public static PrepData PrepSimulation(Excel.Application app, Excel.Workbook wbh, ProgBar pb, bool ignore_parse_errors)
{
// build graph
var dag = new DAG(wbh, app, ignore_parse_errors);
if (dag.containsLoop())
{
throw new DataDebugMethods.ContainsLoopException();
}
pb.IncrementProgress();
// get terminal input and terminal formula nodes once
var terminal_input_nodes = dag.terminalInputVectors();
var terminal_formula_nodes = dag.terminalFormulaNodes(true); ///the boolean indicates whether to use all outputs or not
if (terminal_input_nodes.Length == 0)
{
throw new NoRangeInputs();
}
if (terminal_formula_nodes.Length == 0)
{
throw new NoFormulas();
}
// save original spreadsheet state
CellDict original_inputs = UserSimulation.Utility.SaveInputs(dag);
// force a recalculation before saving outputs, otherwise we may
// erroneously conclude that the procedure did the wrong thing
// based solely on Excel floating-point oddities
UserSimulation.Utility.InjectValues(app, wbh, original_inputs);
// save function outputs
CellDict correct_outputs = UserSimulation.Utility.SaveOutputs(terminal_formula_nodes, dag);
return new PrepData()
{
dag = dag,
original_inputs = original_inputs,
correct_outputs = correct_outputs,
terminal_input_nodes = terminal_input_nodes,
terminal_formula_nodes = terminal_formula_nodes
};
}
// this function returns an address but also updates
// the filtered_high_scores list
public static AST.Address CheckCell_Step(UserResults o,
double significance,
CutoffKind ck,
int nboots,
DAG dag,
Excel.Application app,
bool weighted,
bool all_outputs,
bool run_bootstrap,
HashSet<AST.Address> known_good,
ref List<KeyValuePair<AST.Address, int>> filtered_high_scores,
long max_duration_in_ms,
Stopwatch sw,
ProgBar pb)
{
// Get bootstraps
// The bootstrap should only re-run if there is a correction made,
// not when something is marked as OK (isn't one of the introduced errors)
// The list of suspected cells doesn't change when we mark something as OK,
// we just move on to the next thing in the list
if (run_bootstrap)
{
TreeScore scores = Analysis.DataDebug(nboots, dag, app, weighted, all_outputs, max_duration_in_ms, sw, significance, pb);
// apply a threshold to the scores
filtered_high_scores = ck.applyCutoff(scores, known_good);
}
else //if no corrections were made (a cell was marked as OK, not corrected)
{
//re-filter out cells marked as OK
filtered_high_scores = filtered_high_scores.Where(kvp => !known_good.Contains(kvp.Key)).ToList();
}
if (filtered_high_scores.Count() != 0)
{
// get AST.Address corresponding to most unusual score
return filtered_high_scores[0].Key;
}
else
{
return null;
}
}
public static void RunProportionExperiment(Excel.Application app, Excel.Workbook wbh, int nboots, double significance, double threshold, UserSimulation.Classification c, Random r, String outfile, long max_duration_in_ms, String logfile, ProgBar pb, bool ignore_parse_errors)
{
pb.setMax(5);
// record intitial state of spreadsheet
var prepdata = Prep.PrepSimulation(app, wbh, pb, ignore_parse_errors);
// init error generator
var eg = new ErrorGenerator();
// get inputs as an array of addresses to facilitate random selection
// DATA INPUTS ONLY
AST.Address[] inputs = prepdata.dag.terminalInputCells();
// sanity check: all of the inputs should also be in prepdata.original_inputs
foreach (AST.Address addr in inputs)
{
if (!prepdata.original_inputs.ContainsKey(addr))
{
throw new Exception("Missing address!");
}
}
for (int i = 0; i < 100; i++)
{
// randomly choose an input address
AST.Address rand_addr = inputs[r.Next(inputs.Length)];
// get the value
String input_value = prepdata.original_inputs[rand_addr];
// perturb it
String erroneous_input = eg.GenerateErrorString(input_value, c);
// create an error dictionary with this one perturbed value
var errors = new CellDict();
errors.Add(rand_addr, erroneous_input);
// run simulations; simulation code does insertion of errors and restore of originals
RunSimulation(app, wbh, nboots, significance, threshold, c, r, outfile, max_duration_in_ms, logfile, pb, prepdata, errors);
}
}
public static void RunSimulation(Excel.Application app, Excel.Workbook wbh, int nboots, double significance, double threshold, UserSimulation.Classification c, Random r, String outfile, long max_duration_in_ms, String logfile, ProgBar pb, PrepData prepdata, CellDict errors)
{
// write header if needed
if (!System.IO.File.Exists(outfile))
{
System.IO.File.AppendAllText(outfile, Simulation.HeaderRowForCSV());
}
// CheckCell weighted, all outputs, quantile
