internal static ExecutionReport Export(string fileName, FunctionOutput <Output <int, Peak, PeakData> > results, string header, string separator = "\t")
{
int intervalCount = 0;
_stopWatch.Restart();
if (!File.Exists(fileName))
{
File.Delete(fileName);
}
using (File.Create(fileName)) { }
using (var writter = new StreamWriter(fileName))
{
writter.WriteLine(header);
foreach (var chr in results.Chrs)
{
foreach (var strand in chr.Value)
{
foreach (var interval in strand.Value)
{
writter.WriteLine(
chr.Key + separator +
interval.interval.left.ToString() + separator +
interval.interval.right.ToString() + separator +
interval.count.ToString() + separator +
strand.Key);
intervalCount++;
}
}
}
}
_stopWatch.Stop();
return(new ExecutionReport(intervalCount, _stopWatch.Elapsed));
}
public FunctionOutput<string>[] FastReplace(Excel.Range com, DAG dag, InputSample original, InputSample sample, AST.Address[] outputs, bool replace_original)
{
FunctionOutput<string>[] fo_arr;
if (!_d.TryGetValue(sample, out fo_arr))
{
// replace the COM value
ReplaceExcelRange(com, sample);
// initialize array
fo_arr = new FunctionOutput<string>[outputs.Length];
// grab all outputs
for (var k = 0; k < outputs.Length; k++)
{
// save the output
fo_arr[k] = new FunctionOutput<string>(dag.readCOMValueAtAddress(outputs[k]), sample.GetExcludes());
}
// Add function values to cache
// Don't care about return value
_d.Add(sample, fo_arr);
// restore the COM value
if (replace_original)
{
ReplaceExcelRange(com, original);
}
}
return fo_arr;
}
public bool Cover(string[] args)
{
if (args.Length < 6)
{
Herald.Announce(Herald.MessageType.Error, string.Format("Missing parameter."));
return(false);
}
char strand;
int minAcc, maxAcc;
string coverOrSummit = args[0].ToLower();
string resultFile = "";
if (!ExtractResultsFile(args[1], out resultFile))
{
return(false); // invalid file URI.
}
if (!char.TryParse(args[2], out strand))
{
Herald.Announce(Herald.MessageType.Error, string.Format("Invalid strand parameter."));
return(false);
}
if (!int.TryParse(args[3], out minAcc))
{
Herald.Announce(Herald.MessageType.Error, string.Format("Invalid minimum accumulation parameter."));
return(false);
}
if (!int.TryParse(args[4], out maxAcc))
{
Herald.Announce(Herald.MessageType.Error, string.Format("Invalid maximum accumulation parameter."));
return(false);
}
Aggregate agg = Aggregate.Count;
if (!String2Aggregate(args[5], out agg))
{
return(false);
}
FunctionOutput <Output <int, Peak, PeakData> > result = null;
switch (coverOrSummit)
{
case "cover":
Herald.AnnounceExeReport("Cover", di4B.Cover(CoverVariation.Cover, strand, minAcc, maxAcc, agg, out result, _maxDegreeOfParallelism), Herald.SpeedUnit.bookmarkPerSecond);
break;
case "summit":
Herald.AnnounceExeReport("Summit", di4B.Cover(CoverVariation.Summit, strand, minAcc, maxAcc, agg, out result, _maxDegreeOfParallelism), Herald.SpeedUnit.bookmarkPerSecond);
break;
}
Herald.AnnounceExeReport("Export", Exporter.Export(resultFile, result, "chr\tleft\tright\tcount\tstrand"));
return(true);
}
private FunctionOutput ProcessOutput(FunctionOutput output)
{
if (!IsRationalToIncludeMetaData(output))
{
return(output);
}
return(IncludeMetaData(output));
}
// attempts to convert all of the bootstraps for FunctionOutput[function_idx, input_idx, _] to doubles
