return topDown(startNodes, InsertionOrder.FRONT);
    }

    /**
     * In top-down traversal, an ancestor node is always traversed before any of its descendant
     * nodes. The traversal order among descendant nodes (particularly aunts and nieces) are
     * determined by the {@code InsertionOrder} parameter: nieces are placed at the FRONT before
     * aunts for pre-order; while in BFS they are placed at the BACK after aunts.

 *       not look for classes in the <i>module path</i>.
 *   <li>It understands only {@code file:} URLs. This means that it does not understand <a
 *       href="https://openjdk.java.net/jeps/220">{@code jrt:/} URLs</a>, among <a
 *       href="https://github.com/classgraph/classgraph/wiki/Classpath-specification-mechanisms">others</a>.
 *   <li>It does not know how to look for classes when running under an Android VM. (ClassGraph does

When you run **multiple processes** of the same API program, they are commonly called **workers**.

### Worker Processes and Ports { #worker-processes-and-ports }

        assertTrue(extraFieldNames.contains("aaa_field"));
        assertTrue(extraFieldNames.contains("mmm_field"));
        assertTrue(extraFieldNames.contains("zzz_field"));
        // Verify sorted order among the custom fields
        final int aIdx = extraFieldNames.indexOf("aaa_field");
        final int mIdx = extraFieldNames.indexOf("mmm_field");
        final int zIdx = extraFieldNames.indexOf("zzz_field");

     * when {@link LookupContext#pushedUserProperties} is null will push user properties IF key does not already
     * exist among Java System Properties, and collects all they key it pushes. Second invocation happens AFTER
     * {@link PropertyContributor} SPI invocation, and "refreshes" already pushed user properties by re-writing them

	"regexp"
	"strconv"
	"strings"
	"time"

	"gorm.io/gorm"
	"gorm.io/gorm/clause"
	"gorm.io/gorm/logger"
	"gorm.io/gorm/schema"
)

// This regular expression seeks to find a sequence of digits (\d+) among zero or more non-digit characters (\D*),
// with a possible trailing non-digit character (\D?).

// For example, values that can pass this regular expression are:
// - "123"
// - "abc456"
// -"%$#@789"

That, plus the simple fact that Python is the main language for **Data Science**, Machine Learning and especially Deep Learning, make FastAPI a very good match for Data Science / Machine Learning web APIs and applications (among many others).

To see how to achieve this parallelism in production see the section about [Deployment](deployment/index.md).

## `async` and `await` { #async-and-await }

	//
	// We also set `DenyOnly` arg to false here - this is an IMPORTANT corner
	// case: DenyOnly is used only for allowing an account to do actions related
	// to its own account (like create service accounts for itself, among
	// others). However when a session policy is present, we need to validate
	// that the action is actually allowed, rather than checking if the action
	// is only disallowed.
	sessionPolicyArgs := args

As this component would take the **load** of requests and distribute that among the workers in a (hopefully) **balanced** way, it is also commonly called a **Load Balancer**.

/// tip

The same **TLS Termination Proxy** component used for HTTPS would probably also be a **Load Balancer**.

///

 * Thread2: acquire(LockB) --X acquire(LockA)
 * </pre>
 *
 * <p>Neither thread will progress because each is waiting for the other. In more complex
 * applications, cycles can arise from interactions among more than 2 locks:
 *
 * <pre>
 * Thread1: acquire(LockA) --X acquire(LockB)
 * Thread2: acquire(LockB) --X acquire(LockC)
 * ...
 * ThreadN: acquire(LockN) --X acquire(LockA)
 * </pre>
 *