//var s_1 = new UserSimulation.Simulation();
//s_1.RunFromBatch(nboots, // number of bootstraps
// wbh.FullName, // Excel filename
// significance, // statistical significance threshold for hypothesis test
// app, // Excel.Application
// new QuantileCutoff(0.05), // max % extreme values to flag
// c, // classification data
// r, // random number generator
// UserSimulation.AnalysisType.CheckCell5,// analysis type
// true, // weighted analysis
// true, // use all outputs for analysis
// prepdata.graph, // AnalysisData
// wbh, // Excel.Workbook
// errors, // pre-generated errors
// prepdata.terminal_input_nodes, // input range nodes
// prepdata.terminal_formula_nodes, // output nodes
// prepdata.original_inputs, // original input values
// prepdata.correct_outputs, // original output values
// max_duration_in_ms, // max duration of simulation
// logfile);
//System.IO.File.AppendAllText(outfile, s_1.FormatResultsAsCSV());
pb.IncrementProgress();
// CheckCell weighted, all outputs, quantile
var s_4 = new UserSimulation.Simulation();
s_4.RunFromBatch(nboots, // number of bootstraps
wbh.FullName, // Excel filename
significance, // statistical significance of threshold
app, // Excel.Application
new QuantileCutoff(0.10), // max % extreme values to flag
c, // classification data
r, // random number generator
UserSimulation.AnalysisType.CheckCell10,// analysis type
true, // weighted analysis
true, // use all outputs for analysis
prepdata.dag, // AnalysisData
wbh, // Excel.Workbook
errors, // pre-generated errors
prepdata.terminal_input_nodes, // input range nodes
prepdata.terminal_formula_nodes, // output nodes
prepdata.original_inputs, // original input values
prepdata.correct_outputs, // original output values
max_duration_in_ms, // max duration of simulation
logfile);
System.IO.File.AppendAllText(outfile, s_4.FormatResultsAsCSV());
pb.IncrementProgress();
// Normal, all inputs
var s_2 = new UserSimulation.Simulation();
s_2.RunFromBatch(nboots, // irrelevant
wbh.FullName, // Excel filename
significance, // normal cutoff?
app, // Excel.Application
new NormalCutoff(threshold), // ??
c, // classification data
r, // random number generator
UserSimulation.AnalysisType.NormalAllInputs, // analysis type
true, // irrelevant
true, // irrelevant
prepdata.dag, // AnalysisData
wbh, // Excel.Workbook
errors, // pre-generated errors
prepdata.terminal_input_nodes, // input range nodes
prepdata.terminal_formula_nodes, // output nodes
prepdata.original_inputs, // original input values
prepdata.correct_outputs, // original output values
max_duration_in_ms, // max duration of simulation
logfile);
System.IO.File.AppendAllText(outfile, s_2.FormatResultsAsCSV());
pb.IncrementProgress();
// Normal, range inputs
//var s_3 = new UserSimulation.Simulation();
//s_3.RunFromBatch(nboots, // irrelevant
// wbh.FullName, // Excel filename
// significance, // normal cutoff?
// app, // Excel.Application
// new NormalCutoff(threshold), // ??
// c, // classification data
// r, // random number generator
// UserSimulation.AnalysisType.NormalPerRange, // analysis type
// true, // irrelevant
// true, // irrelevant
// prepdata.graph, // AnalysisData
// wbh, // Excel.Workbook
// errors, // pre-generated errors
// prepdata.terminal_input_nodes, // input range nodes
// prepdata.terminal_formula_nodes, // output nodes
// prepdata.original_inputs, // original input values
// prepdata.correct_outputs, // original output values
// max_duration_in_ms, // max duration of simulation
// logfile);
//System.IO.File.AppendAllText(outfile, s_3.FormatResultsAsCSV());
pb.IncrementProgress();
}
public static bool RunSubletyExperiment(Excel.Application app, Excel.Workbook wbh, int nboots, double significance, double threshold, UserSimulation.Classification c, Random r, String outfile, long max_duration_in_ms, String logfile, ProgBar pb, bool ignore_parse_errors)
{
pb.setMax(5);
// record intitial state of spreadsheet
var prepdata = Prep.PrepSimulation(app, wbh, pb, ignore_parse_errors);
// init error generator
var eg = new ErrorGenerator();
// get inputs as an array of addresses to facilitate random selection
// DATA INPUTS ONLY
AST.Address[] inputs = prepdata.dag.terminalInputCells();
for (int i = 0; i < 100; i++)
{
// randomly choose a *numeric* input
// TODO: use Fischer-Yates and take values until
// either we have a satisfactory input value or none
// remain
var rnd_addrs = inputs.Shuffle().ToList();
bool num_found = false;
String input_string;
double input_value;
AST.Address rand_addr;
do
{
// randomly choose an address; if there are none left, fail
