Skip to main content

Developer documentation

Word of caution​

It is possible that the gap between the current development version and the current stable version will render this documentation obsolete in parts. It is highly recommended to use the next branch of elabftw/documentation.

Another thing is that this documentation is targeted towards GNU/Linux users. If you are on Windows or MacOS, you will need to adapt some things. We currently do not provide detailed documentation for Windows or MacOS users, as we are avid open source software aficionados, and consider these operating systems as spyware.

Introduction​

Depending on your background, the eLabFTW project might seem daunting at first, because it uses a lot of different technologies: Docker, Yarn, Webpack, Composer, TypeScript, Scss, ...

But fear not, because there is a whole documentation about getting started, and you're already reading it ;)

Note about repositories​

The eLabFTW project is split in different repositories. The main one with the actual PHP code is elabftw/elabftw. The present document is generated from markdown files in elabftw/documentation. So if you need to change the documentation, it will be in there.

The Docker image is built from the code at elabftw/elabimg.

Other interesting repositories are:

The rest of this documentation is about elabftw/elabftw.

Note about branches​

The repository contains (at least) 3 branches:

  • The master branch points to the latest released version and should always be in working state. It can contain alpha, beta or rc versions.
  • The next is used to work on bugfixes for the stable release. If you wish to make a bugfix PR, this is the branch that you should target (if hypernext is far from it).
  • The hypernext branch is the dev branch, it might contain bugs and unfinished work, never use it in production! It is the latest version of the code and the one you should work against for new features or non critical bugfixes. It is a very active branch, so make sure to pull from upstream often so your fork doesn't stray too far.

Code organization​

  • Real accessible pages are in the web/ directory (experiments.php, database.php, login.php, etc…)
  • The rest is in app/ or src/ for PHP classes
  • src/models will contain classes with CRUD (Create, Read, Update, Destroy)
  • src/classes will contain services or utility classes
  • A new class will be loaded automagically thanks to the use of PSR-4 with composer (namespace Elabftw\Elabftw)
  • app/controllers will contain pages that send actions to models (like destroy something), and generally output json for an ajax request, or redirect the user.
  • Check out the scripts in src/tools too

Working with JavaScript​

All JavaScript code is written in TypeScript <https://www.typescriptlang.org/>_ in src/ts. During build, it is converted in JS by tsc. It is then bundled by Webpack. A full build can be quite time consuming, especially on hardware with limited CPU power.

When working on some JS, what you want is to be able to save the file and immediately see the changes. For that, use yarn watchjs to build the JS and watch for changes. Now changes will take a very small time to compile and be visible.

You'll also want to configure your favorite text editor to display TypeScript errors when writing the code.

Use vanilla JS and ban the use of jQuery selectors or functions.

Miscellaneous​

  • if you make a change to the SQL structure, you need to add a schema file in src/sql. See the existing files for an example. Then increment the required version in src/classes/Update. Modify src/sql/structure.sql so new installs will get the correct structure. See also dev:genschema command.
  • comment your code wisely, what is important is the why, not the what
  • your code must follow the PSR standards <https://github.com/php-fig/fig-standards/blob/master/accepted/PSR-1-basic-coding-standard.md>_
  • add a plugin to your editor to show trailing whitespaces in red
  • add a plugin to your editor to show PSR-1 errors
  • see editorconfig.org <https://editorconfig.org/>_ and configure your editor to follow the settings from .editorconfig
  • remove BOM
  • if you want to work on the documentation, clone the documentation repo <https://github.com/elabftw/documentation>_
  • if you want to make backups of your dev install, you'll need to edit elabctl.conf to point to the correct folders/config files. See example <https://github.com/elabftw/elabctl/blob/master/elabctl.conf>_
  • in php camelCase; in html, dash separation for CSS stuff, camelCase for JS
  • check the commands in the "scripts" part of the package.json file, a lot of nice things in there ;)

Glossary​

  • Experiments + Database items + Experiments Templates = Entities. So when you see Entity it means it can be an experiment/template or a database item.

Build​

The javascript and css files are stored unminified in the source code. But the app uses the minified versions, so if you make a change to the javascript or css files, you need to rebuild them.

  • To minify files:
# install the packages first
yarn install
yarn buildall

Other commands exist, see builder.js (webpack), the scripts part of package.json (yarn). If you just want to rebuild the CSS, use yarn buildcss.

When working on the code, it is best to have yarn watchjs and yarn watchcss running so changes are immediately picked up.

Tests​

The tests run on the Codeception framework for unit and api tests. End to end testing is done with Cypress.

yarn unit # will run the unit tests
yarn test # will run the full test suite

A good contribution you can make would be adding Cypress tests.

In order for the tests to run successfully, you'll want to have a file in tests/elabftw-user.env with the following content:

ELABFTW_USER=sam
ELABFTW_GROUP=wheel
ELABFTW_USERID=1000
ELABFTW_GROUPID=1000

In the example above, the user is sam and the main group is wheel. Find out this info with id command. This file will make the test container run as your user and prevent permissions issues.

Exceptions handling​

Here are some ground rules for exceptions thrown in the code:

  • Code should not throw a generic Exception, but one of Elabftw\Exceptions
  • ImproperActionException when something forbidden happens but it's not suspicious. Error is not logged, and message is shown to user
  • DatabaseErrorException when a SQL query failed, the error is logged and message is shown to user
  • IllegalActionException when something should not happen in normal conditions unless someone is poking around by editing the requests. Error is logged and generic permission error is shown
  • FilesystemErrorException, same as DatabaseErrorException but for file operations
  • For the rest, the error is logged and a generic error message is shown to user
  • Code should throw an Exception as soon as something goes wrong
  • Exceptions should not be caught in the code (models), only in the controllers
  • Instead of returning bool, functions should throw exception if something goes wrong. This removes the need to check for return value in consuming code (something often forgotten!)

