* will typically have been set to an appropriate loader for the current thread.
   *
   * <p>In the unusual case where the context class loader is null, the class loader that loaded
   * this class ({@code Resources}) will be used instead.
   *
   * @throws IllegalArgumentException if the resource is not found
   */
  @CanIgnoreReturnValue // being used to check if a resource exists

## Enabling Parquet Format

Parquet is DISABLED by default since hostile crafted input can easily crash the server.

If you are in a controlled environment where it is safe to assume no hostile content can be uploaded to your cluster you can safely enable Parquet.
To enable Parquet set the environment variable `MINIO_API_SELECT_PARQUET=on`.

## Example using Python API

### 1. Prerequisites

	var opts ObjectOptions

	err = objLayer.MakeBucket(ctx, bucket, MakeBucketOptions{})
	if err != nil {
		t.Fatalf("Failed to make a bucket - %v", err)
	}

	// Create an object with multiple parts uploaded in decreasing
	// part number.
	res, err := objLayer.NewMultipartUpload(ctx, bucket, object, opts)
	if err != nil {
		t.Fatalf("Failed to create a multipart upload - %v", err)
	}

        assertNotNull(hash);
        // For lmCompatibility 3,4,5 (NTLMv2), returns empty array
        assertEquals(0, hash.length);
    }

    // Test that changing lmCompatibility at runtime doesn't affect already loaded static value
    @ParameterizedTest
    @ValueSource(strings = { "0", "1", "2", "3", "4", "5" })
    void testLmCompatibilityStaticInitialization(String lmCompatibility) {

After that, all of the processing logic is the same.

But because of our changes in `GzipRequest.body`, the request body will be automatically decompressed when it is loaded by **FastAPI** when needed.

## Accessing the request body in an exception handler { #accessing-the-request-body-in-an-exception-handler }

/// tip

    return statsCounterSupplier;
  }

  /**
   * Builds a cache, which either returns an already-loaded value for a given key or atomically
   * computes or retrieves it using the supplied {@code CacheLoader}. If another thread is currently
   * loading the value for this key, simply waits for that thread to finish and returns its loaded
   * value. Note that multiple threads can concurrently load values for distinct keys.
   *

          }
          // TODO(cpovirk): if reference has been cleared, remove it?
        }
        checkExceptionClassValidity(exceptionClass);

        /*
         * It's very unlikely that any loaded Futures class will see getChecked called with more
         * than a handful of exceptions. But it seems prudent to set a cap on how many we'll cache.

                    if (!info.isExpired()) {
                        handles.put(info.getPath(), info);
                        guidToHandle.put(info.getCreateGuid(), info);
                        log.info("Loaded persisted handle: {}", info.getPath());
                    } else {
                        Files.delete(handleFile);
                        log.debug("Deleted expired persisted handle: {}", handleFile);

That's what we'll solve, let's load the model before the requests are handled, but only right before the application starts receiving requests, not while  the code is being loaded.

## Lifespan { #lifespan }

You can define this *startup* and *shutdown* logic using the `lifespan` parameter of the `FastAPI` app, and a "context manager" (I'll show you what that is in a second).

<div class="termy">

```console
$ uvicorn main:app --reload

<span style="color: green;">INFO</span>:     Uvicorn running on http://127.0.0.1:8000 (Press CTRL+C to quit)
<span style="color: green;">INFO</span>:     Started reloader process [28720]
<span style="color: green;">INFO</span>:     Started server process [28722]
<span style="color: green;">INFO</span>:     Waiting for application startup.