Code Contribition Guideline

Code Contribition Guideline#

Presently, Wildbook is a tomcat application leveraging jsp as the frontend. We are transitioning to a react frontend with a standardized REST API framework. We are iteratively removing jsp pages. For more details on the trajectory of page updates, see the roadmap.

General#

  • DRY Principle: Do Not Repeat Yourself. Avoid code duplication by reusing components, functions, and styles. This not only minimizes the codebase but also simplifies maintenance and updates.

  • If you are adding code that will be custom to a single platform, the feature needs to be approved by the Wild Me team. If approved, add all custom code in a block at the bottom of the page to prevent merge conflicts.

Dependencies#

  • All dependencies must be on a locked version. We do not accept PRs that use the latest version of anything.

  • We use dependabot for dependency management; however, due to inconsistencies in semantic versioning and behavioral changes in packages supporting machine learning, we do system testing for all PRs suggested by dependabot. If you want to update or add a new dependency, contact the Wild Me team to discuss the changes before submitting a PR.

In-line styling#

Generally speaking, we want to avoid in-line styling to allow for greater consistency and understandability. If you have concerns with what theme styling to use, check the theme figma.

Internationalization#

All Wild Me tools are intended for an international audience, so we want to provide translated versions. An active goal for 2024 is to get ensure that all product copy has been internationalized according to the following standard.

At present, in Wildbook, all copy except for site name, notification messages, and species names are wrapped and localized to 4 languages: German, English, French, Spanish, and Italian. Language files differ in location based on whether you are contributing to a react or jsp page. For all language updates, make sure that the update is addressed across all locales.

React#

Language files are located in frontend/src/locale/. If you are adding a new string to the system, make sure the string is available across the different languages by adding the key and translated string to

  • de.json - German

  • en.json - English

  • es.json - Spanish

  • fr.json - French

  • it.json - Italian Ensure that the key is all caps with underscores as separators and is human-legible. Ex: THISKEY_IS_GOOD vs this.key.NeedsToBe-fixed vs TKNTBF

Jsp#

.properties files in the resources/bundles language folders are used to generate translations for different supported languages. If you add a new string to the system, add the string to the appropriate .properties file, and make sure the string is available across en, de, es, fr, and it. Either provide the English string in all files, or provide the translated string in each one.

Variable naming conventions#

  • Camel case

  • Don’t use single-letter variable names (no matter how temporary you think the code is)

  • Code should be clear enough to speak for itself without comments, but use your judgement on if a comment is necessary

  • Code for clarity rather than for efficiency (one-liners are cool, but not at the expense of future obfuscation)

Notes during modernization#

Because we are transitioning between tech stacks, there are some places where code is weird or where duplicate effort is necessary to maintain user experience.

  • If updating the header, you will need to update:

    • header.jsp

    • header.jsx

    • servletResponseTemplate.htm

  • If changing a dropdown in a .jsp page, use the library select 2. It provides a single line component update that adds in the ability to filter the dropdown. For example: $('#selectCode').select2({width: '100%'});

React#

Plugins used#

  • eslint-plugin-react: For React best practices.

  • eslint-plugin-react-hooks: Ensures correct use of hooks.

Relevant rules#

  • Semicolons are required (“semi”: 2).

  • JSX is allowed in both .js and .jsx files (“react/jsx-filename-extension”: 0).

  • Console statements show warnings (“no-console”: 1).

  • React Hooks rules are enforced (“react-hooks/rules-of-hooks”: “error”).

Best practices#

  • Prefer Bootstrap Classes: Always use Bootstrap’s utility classes for styling when possible. This ensures consistency across the project and leverages Bootstrap’s responsive features. Only use custom CSS or inline styles when the desired styling cannot be achieved with Bootstrap.

  • Code Consistency: We are using eslint and prettier to enforce a consistent coding style throughout the project. This includes consistent naming conventions, file and folder structure, and code formatting.

  • Componentization: Break down the UI into reusable components to maximize code reusability and clarity. Each component should have a well-defined purpose and should operate independently as much as possible.

  • State Management: Carefully manage state to avoid unnecessary re-renders. Use local state (with useState) for data that affects only one component, and consider using context or state management libraries (like Redux or MobX) for shared state across many components.