public static FunctionOutput<double>[] ConvertToNumericOutput(FunctionOutput<string>[] boots)
{
var fi_boots = new FunctionOutput<double>[boots.Length];
for (int b = 0; b < boots.Length; b++)
{
FunctionOutput<string> boot = boots[b];
double value = System.Convert.ToDouble(boot.GetValue());
fi_boots[b] = new FunctionOutput<double>(value, boot.GetExcludes());
}
return fi_boots;
}
private FunctionOutput IncludeMetaData(FunctionOutput output)
{
output.MetaData = new Dictionary <string, object>();
output.MetaData["resultType"] = RpcDataType.ResolveRpcDataType(output.Result);
if (RpcService.Configuration.DebugMode)
{
output.MetaData["clrResultType"] = output.Result?.GetType().FullName;
output.MetaData["internalException"] = internalException?.GetType().FullName;
output.MetaData["exceptionStackTrace"] = internalException?.StackTrace;
}
return(output);
}
private void UseFunctionResults(IReadOnlyList <Result> traceResults)
{
#region Using Function Results
// Get the FunctionOutputResult from the trace results
FunctionOutputResult functionOutputResult = traceResults.OfType <FunctionOutputResult>().First();
// First() can be used here if only one Function was included in the TraceConfiguration.Functions collection.
// Otherwise you will have to search the list for the correct FunctionOutput object.
FunctionOutput functionOutput = functionOutputResult.FunctionOutputs.First();
// Extract the total load from the GlobalValue property
double totalLoad = (double)functionOutput.Value;
#endregion
}
private async Task <FunctionOutput> TryCallFunction(MethodInfo callMethod, object[] args)
{
object result;
try
{
if (functionInfo.IsCallMethodAsync)
{
result = await ExecuteFunctionAsync(callMethod, args);
}
else
{
result = ExecuteFunctionSync(callMethod, args);
}
}
catch (ArgumentException)
{
return(FunctionOutput.WithError(RpcErrorCode.InvalidArguments));
}
catch (TargetParameterCountException)
{
return(FunctionOutput.WithError(RpcErrorCode.InvalidArguments));
}
catch (FunctionException ex)
{
return(FunctionOutput.WithError(ex.Code));
}
catch (TargetInvocationException ex)
{
if (ex.InnerException is FunctionException fEx)
{
return(FunctionOutput.WithError(fEx.Code));
}
internalException = ex.InnerException;
return(FunctionOutput.WithError(RpcErrorCode.InternalServerError));
}
catch (Exception ex)
{
internalException = ex;
return(FunctionOutput.WithError(RpcErrorCode.InternalServerError));
}
return(FunctionOutput.WithResult(result));
}
public async Task <FunctionOutput> Post()
{
Prepare();
Call = await GetFunctionCall();
var callMethod = functionInfo.CallMethod;
if (functionInfo.RequiresAuthorization && !ValidateAuthorizationToken())
{
return(FunctionOutput.WithError(RpcErrorCode.AuthorizationError));
}
var args = GetArguments(callMethod);
var output = await TryCallFunction(callMethod, args);
return(ProcessOutput(output));
}
private void BenchmarkCover()
{
int tries = 10;
var agg = Aggregate.Count;
FunctionOutput <Output <int, Peak, PeakData> > result = null;
var combinations = new List <int[]>
{
new int[] { 1, 2 },
new int[] { 5, 10 },
new int[] { 10, 20 },
new int[] { 50, 60 },
new int[] { 80, 90 },
new int[] { 100, 200 },
new int[] { 200, 220 },
new int[] { 300, 320 },
new int[] { 400, 500 },
new int[] { 550, 1000 },
new int[] { 1, 1 },
new int[] { 10, 10 },
new int[] { 25, 25 },
new int[] { 50, 50 },
new int[] { 60, 60 },
