/**
* Returns a list of all jobs in a system.
*
* @return list of all jobs including running, done, cancelled, crashed jobs.
*/
public static Job[] all() {
List list = UKV.get(LIST);
Job[] jobs = new Job[list == null ? 0 : list._jobs.length];
int j = 0;
for (int i = 0; i < jobs.length; i++) {
Job job = UKV.get(list._jobs[i]);
if (job != null) jobs[j++] = job;
}
if (j < jobs.length) jobs = Arrays.copyOf(jobs, j);
return jobs;
}
public void set(Argument arg, String input, Object value) {
if (arg._field.getType() != Key.class && value instanceof Key) value = UKV.get((Key) value);
try {
if (arg._field.getType() == Key.class && value instanceof ValueArray)
value = ((ValueArray) value)._key;
//
else if (arg._field.getType() == int.class && value instanceof Long)
value = ((Long) value).intValue();
//
else if (arg._field.getType() == float.class && value instanceof Double)
value = ((Double) value).floatValue();
//
else if (arg._field.getType() == Frame.class && value instanceof ValueArray)
value = ((ValueArray) value).asFrame(input);
//
else if (value instanceof NumberSequence) {
double[] ds = ((NumberSequence) value)._arr;
if (arg._field.getType() == int[].class) {
int[] is = new int[ds.length];
for (int i = 0; i < is.length; i++) is[i] = (int) ds[i];
value = is;
} else value = ds;
}
arg._field.set(this, value);
} catch (Exception e) {
throw new RuntimeException(e);
}
}
@Override
protected Response serve() {
Response response = super.serve();
if (destination_key != null) {
GridSearch grid = UKV.get(destination_key);
if (grid != null) jobs = grid.jobs;
}
return response;
}
public void onException(Throwable ex) {
UKV.remove(dest());
Value v = DKV.get(progressKey());
if (v != null) {
ChunkProgress p = v.get();
p = p.error(ex.getMessage());
DKV.put(progressKey(), p);
}
cancel(ex);
}
protected String[] getVectorDomain(final Vec v) {
assert v == null || v.isInt() || v.isEnum() : "Cannot get vector domain!";
if (v == null) return null;
String[] r;
if (v.isEnum()) {
r = v.domain();
} else {
Vec tmp = v.toEnum();
r = tmp.domain();
UKV.remove(tmp._key);
}
return r;
}
@Override
protected Status exec() {
UKV.put(destination_key, this);
int max = jobs[0].gridParallelism();
int head = 0, tail = 0;
while (head < jobs.length && !cancelled()) {
if (tail - head < max && tail < jobs.length) jobs[tail++].fork();
else {
try {
jobs[head++].get();
} catch (Exception e) {
throw new RuntimeException(e);
}
}
}
return Status.Done;
}
/**
* Start this task based on given top-level fork-join task representing job computation.
*
* @param fjtask top-level job computation task.
* @return this job in {@link JobState#RUNNING} state
* @see JobState
* @see H2OCountedCompleter
*/
public
/** FIXME: should be final or at least protected */
Job start(final H2OCountedCompleter fjtask) {
assert state == JobState.CREATED : "Trying to run job which was already run?";
assert fjtask != null : "Starting a job with null working task is not permitted! Fix you API";
_fjtask = fjtask;
start_time = System.currentTimeMillis();
state = JobState.RUNNING;
// Save the full state of the job
UKV.put(self(), this);
// Update job list
new TAtomic<List>() {
@Override
public List atomic(List old) {
if (old == null) old = new List();
Key[] jobs = old._jobs;
old._jobs = Arrays.copyOf(jobs, jobs.length + 1);
old._jobs[jobs.length] = job_key;
return old;
}
}.invoke(LIST);
return this;
}
@Override
public boolean toHTML(StringBuilder sb) {
if (jobs != null) {
DocGen.HTML.arrayHead(sb);
sb.append("<tr class='warning'>");
ArrayList<Argument> args = jobs[0].arguments();
// Filter some keys to simplify UI
args = (ArrayList<Argument>) args.clone();
filter(
args,
"destination_key",
"source",
"cols",
"ignored_cols_by_name",
"response",
"classification",
"validation");
for (int i = 0; i < args.size(); i++)
sb.append("<td><b>").append(args.get(i)._name).append("</b></td>");
sb.append("<td><b>").append("run time").append("</b></td>");
String perf = jobs[0].speedDescription();