  • UseEffect Best Practices: When using useEffect, always include all dependencies in the dependency array to avoid bugs related to stale state and props. Also, be sure to return a cleanup function to avoid memory leaks, especially when subscribing to external data sources.

  • Props Validation: Use PropTypes to validate props passed to a component, or consider using TypeScript for static type checking throughout the application. This can prevent many runtime errors and improve developer productivity.

Tip and Tricks#

  • If bootstrap components are stubbornly refusing to change color, try accent-color.

Customizing copy#

If you are updating a translation for general use, do your PR against the main branch. If you are updating a language file to have specific copy associated with a specific platform, do your PR against the specific platform’s branch.

Coming changes#

10.7: 867 Linter will recognize node.js. Until then, you may need to use the following to cancel out unused reference errors from your linter

  • this line used before process.env // eslint-disable-next-line no-undef`

  • this line at the top of files with a lot of node.js calls /* eslint-disable no-undef */

Java/jsp style#

  • Initialize variables and type signatures at the abstract/interface level when possible.

Instead of:

ArrayList encounters = new ArrayList<Encounter>();
...
public int getMax(ArrayList<int> numbers) {

Try:

List encounters = new ArrayList<Encounter>();
...
public int getMax(Collection<int> numbers) {

It’s easier to read and more intuitive for a function to take a Map or List than a HashMap or ArrayList.

The List interface defines how we want that variable to behave, and whether it’s an ArrayList or LinkedList is incidental. Keeping the variable and method signatures abstract means we can change the implementation later (eg swapping ArrayList->LinkedList) without changing the rest of our code. https://stackoverflow.com/questions/2279030/type-list-vs-type-arraylist-in-java

Related: when writing utility methods, making the input type as abstract as possible makes the method versatile. See Util.asSortedList in Wildbook: since the input is an abstract Collection, it can accept a List, Set, PriorityQueue, or Vector as input, and return a sorted List.

  • Runtime (not style): Use Sets (not Lists or arrays) if you’re only keeping track of collection membership / item uniqueness.

Instead of:

    	List<MarkedIndividual> uniqueIndividuals = new ArrayList<MarkedIndividual>();
    	for(Encounter currentEncounter: encounters){
		MarkedIndividual currentInd = enc.getIndividual();
		if !(uniqueIndividuals.contains(currentInd) {
			uniqueIndividuals.add(currentInd);
			doStuff();

Try:

Set<MarkedIndividual> uniqueIndividuals = new HashSet<MarkedIndividual>();	
    	for(Encounter currentEncounter: encounters){
		MarkedIndividual currentInd = enc.getIndividual();
		if !(uniqueIndividuals.contains(currentInd) {
			uniqueIndividuals.add(currentInd);
			doStuff();

The reason is a little deep in the data types. Sets are defined as unordered collections of unique elements; and Lists/arrays are ordered collections with no bearing on element-uniqueness. If the order of a collection doesn’t matter and you’re just checking membership, you’ll have faster runtime using a Set.

Sets implement contains, add, and remove methods much faster than lists [contains is O(log(n)) vs O(n) runtime]. A list has to iterate through the entire list every time it runs contains (it checks each item once at a time) while a set (especially a HashSet) keeps track of an item index for quick lookup.

  • Use for-each loops aka “enhanced for loops” to make loops more concise and readable.

Instead of:

for (int i=0; i<encounters.length(); i++) {
	Encounter enc = encounters.get(i)
	doStuff();

try:

for (Encounter enc: encounters) {
	doStuff();

Note that in both cases you might want to check if encounters == null if relevant, but you rarely need to check if encounters.length()>0 because the for-loops take care of that.

Conversely, if you want access to the i variable for logging or otherwise, the classic for-loop is best.

  • Util.stringExists is shorthand for a common string check:

Instead of:

	if (str!=Null && !str.equals("")) {
		doStuff();

Try:

	if (Util.stringExists(str)) {
		doStuff();

This method also checks for the strings “none” and “unknown” which have given us trouble in displays in the past.

  • When dealing with any functionality that handles data persistence, make sure to use try catch exceptions. Data loss is considered a critical failure and an emergency to address.