* memory, so it is only likely to be worthwhile if the precomputed matcher is queried very often.
   *
   * <p>This method has no effect (returns {@code this}) when called in GWT: it's unclear whether a
   * precomputed matcher is faster, but it certainly would consume more memory (which doesn't seem
   * like a worthwhile tradeoff in a browser).
   */
  public CharMatcher precomputed() {
    return Platform.precomputeCharMatcher(this);
  }

/**
 * @author Jesse Wilson
 */
final class Platform {
  static CharMatcher precomputeCharMatcher(CharMatcher matcher) {
    // CharMatcher.precomputed() produces CharMatchers that are maybe a little
    // faster (and that's debatable), but definitely more memory-hungry. We're
    // choosing to turn .precomputed() into a no-op in GWT, because it doesn't
    // seem to be a worthwhile tradeoff in a browser.
    return matcher;
  }

 *
 * <h3>Examples</h3>
 *
 * <p>To compute the median:
 *
 * {@snippet :
 * double myMedian = median().compute(myDataset);
 * }
 *
 * where {@link #median()} has been statically imported.
 *
 * <p>To compute the 99th percentile:
 *
 * {@snippet :
 * double myPercentile99 = percentiles().index(99).compute(myDataset);
 * }
 *

from fastapi import APIRouter, Body

router = APIRouter()


@router.post("/compute/")
def compute(a: int = Body(), b: str = Body()):

  TF_Status* s = TF_NewStatus();
  TF_OpKernelConstruction_GetAttrType(ctx, "foobar", &type, s);
  TF_DeleteStatus(s);
  return NULL;
}

// A compute function. This will never actually get called in this test, it's
// just nice to know that it compiles.
void compute(void* kernel, TF_OpKernelContext* ctx) {
  TF_Tensor* input;
  TF_Status* s = TF_NewStatus();
  TF_GetInput(ctx, 0, &input, s);
  TF_DeleteTensor(input);

export DEBIAN_FRONTEND=noninteractive

# Fetch the NVIDIA key.
apt-key adv --fetch-keys https://developer.download.nvidia.com/compute/cuda/repos/ubuntu2204/x86_64/3bf863cc.pub;

# Set up sources for NVIDIA CUDNN.
cat >/etc/apt/sources.list.d/nvidia.list <<SOURCES
# NVIDIA
deb https://developer.download.nvidia.com/compute/cuda/repos/ubuntu2204/x86_64/ /

}


fun Test.includeSpockAnnotation(fqcn: String) {
    systemProperties.compute("include.spock.annotation") { _, oldValue ->
        if (oldValue == null) fqcn else "$oldValue,$fqcn"
    }
}


fun Test.excludeSpockAnnotation(fqcn: String) {
    systemProperties.compute("exclude.spock.annotation") { _, oldValue ->
        if (oldValue == null) fqcn else "$oldValue,$fqcn"
    }

    // Compute the filter.
    long filter = 0;
    int size = chars.cardinality();
    boolean containsZero = chars.get(0);
    // Compute the hash table.
    char[] table = new char[chooseTableSize(size)];
    int mask = table.length - 1;
    for (int c = chars.nextSetBit(0); c != -1; c = chars.nextSetBit(c + 1)) {
      // Compute the filter at the same time.
      filter |= 1L << c;

    // Compute the filter.
    long filter = 0;
    int size = chars.cardinality();
    boolean containsZero = chars.get(0);
    // Compute the hash table.
    char[] table = new char[chooseTableSize(size)];
    int mask = table.length - 1;
    for (int c = chars.nextSetBit(0); c != -1; c = chars.nextSetBit(c + 1)) {
      // Compute the filter at the same time.
      filter |= 1L << c;

        return (logX2Floor < 2 * logFloor + 1) ? logFloor : logFloor + 1;
    }
    throw new AssertionError();
  }

  /*
   * The maximum number of bits in a square root for which we'll precompute an explicit half power
   * of two. This can be any value, but higher values incur more class load time and linearly
   * increasing memory consumption.
   */