if (perf != null) sb.append("<td><b>").append(perf).append("</b></td>");
sb.append("<td><b>").append("model key").append("</b></td>");
sb.append("<td><b>").append("prediction error").append("</b></td>");
sb.append("<td><b>").append("F1 score").append("</b></td>");
sb.append("</tr>");
ArrayList<JobInfo> infos = new ArrayList<JobInfo>();
for (Job job : jobs) {
JobInfo info = new JobInfo();
info._job = job;
Object value = UKV.get(job.destination_key);
info._model = value instanceof Model ? (Model) value : null;
if (info._model != null) info._cm = info._model.cm();
if (info._cm != null) info._error = info._cm.err();
infos.add(info);
}
Collections.sort(
infos,
new Comparator<JobInfo>() {
@Override
public int compare(JobInfo a, JobInfo b) {
return Double.compare(a._error, b._error);
}
});
for (JobInfo info : infos) {
sb.append("<tr>");
for (Argument a : args) {
try {
Object value = a._field.get(info._job);
String s;
if (value instanceof int[]) s = Utils.sampleToString((int[]) value, 20);
else s = "" + value;
sb.append("<td>").append(s).append("</td>");
} catch (Exception e) {
throw new RuntimeException(e);
}
}
String runTime = "Pending", speed = "";
if (info._job.start_time != 0) {
runTime = PrettyPrint.msecs(info._job.runTimeMs(), true);
speed = perf != null ? PrettyPrint.msecs(info._job.speedValue(), true) : "";
}
sb.append("<td>").append(runTime).append("</td>");
if (perf != null) sb.append("<td>").append(speed).append("</td>");
String link = info._job.destination_key.toString();
if (info._job.start_time != 0 && DKV.get(info._job.destination_key) != null) {
if (info._model instanceof GBMModel)
link = GBMModelView.link(link, info._job.destination_key);
else if (info._model instanceof NeuralNetModel)
link = NeuralNetProgress.link(info._job.self(), info._job.destination_key, link);
if (info._model instanceof KMeans2Model)
link = KMeans2ModelView.link(link, info._job.destination_key);
else link = Inspect.link(link, info._job.destination_key);
}
sb.append("<td>").append(link).append("</td>");
String pct = "", f1 = "";
if (info._cm != null) {
pct = String.format("%.2f", 100 * info._error) + "%";
if (info._cm._arr.length == 2)
f1 = String.format("%.2f", info._cm.precisionAndRecall());
}
sb.append("<td><b>").append(pct).append("</b></td>");
sb.append("<td><b>").append(f1).append("</b></td>");
sb.append("</tr>");
}
DocGen.HTML.arrayTail(sb);
}
return true;
}
@Override
public void remove() {
super.remove();
UKV.remove(_progress);
}
public ChunkProgressJob(long chunksTotal, Key destinationKey) {
destination_key = destinationKey;
_progress =
Key.make(Key.make()._kb, (byte) 0, Key.DFJ_INTERNAL_USER, destinationKey.home_node());
UKV.put(_progress, new ChunkProgress(chunksTotal));
}
/**
* Finds a job with given key or returns null.
*
* @param jobkey job key
* @return returns a job with given job key or null if a job is not found.
*/
public static Job findJob(final Key jobkey) {
return UKV.get(jobkey);
}
/**
* Check if given job is running.
*
* @param job_key job key
* @return true if job is still running else returns false.
*/
public static boolean isRunning(Key job_key) {
Job j = UKV.get(job_key);
assert j != null : "Job should be always in DKV!";
return j.isRunning();
}
/**
* Blocks and get result of this job.
*
* <p>The call blocks on working task which was passed via {@link #start(H2OCountedCompleter)}
* method and returns the result which is fetched from UKV based on job destination key.
*
* @return result of this job fetched from UKV by destination key.
* @see #start(H2OCountedCompleter)
* @see UKV
*/
public <T> T get() {
_fjtask.join(); // Block until top-level job is done
T ans = (T) UKV.get(destination_key);
remove(); // Remove self-job
return ans;
}
/**
* Return progress of this job.
*
* @return the value in interval <0,1> representing job progress.
*/
public float progress() {
Freezable f = UKV.get(destination_key);
if (f instanceof Progress) return ((Progress) f).progress();
return 0;
}