new int[] { 80, 80 },
new int[] { 100, 100 },
new int[] { 150, 150 },
new int[] { 200, 200 },
new int[] { 500, 500 },
};
foreach (var c in combinations)
{
for (int i = 0; i < tries; i++)
{
Herald.AnnounceExeReport(c[0] + "," + c[1], di4B.Cover(CoverVariation.Cover, '*', c[0], c[1], agg, out result, _maxDegreeOfParallelism), Herald.SpeedUnit.bookmarkPerSecond);
}
}
}
// Sort numeric bootstrap values
public static FunctionOutput<double>[] SortBootstraps(FunctionOutput<double>[] boots)
{
return boots.OrderBy(b => b.GetValue()).ToArray();
}
private async Task SendErrorOutput(HttpResponse response, RpcErrorCode code)
{
response.ContentType = "application/json";
await response.WriteAsync(JsonConvert.SerializeObject(FunctionOutput.WithError(code)));
}
public void threadPoolCallback(Object threadContext)
{
// perform hypothesis tests
hypothesisTests();
// OK to dealloc fields; this object lives on because it is
// needed for job control
_bs = null;
_initial_outputs = null;
_input = null;
_outputs = null;
// notify
_mre.Set();
}
public DataDebugJob(
DAG dag,
FunctionOutput<String>[][] bs,
Dictionary<AST.Address, string> initial_outputs,
AST.Range input,
AST.Address[] output_arr,
bool weighted,
double significance,
ManualResetEvent mre)
{
_dag = dag;
_bs = bs;
_initial_outputs = initial_outputs;
_input = input;
_outputs = output_arr;
_weighted = weighted;
_significance = significance;
_mre = mre;
_score = new TreeScore();
}
public static TreeScore StringHypothesisTest(DAG dag, AST.Range rangeNode, AST.Address functionNode, FunctionOutput<string>[] boots, string initial_output, bool weighted, double significance)
{
// this function's input cells
var input_cells = rangeNode.Addresses();
// scores
var iexc_scores = new TreeScore();
var inputs_sz = input_cells.Count();
// exclude each index, in turn
for (int i = 0; i < inputs_sz; i++)
{
// default weight
int weight = 1;
// add weight to score if test fails
AST.Address xtree = input_cells[i];
if (weighted)
{
// the weight of the function value of interest
weight = dag.getWeight(functionNode);
}
if (RejectNullHypothesis(boots, initial_output, i, significance))
{
if (iexc_scores.ContainsKey(xtree))
{
iexc_scores[xtree] += weight;
}
else
{
iexc_scores.Add(xtree, weight);
}
}
else
{
// we need to at least add the value to the tree
if (!iexc_scores.ContainsKey(xtree))
{
iexc_scores.Add(xtree, 0);
}
}
}
return iexc_scores;
}
public FunctionOutput Get()
{
return(FunctionOutput.WithError(RpcErrorCode.PreCallChecksFailed));
}
// are all of the values numeric?
public static bool FunctionOutputsAreNumeric(FunctionOutput<string>[] boots)
{
for (int i = 0; i < boots.Length; i++)
{
double d;
if (!Double.TryParse(boots[i].GetValue(), out d))
{
return false;
}
};
return true;
}
// Exclude specified input index, compute multinomial probabilty vector, and return true if probability is below threshold
public static bool RejectNullHypothesis(FunctionOutput<string>[] boots, string original_output, int exclude_index, double significance)
{
// get bootstrap fingerprint for exclude_index
var xfp = BigInteger.One << exclude_index;
// filter bootstraps which include exclude_index
var boots_exc = boots.Where(b => (b.GetExcludes() & xfp) == xfp);
// get p_value vector
var freq = BootstrapFrequency(boots_exc);
// what is the probability of seeing the original output?