Making a pull request​

  1. Before working on a feature, it's a good idea to open an issue first to discuss its implementation
  2. Create a branch from hypernext
  3. Work on a feature
  4. Make sure yarn full exits with no errors
  5. Make a pull request on GitHub to include it in hypernext
cd $dev/elabftw
# create your feature branch from the hypernext branch
git checkout -b my-feature
# modify the code, commit and push to your fork
# go to github.com and create a pull request

Adding a lang​

  • Add lang on poeditor.com
  • Get .po
  • Save it in a new folder in src/langs
  • Open with poeditor and fix issues
  • Save the .mo
  • Upload .po fixed to poeditor
  • Add it to Enums/Language.php
  • Get the tinymce translation from this repository <https://github.com/mklkj/tinymce-i18n/tree/master/langs6>_
  • Rename file to 4 letters code in src/js/tinymce-langs
  • Edit first line of file to match code
  • Import it in tinymce.ts
  • Run bin/console dev:i18n4js
  • Import it in ts/i18n.ts

Adding a new term for js i18n​

These steps are overly complicated and should be made automatically ideally.

  • Add the new term to src/langs/js-strings.php and give it an identifier
  • Open all files in src/ts/langs/*.ts and add it there with translation for all
  • Import i18next in the corresponding ts file and use i18next.t('string-id')

Accessing Docker MySQL database with phpmyadmin​

You might be used to access your local MySQL dev database with PHPMyadmin. Just uncomment the part related to phpmyadmin in the config file and elabctl restart.

This will launch a docker container with phpmyadmin that you can reach on port 8080. Go to localhost:8080. Login with your mysql user (elabftw by default) and your mysql password found in the .yml configuration file. You should see the elabftw database now.

Using a trusted certificate for local dev​

When working locally, the docker image will generate a self-signed TLS certificate. This will show a warning in the browser address bar and multiple warnings in the console (when you press F12). To fix this, it is possible to generate certificates that are trusted by your local browser.

We'll use FiloSottile/mkcert project to achieve this.

Step 1: use a real domain name​

I like to use elab.local on port 3148. Edit /etc/hosts and add a line with elab.local pointing to localhost like this:

127.0.0.1 elab.local

Step 2: get certs​

Install mkcert and generate certificates for elab.local. Create a new folder somewhere to hold them:

mkdir -p $dev/certs/live/elab.local
mv elab.local+3.pem $dev/certs/live/elab.local/fullchain.pem
mv elab.local+3-key.pem $dev/certs/live/elab.local/privkey.pem

Step 3: edit config to use certificates​

Edit the .yml file for elabftw, change ENABLE_LETSENCRYPT to true. Uncomment the volume line with /ssl and make it point to where you have the certs.

Example:

volumes:
    - /home/user/.dev/elabftw:/elabftw
    - /home/user/.dev/certs:/ssl

Step 4: restart containers​

elabctl restart, and you should now have a valid certificate on your local dev install of elabftw :)

How to test external auth​

To easily test external authentication, edit in the container /etc/php8/php-fpm.d/elabpool.conf and at the end add:

env[auth_user] = ntesla
env[auth_username] = Nicolas
env[auth_lastname] = Tesla
env[auth_email] = "nico@example.com"
env[auth_team] = "Alpha"

Restart the php process with: s6-svc -r /var/run/s6/services/php.

Next, configure the correct keys in the Sysconfig panel and external authentication should be working as expected.

How to test ldap​

Uncomment the ldap and ldap-admin containers definitions in the config file. Then use the ldap-admin (running on port 6443 by default) to login with "cn=admin,dc=example,dc=org" and password "admin". Then click the "dc=example,dc=org" in the left menu and "Create a child entry". Create a "Generic: Posix Group". We don't care about the name but it is necessary to have one before creating our test user.

Click again the "dc=example,dc=org" in the left to be at the root, "Create a child entry" and select "Generic: User Account". In GID Number you can assign the previously created group. Once the user is created, go select it in the left menu and "Add new attribute": Email. And add the email for that user. Now you should be able to login with that user after activating ldap from the sysconfig menu. Default values from the populate script should be good to go without changes.

Install a pre-commit hook​

It is a good idea to use a pre-commit hook to run linters before the commit is actually done. It prevents doing another commit afterwards for "fix phpcs" or "fix linting". Go into .git/hooks. And cp pre-commit.sample pre-commit. Edit it and before the last line with the "exec", add this:

# eLabFTW linting pre-commit hook
reset="\e[0m"
red="\e[0;31m"
set -e
if ! docker exec elabftw yarn pre-commit
then
    printf "${red}error${reset} Pre-commit script found a problem!.\n"
    exit 1
fi

Now when you commit it should run this script and prevent the commit if there are errors.

Running cypress locally​

In docker: yarn run cy (where yarn is the local command, not the one in container, because this starts docker images)

Locally: current workaround:

cd /tmp
git clone -b hypernext https://github.com/elabftw/elabftw
cd elabftw
npm i --no-save --no-lockfile cypress cypress-terminal-report
./node_modules/.bin/cypress open
# then once it's fixed
git checkout -- yarn.lock
rm -rf node_modules

Not great, not terrible.

Debugging mysql queries​

Connect as root in the MySQL container:

docker exec -it mysql
mysql -uroot -p$MYSQL_ROOT_PASSWORD
mysql> SET GLOBAL general_log = ON;
# check where it is saved
mysql> SHOW VARIABLES WHERE Variable_name IN ('general_log','log_output','general_log_file');
# exit mysql
tail -f /var/lib/mysql/abcd1234.log

This will log ALL queries.