开源软件名称：

开源软件地址：

开源软件介绍：

• What is Eva?
  • Getting Started
• Development
  • Required Tools
  • Example: Hello World
  • Project Structure
  • Development Tasks
    • Running the Test Suite
• Configuration
• About the Eva Data Model
  • Entity-Attribute-Value (EAV)
  • Time-Aware
  • Accumulative
  • Atomic Consistency
    • Transactions
      • Using Object/Map form in transactions
  • Schemas
    • Defining the schema for :attribute1:
• Components
• Running with Docker
• Additional Resources
  • FAQ
    • Is this project or Workiva in any way affiliated with Cognitect?
    • Should I use Eva instead of Datomic?
    • What are the key differences between Eva and Datomic?
    • Why did Workiva build Eva?
    • Why is Workiva open sourcing Eva?
    • What will Workiva's ongoing contributions to Eva be?
  • Maintainers and Contributors
    • Active Maintainers
    • Previous Contributors

What is Eva?

Eva is a distributed database-system implementing an entity-attribute-value data-modelthat is time-aware, accumulative, and atomically consistent. Its API isby-and-large compatible with Datomic's. This software should be considered alphafor the purposes of quality and stability. Check out the FAQ for more info.

Getting Started

If you are brand new to Eva, we suggest reading through this entire readme to familiarize yourself with Eva as a whole. Afterwards, be sure to check out the Eva tutorial series, which break down and go over almost everything you will want to know.

• Eva 101
• Eva 102

Development

Required Tools

1. Java Development Kit (JDK) v8
2. Leiningen Build Tool

Example: Hello World

First we kick off the repl with:

lein repl

Next we create a connection (conn) to an in-memory Eva database. We also need to define the fact (datom) we want to add to Eva. Finally we use the transact call to add the fact into the system.

(def conn (eva/connect {:local true}))(def datom [:db/add (eva/tempid :db.part/user)            :db/doc "hello world"])(deref (eva/transact conn [datom]))

Note: deref can be used interchangeably with the @ symbol.

Now we can run a query to get this fact out of Eva. We don't use conn to make a query but rather we obtain an immutable database value like so:

(def db (eva/db conn))

Next we execute a query that returns all entity ids in the system matching the doc string "hello world".

(eva/q '[:find ?e :where [?e :db/doc "hello world"]] db)

If we want to return the full representation of these entities, we can do that by adding pull to our query.

(eva/q '[:find (pull ?e [*]) :where [?e :db/doc "hello world"]] db)

Project Structure

1. project.clj contains the project build configuration
2. core/* primary releasable codebase for Eva Transactor and Peer-library
   1. core/src clojure source files
   2. core/java-src java source files
   3. core/test test source files
   4. core/resources non-source files
3. dev/* codebase used during development, but not released
   1. dev/src clojure source files
   2. dev/java-src java source files
   3. dev/test test source files
   4. dev/resources non-source files
   5. dev/test-resources non-source files used to support integration testing

Development Tasks

Running the Test Suite

lein test

Configuration

Eva exposes a number of configuration-properties that can be configuredusing java system-properties. Some specific configuration-properties canalso be configured using environment-variables.

The eva.config namespace, linked here, containsdescriptions and default values for the config vars.

About the Eva Data Model

Entity-Attribute-Value (EAV)

EAV data-entities consist of:

1. a distinct entity-id
2. 1-or-more attribute-value pairs associated with a single entity-id

EAV data can be represented in the following (equivalent) forms:

1. as an object or map:

   {:db/id 12345, :attribute1 "value1", :attribute2 "value2"}
2. as a list of EAV tuples:

   [  [12345, :attribute1, "value1"],  [12345, :attribute2, "value2"]]

Time-Aware

To make the EAV data-model time-aware, we extend the EAV-tupleinto an EAVT-tuple containing the transaction-id (T) that introduced the tuple:

[;;  E      A            V         T   [12345, :attribute1, "value1", 500],   [12345, :attribute2, "value2", 500]]

Accumulative

To make the EAVT data-model accumulative, we extend the EAVT-tuplewith a final flag that indicates if the EAV information was added or removedat the transaction-id (T).

[;;  E      A            V         T    added?   [12345, :attribute1, "value1", 500, true],   [12345, :attribute2, "value2", 500, true]]

Under this model, common data operations (create, update, delete) are represented like this:

• Create: a single tuple with added? == true

[[12345, :attribute1, "create entity 12345 with field :attribute1 at transaction 500", 500, true]]

• Delete: a single tuple with added? == false

[[12345, :attribute1, "create entity 12345 with field :attribute1 at transaction 500", 501, false]]

• Update: a pair of deletion and creation tuples

[ ;; At transaction 502 ;;   invalidate the old entry for :attribute2      [12345, :attribute2, "old-value", 502, false] ;;   add a new entry for :attribute2      [12345, :attribute2, "new-value", 502, true]]

The complete history of the database is the cumulative list of these tuples.