double p_val;
if (!freq.TryGetValue(original_output, out p_val))
{
p_val = 0.0;
}
// test H_0
return p_val < 1.0 - significance;
}
public static TreeScore NumericHypothesisTest(DAG dag, AST.Range rangeNode, AST.Address functionNode, FunctionOutput<string>[] boots, string initial_output, bool weighted, double significance)
{
// this function's input cells
var input_cells = rangeNode.Addresses();
var inputs_sz = input_cells.Count();
// scores
var input_exclusion_scores = new TreeScore();
// convert to numeric
var numeric_boots = ConvertToNumericOutput(boots);
// sort
var sorted_num_boots = SortBootstraps(numeric_boots);
// for each excluded index, test whether the original input
// falls outside our bootstrap confidence bounds
for (int i = 0; i < inputs_sz; i++)
{
// default weight
int weight = 1;
// add weight to score if test fails
AST.Address xtree = input_cells[i];
if (weighted)
{
// the weight of the function value of interest
weight = dag.getWeight(functionNode);
}
double outlieriness = RejectNullHypothesis(sorted_num_boots, initial_output, i, significance);
if (outlieriness != 0.0)
{
// get the xth indexed input in input_rng i
if (input_exclusion_scores.ContainsKey(xtree))
{
input_exclusion_scores[xtree] += (int)(weight * outlieriness);
}
else
{
input_exclusion_scores.Add(xtree, (int)(weight * outlieriness));
}
}
else
{
// we need to at least add the value to the tree
if (!input_exclusion_scores.ContainsKey(xtree))
{
input_exclusion_scores.Add(xtree, 0);
}
}
}
return input_exclusion_scores;
}
public static TreeScore Inference(
int num_bootstraps,
InputSample[][] resamples,
Dictionary<AST.Range, InputSample> initial_inputs,
Dictionary<AST.Address, string> initial_outputs,
AST.Range[] input_arr,
AST.Address[] output_arr,
DAG dag,
bool weighted,
double significance,
ProgBar pb)
{
// synchronization token
object lock_token = new Object();
// init thread event notification array
var mres = new ManualResetEvent[input_arr.Length];
// init job storage
var ddjs = new DataDebugJob[input_arr.Length];
// init started jobs count
var sjobs = 0;
// init completed jobs count
var cjobs = 0;
// last-ditch effort flag
bool last_try = false;
// init score storage
var scores = new TreeScore();
for (int i = 0; i < input_arr.Length; i++)
{
try
{
#region BOOTSTRAP
// bootstrapping is done in the parent STA thread because
// the .NET threading model prohibits thread pools (which
// are MTA) from accessing STA COM objects directly.
// alloc bootstrap storage for each output (f), for each resample (b)
FunctionOutput<string>[][] bs = new FunctionOutput<string>[initial_outputs.Count][];
for (int f = 0; f < initial_outputs.Count; f++)
{
bs[f] = new FunctionOutput<string>[num_bootstraps];
}
// init memoization table for input vector i
var memo = new BootMemo();
// fetch the input range TreeNode
var input = input_arr[i];
// fetch the input range COM object
var com = dag.getCOMRefForRange(input).Range;
// compute outputs
// replace the values of the COM object with the jth bootstrap,
// save all function outputs, and
// restore the original input
for (var b = 0; b < num_bootstraps; b++)
{
// lookup outputs from memo table; otherwise do replacement, compute outputs, store them in table, and return them
FunctionOutput<string>[] fos = memo.FastReplace(com, dag, initial_inputs[input], resamples[i][b], output_arr, false);
for (var f = 0; f < output_arr.Length; f++)
{
bs[f][b] = fos[f];
}
}
// restore the original inputs; faster to do once, after bootstrapping is done
BootMemo.ReplaceExcelRange(com, initial_inputs[input]);
// TODO: restore formulas if it turns out that they were overwrittern
// this should never be the case
#endregion BOOTSTRAP
#region HYPOTHESIS_TEST
// cancellation token
mres[i] = new ManualResetEvent(false);
// set up job
ddjs[i] = new DataDebugJob(
dag,
bs,
initial_outputs,
input_arr[i],
output_arr,
weighted,
significance,
mres[i]
);
sjobs++;
// hand job to thread pool
ThreadPool.QueueUserWorkItem(ddjs[i].threadPoolCallback, i);
#endregion HYPOTHESIS_TEST
// update progress bar
pb.IncrementProgress();
}
catch (System.OutOfMemoryException e)
{
if (!last_try)
{
// If there are no more jobs running, but
// we still can't allocate memory, try invoking
// GC and then trying again
cjobs = mres.Count(mre => mre.WaitOne(0));
if (sjobs - cjobs == 0)
{
GC.Collect();
last_try = true;
}
}
else
{
// we just don't have enough memory
throw e;
}
// wait for any of the 0..i-1 work items
// to complete and try again
WaitHandle.WaitAny(mres.Take(i).ToArray());
}
}
// Do not proceed until all hypothesis tests are done.
