Skip to content


Switch branches/tags

Name already in use

A tag already exists with the provided branch name. Many Git commands accept both tag and branch names, so creating this branch may cause unexpected behavior. Are you sure you want to create this branch?

Latest commit


Git stats


Failed to load latest commit information.
Latest commit message
Commit time

MoleculeAdmin for Datomic dbs


An advanced web app (built with ScalaJS) to administrate Datomic databases:

  • Create and delete Datomic databases.
  • Build and update Datomic database schemas.
  • Build and run complex queries from a graphical query builder or directly with molecules.
  • Sort, filter, group and browse large data sets easily and fast.
  • Add/edit data directly in table cells with easy key-command navigation.
  • Perform custom group edits on any number of entities.
  • Recursively expand entities with simple mouseovers for fast exploration.
  • Watch/undo transaction history.
  • See/edit grouped values.
  • Shows queries as molecules and datalog queries (for easy copy/paste).
  • Automatic query optimization for fast query performance.

MoleculeAdmin uses molecules as a query representation. But it can be used without knowing about/using molecules or Scala for that matter (Clojure/Java folks are welcome!).

For performing Datalog queries, please use the official Datomic Console (requires registration or using a Pro version of Datomic).


To explore MoleculeAdmin with the mBrainz sample database,

  1. Download a Datomic distribution like the free version
  2. Download and untar the mBrainz sample database
    cd datomic-free-0.9.5703.21
    wget -O mbrainz.tar
    tar -xvf mbrainz.tar
  3. Start the transactor in its own process
    bin/transactor config/samples/
  4. In another process (open new tab/window in your terminal), install the mbrainz database
    bin/datomic restore-db file:mbrainz-1968-1973 datomic:free://localhost:4334/mbrainz-1968-1973
  5. Download MoleculeAdmin
    cd ..
    git clone

Use MoleculeAdmin

Having downloaded and installed all the necessary parts listed above,

  1. Start the Datomic transactor in its own process, something like this (might already be running):
    cd datomic-free-0.9.5703.21
    bin/transactor -Xmx4g -Xms4g -Ddatomic.txTimeoutMsec=120000 config/samples/
  2. In another process, start the application:
    cd molecule-admin
    sbt run
  3. Wait for the server to have started:
    (Server started, use Enter to stop and go back to the console...)
  4. Then open localhost:9001 - this can take a while the first time since all dependencies are resolved and the whole project compiled (check progress in terminal).

You should arrive at a list of current databases, something like this (if you haven't already created other databases):

A few more sample databases were also created that you can freely play around with by exploring their Schema or perform a Query:

  • CoreTest - contains all possible data types. Note how the Schema defines those and how the query builder takes each type into consideration and only allow valid settings.
  • Partition/Partition1 - exploring multiple partitions that each contains one or more namespaces.
  • Tree - Used to test building complex hierarchies with the query builder.
  • mBrainz - large real data set that can give a sense of the query builder capabilities and the paging/editing features.

It's recommended to use MoleculeAdmin in the Google Chrome browser for expected look and behaviour.

Group edits with custom Scala code

An advanced feature of MoleculeAdmin is to make group edits on a whole column of data. In the input cell above the group edit column, you define a Scala transformation lambda that will be applied to each value in turn. The lambda code snippet will be automatically compiled and applied once you hit enter and transform all current values of the column, both visible on the current page and subsequent values. Make sure you have filtered exactly the values that you want to transform!

Your Scala lambda code snippet is compiled by a separately running ScalaFiddle server to javascript code that is then applied to the current values. Both the browser table and memory and the Datomic database is updated.

If you want to use this group edit functionality, you'll need to once initially clone the adapted ScalaFiddle server for MoleculeAdmin:

git clone

From here on you only need to start the server and have it running in the background to allow group edits:

cd scalafiddle-core
sbt "~; router/reStart; compilerServer06x/reStart"

MoleculeAdmin can then talk automatically in the background with the scalafiddle-core server through http://localhost:8880/embed.

See examples of group edits here.

Enable your own Datomic database

For MoleculeAdmin to be able to work with a database it needs a path to a schema definition file which is a molecule representation of a Datomic database schema. You don't need to use molecules or Scala to use MoleculeAdmin, as long as you provide a schema definition file that defines your Datomic database schema. Once you have created such file you can enter its full path into MoleculeAdmin:

Schema development in Scala

Molecule uses the schema definition file as a blueprint for generating Scala boilerplate code that you can use to make type safe queries in Scala against a Datomic database.

In the first life cycles of a Scala project, you'll likely develop your database schema by writing a lot of changes directly to the schema definition file and re-compile the project regularly to have the boilerplate code reflecting your changes.

But when the schema stabilizes and you start to enter real data into the database, you'll want a more flexible way to make changes to the schema and at the same time have those changes continuously propagated to your generated boilerplate. This is when MoleculeAdmin can "take over" the responsibility of keeping your schema definition file in sync with your database. So instead of you writing manually to the schema definition file, you let MoleculeAdmin write changes directly to it.

Clojure/Java users

For Clojure/Java users not using Scala, it would be cool if we could use the newer Datomic feature of exporting the schema definition for a database in order to automatically generate a schema definition file. Since the free version of Datomic is not up to date and the export feature not available in the current older version, we can't yet automate this process and you'll have to make such schema definition file manually for now.


If you have questions or ideas, please feel free to discuss in the gitter channel:


Or you can submit issues / pull requests too.


Run all internal MoleculeAdmin test by platform:

server> server/test
server> sharedJVM/test
server> sharedJS/test

.. or run individual tests:

server> server/testOnly moleculeadmin.servertest.log.LatestTxs
server> sharedJVM/testOnly moleculeadmin.sharedtest.util.DateTransformation
server> sharedJS/testOnly moleculeadmin.sharedtest.util.DateTransformation


> sbt clean compile
> sbt sharedJS/fastOptJS


  • Big credits go to Datomic's own Console that can make all possible queries with Datalog against Datomic databases. MoleculeAdmin can make a subset of those queries, but hopefully enough to be useful, also given its capability to edit data and schema.
  • OpenRefine (earlier "GoogleRefine") - a sophisticated editor of tabular data that has served as a great inspiration for MoleculeAdmin!


Marc Grue


Molecule is licensed under the Apache License 2.0


No releases published


No packages published