Atomic Consistency

Data-updates are submitted as transactions that are processed atomically.This means that when you submit a transaction, either all the changes inthe transaction are applied, or none of the changes are applied.

Transactions

Transactions are submitted as a list of data-modification commands.

The simplest data-modification commands (:db/add, :db/retract) correspondto the accumulative tuples described above:

[  [:db/retract 12345 :attribute2 "old-value"]  [:db/add 12345 :attribute2 "new-value"]]

When this transaction is committed it will produce the following tuples in the database history(where <next-tx> is the next transaction-number):

[  [12345, :attribute2, "old-value", <next-tx>, false]  [12345, :attribute2, "new-value", <next-tx>, true]]

Using Object/Map form in transactions

In addition to the command-form, you can also create/update data using theobject/map form of an entity:

[  {:db/id 12345   :attribute1 "value1"   :attribute2 "value2"}]

This form is equivalent to the command-form:

[  [:db/add 12345 :attribute1 "value1"]  [:db/add 12345 :attribute2 "value2"]]

Schemas

Because all stored data reduces to EAVT tuples,schemas are defined per Attribute, rather than per Entity.

Schemas definitions are simply Entities that have special schema-attributes.

Defining the schema for :attribute1:

[  {:db/id #db/id[:db.part/db]   :db/ident :attribute1   :db/doc "Schema definition for attribute1"   :db/valueType :db.type/string   :db/cardinality :db.cardinality/one   :db.install/_attribute :db.part/db}]

Taking each key-value pair of the example in turn:

• :db/id #db/id[:db.part/db]: declares a new entity-id in the :db.part/db id-partition
• :db/ident :attribute1: declares that :attribute1 is an alias for the entity-id
• :db/doc "Schema definition for attribute1": human-readable string documenting the purpose of :attribute1
• :db/valueType :db.type/string: declares that only string values are allowed for :attribute1
• :db/cardinality :db.cardinality/one: declares that an entity may no-more-than one :attribute1.This means that for an given entity-id, there will only ever be one current tuple of [<entit-id> :attribute1 <value>].Adding a new tuple with this attribute will cause any existing tuple to be removed.
• :db.install/_attribute :db.part/db: declares that this :attribute1 is registered with the database asan installed attribute

Components

• Eva Catalog
• Eva Client Service

Running with Docker

The included docker compose can be used to spin up a completely integrated Eva environment. This includes:

• transactor
• eva-catalog
• activemq
• maria-db image

To spin up said environment run the following commands:

make gen-docker-no-tests # to build Eva with the latest changesmake run-docker

To shut down the the environment use the following command:

make stop-docker

In order to open a repl container that can talk to the environment use:

make repl

And run the following to initially setup the repl environment:

(require '[eva.catalog.client.alpha.client :as catalog-client])(def config (catalog-client/request-flat-config "http://eva-catalog:3000" "workiva" "eva-test-1" "test-db-1"))(def conn (eva/connect config))

Finally, test that everything is working with an empty transaction:

(deref (eva/transact conn []))

A similar result to this should be expected:

{:tempids {}, :tx-data (#datom[4398046511105 15 #inst "2018-06-06T17:35:07.516-00:00" 4398046511105 true]), :db-before #DB[0], :db-after #DB[1]}

Additional Resources

• Additional Documentation.
• The reference documentation at http://docs.datomic.com/ is highly relevant for development within Eva.
• http://www.learndatalogtoday.org/ is a great resource for learning Datalog, the underlying query language in Eva.
• https://github.com/kristianmandrup/datascript-tutorial is another great resource for learning Datalog and how it applies to Datomic (or Eva).

FAQ

Is this project or Workiva in any way affiliated with Cognitect?

No. Eva is its own project we built from the ground up. The API and high-levelarchitecture are largely compatible with Datomic, but the database, up tosome EPL code, was entirely built in-house. We have a list of the mostnotable API differences here.

Should I use Eva instead of Datomic?

If you are looking for an easy system to move to production quickly, almostcertainly not. Eva is far less mature and has seenfar less time in battle. Datomic Cloud is an amazing (and supported) productthat is far easier to stand up and run with confidence. Eva is provided as-is.

What are the key differences between Eva and Datomic?

There are a handful of small API differences in the Peer, whereas the Clientsare quite distinct. For example, a Connection in Eva is constructed using abaroque configuration map, not a string. From an operational standpoint,Datomic is far more turn-key. There are also likely some low-levelarchitectural differences between the systems that will cause them toexhibit different run-time characteristics. For example, our indexes arebacked by a persistent B^