts (from the Demo project):

(->> (table->rows :account/id)
     (filter :account/active?)
     (mapv #(select-keys % [:account/id])))

Here’s one way you could do the same thing using Konserve’s store directly:

(<!!
  (go
    (->> (vals (<! (k/get-in store [:account/id])))
         (filter :account/active?)
         (mapv #(select-keys % [:account/id])))))

Here both are equally 'async', but when there are many fetches you are going to get better performance using ::kv-key-store/store directly.

Additionally there are some useful functions in the ::kv-write namespace: import, remove-table-rows and write-tree. And in the ::key-value namespace there is export.

As with all RAD Database plugins:
  • It comes with delete and save middleware so you don’t have to write your own Pathom mutations.

  • Is capable of generating the 'id → entity' Pathom resolvers.

  • Injects values into the Pathom env that are then picked up by these resolvers.

  • Has a function to start the plugin that picks up values set in the RAD configuration.

Table 1. Entry points from RAD Demo into Key Value plugin
Description Demo Mount component Function call

Generation of Pathom resolvers

com.example.components.auto-resolvers/automatic-resolvers

kv-pathom/generate-resolvers

Save Middleware

::save/middleware

kv-pathom/wrap-save

Delete Middleware

::delete/middleware

kv-pathom/wrap-delete

Injection into Pathom’s env

::parser/parser

kv-pathom/pathom-plugin

Create connection

::config/config,com.example.components.seeded-connection/kv-connections

key-value/start

An area of integration not included on this table is where your hand written resolvers need to query the database. Here queries will receive env, which they can then pass to the function kv-pathom/env→key-store to get the ::key-value/key-store. With this key store in hand either use one of the pre-written functions the map contains, or use Konserve directly.

{::kv-key-store/keys [store]} (kv-pathom/env->key-store env)

Compared to the Datomic plugin some things have been left undone.

Even although the Pathom env of this plugin has ::key-value/connections and ::key-value/databases only one connection/database is ever used. (With this type of database there is no difference between a connection and a database). So with current functionality we could get away with just having ::key-value/database. The Datomic plugin requires this setup to support sharding, which has been left undone for the Key Value plugin. Note that even if we invented a new key such as ::key-value/database to go into the env, we could still keep ::key-value/databases in the config file. You’ll probably never need to worry about all this however, as the function kv-pathom/env→key-store abstracts it away, and you’ll usually be editing .edn config files rather than creating them from scratch.

There is no automatic schema generation. Unlike Datomic, Key Value databases do not have schemas to generate.

This plugin currently eschews looking to RAD attributes to ascertain the primary key of entities, instead making the assumption that your entities are strict (according to the namespace ::strict-entity). Thus if you do not need automatic Pathom resolver generation then this plugin can be used outside of RAD (it is then not actually a plugin).

The last thing this plugin lacks is the function datomic/empty-db-connection that gives a data-less database - good for making tests that build up just the data they need, not touching existing databases. The closest equivalent we have is kv-key-store/import which requires an existing database and can be used to destroy the existing data (so not actually importing anything).

This section more for maintainers, just outlining steps to be taken when copying in the latest RAD Demo.

Apart from com.example.components and config, overwrite all with new files. So com.example: client.cljs, model.cljc and ui.cljc, then com.example.model and com.example.ui. The mount components should not change but you might want to check the .edn config files, as well as deps.edn for library version changes or replacements.

Put back com.example.ui.landing-page and com.example.ui.landing-page-2 if you overwrote the entire com.example.ui directory. Then alter the new version of com.example.ui requiring [com.example.ui.landing-page :as lp] and prefixing the LandingPage component with lp. Check the local LandingPage is the same and remove it. com.example.client will need a similar change to that of com.example.ui - require lp and reference it.

The Demo project should now work; the following changes are optional.

time-zone is Datomic-specific so remove it by commenting out timezone/attributes from com.example.model and on whatever UIs TZ appears - in fo/attributes in AccountForm for example (timezone/zone-id). Remove the use of :db/ident in com.example.model.authorization/login! - it is Datomic specific and unnecessary.

Ensure that deps.edn, package.json and shadow-cljs.edn are up to date.

There will be a base Maven artifact and one for each substantial adaptor. Thus :memory, :filestore and :indexeddb will be covered by the base artifact.

Please see the Makefile to produce your own local jars. But really all you need to do is implement the multimethod key-value/make-adaptor. Many defmethod examples in this source code.

Copyright (c) 2017-2019, Fulcrologic, LLC The MIT License (MIT)

Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions:

The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software.

THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.

About

A RAD database adapter that is generalized to work with key-value databases

Resources

Stars

Watchers

Forks

Packages

No packages published