// WaitHandle.WaitAll cannot be called on an STA thread which
// is why we call WaitOne in a loop.
// Merge scores as data becomes available.
for (int i = 0; i < input_arr.Length; i++)
{
mres[i].WaitOne();
scores = DictAdd(scores, ddjs[i].Result);
}
return scores;
}
// Exclude a specified input index, compute quantiles, and check position of original input
public static double RejectNullHypothesis(FunctionOutput<double>[] boots, string original_output, int exclude_index, double significance)
{
// low
double low_thresh = (1.0 - significance) / 2.0;
// high
double hi_thresh = significance + low_thresh;
// get bootstrap fingerprint for exclude_index
var xfp = BigInteger.One << exclude_index;
// filter bootstraps that include exclude_index
var boots_exc = boots.Where(b => (b.GetExcludes() & xfp) == xfp);
var exc_count = boots_exc.Count();
// return neutral (0.5) if we are having a sparsity problem
if (exc_count == 0)
{
return 0.5;
}
// index for value greater than 2.5% of the lowest values; we want to round down here
var low_index = System.Convert.ToInt32(Math.Floor((float)(exc_count - 1) * low_thresh));
// index for value greater than 97.5% of the lowest values; we want to round up here
var high_index = System.Convert.ToInt32(Math.Ceiling((float)(exc_count - 1) * hi_thresh));
var low_value = boots_exc.ElementAt(low_index).GetValue();
var high_value = boots_exc.ElementAt(high_index).GetValue();
var lowest_value = boots_exc.ElementAt(0).GetValue();
var highest_value = boots_exc.ElementAt(exc_count - 1).GetValue();
double original_output_d;
Double.TryParse(original_output, out original_output_d);
// truncate the values to deal with floating point imprecision
var low_value_tr = Math.Truncate(low_value * 10000) / 10000;
var high_value_tr = Math.Truncate(high_value * 10000) / 10000;
var original_tr = Math.Truncate(original_output_d * 10000) / 10000;
var lowest_value_tr = Math.Truncate(lowest_value * 10000) / 10000;
var highest_value_tr = Math.Truncate(highest_value * 10000) / 10000;
// reject or fail to reject H_0
if (original_tr > high_value_tr)
{
if (highest_value_tr != high_value_tr)
{
return Math.Abs((original_tr - high_value_tr) / Math.Abs(high_value_tr - highest_value_tr)); //normalize by the highest 2.5%
}
else //can't normalize
{
return Math.Abs(original_tr - high_value_tr);
}
}
else if (original_tr < low_value_tr)
{
if (lowest_value_tr != low_value_tr)
{
return Math.Abs((original_tr - low_value_tr) / Math.Abs(low_value_tr - lowest_value_tr)); //normalize by the lowest 2.5%
}
else //can't normalize
{
return Math.Abs(original_tr - low_value_tr);
}
}
return 0.0;
}
private bool IsRationalToIncludeMetaData(FunctionOutput output)
{
return(RpcService.Configuration.AlwaysIncludeMetadata ||
RpcService.Configuration.DebugMode ||
output.Result is JwtAuthorizationTicket);
}