if (rnd_addrs.Count == 0) {
return false;
}
rand_addr = rnd_addrs.First();
rnd_addrs = rnd_addrs.Skip(1).ToList();
// get the value
input_string = prepdata.original_inputs[rand_addr];
// try parsing it
if (Double.TryParse(input_string, out input_value))
{
num_found = true;
}
} while (!num_found);
// perturb it
String erroneous_input = eg.GenerateSubtleErrorString(input_value, c);
// create an error dictionary with this one perturbed value
var errors = new CellDict();
errors.Add(rand_addr, erroneous_input);
// run simulations; simulation code does insertion of errors and restore of originals
RunSimulation(app, wbh, nboots, significance, threshold, c, r, outfile, max_duration_in_ms, logfile, pb, prepdata, errors);
}
return true;
}
public static void RunSimulationPaperMain(Excel.Application app, Excel.Workbook wbh, int nboots, double significance, double threshold, UserSimulation.Classification c, Random r, String outfile, long max_duration_in_ms, String logfile, ProgBar pb, bool ignore_parse_errors)
{
pb.setMax(5);
// record intitial state of spreadsheet
var prepdata = Prep.PrepSimulation(app, wbh, pb, ignore_parse_errors);
// generate errors
CellDict errors = UserSimulation.Utility.GenImportantErrors(prepdata.terminal_formula_nodes,
prepdata.original_inputs,
5,
prepdata.correct_outputs,
app,
wbh,
c,
prepdata.dag);
// run paper simulations
RunSimulation(app, wbh, nboots, significance, threshold, c, r, outfile, max_duration_in_ms, logfile, pb, prepdata, errors);
}
// num_bootstraps: the number of bootstrap samples to get
// inputs: a list of inputs; each TreeNode represents an entire input range
// outputs: a list of outputs; each TreeNode represents a function
public static TreeScore DataDebug(int num_bootstraps,
DAG dag,
Excel.Application app,
bool weighted,
bool all_outputs,
long max_duration_in_ms,
Stopwatch sw,
double significance,
ProgBar pb)
{
// this modifies the weights of each node
PropagateWeights(dag);
// filter out non-terminal functions
var output_fns = dag.terminalFormulaNodes(all_outputs);
// filter out non-terminal inputs
var input_rngs = dag.terminalInputVectors();
// first idx: the index of the TreeNode in the "inputs" array
// second idx: the ith bootstrap
var resamples = new InputSample[input_rngs.Length][];
// RNG for sampling
var rng = new Random();
// we save initial inputs and outputs here
var initial_inputs = StoreInputs(input_rngs, dag);
var initial_outputs = StoreOutputs(output_fns, dag);
// Set progress bar max
pb.setMax(input_rngs.Length * 2);
#region RESAMPLE
// populate bootstrap array
// for each input range (a TreeNode)
for (int i = 0; i < input_rngs.Length; i++)
{
// this TreeNode
var t = input_rngs[i];
// resample
resamples[i] = Resample(num_bootstraps, initial_inputs[t], rng);
// update progress bar
pb.IncrementProgress();
}
#endregion RESAMPLE
#region INFERENCE
return Inference(
num_bootstraps,
resamples,
initial_inputs,
initial_outputs,
input_rngs,
output_fns,
dag,
weighted,
significance,
pb);
#endregion INFERENCE
}
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;
}
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);
}
// num_bootstraps: the number of bootstrap samples to get
// inputs: a list of inputs; each TreeNode represents an entire input range
// outputs: a list of outputs; each TreeNode represents a function
public static TreeScore DataDebug(int num_bootstraps,
DAG dag,
Excel.Application app,
bool weighted,
bool all_outputs,
long max_duration_in_ms,
Stopwatch sw,
double significance,
ProgBar pb)
{
// this modifies the weights of each node
PropagateWeights(dag);
// filter out non-terminal functions
var output_fns = dag.terminalFormulaNodes(all_outputs);
// filter out non-terminal inputs
var input_rngs = dag.terminalInputVectors();
// first idx: the index of the TreeNode in the "inputs" array
// second idx: the ith bootstrap
var resamples = new InputSample[input_rngs.Length][];
// RNG for sampling
var rng = new Random();
// we save initial inputs and outputs here
var initial_inputs = StoreInputs(input_rngs, dag);
var initial_outputs = StoreOutputs(output_fns, dag);
// Set progress bar max
pb.setMax(input_rngs.Length * 2);
#region RESAMPLE
// populate bootstrap array
// for each input range (a TreeNode)
for (int i = 0; i < input_rngs.Length; i++)
{
// this TreeNode
var t = input_rngs[i];
// resample
resamples[i] = Resample(num_bootstraps, initial_inputs[t], rng);
// update progress bar
pb.IncrementProgress();
}
#endregion RESAMPLE
#region INFERENCE
return(Inference(
num_bootstraps,
resamples,
initial_inputs,
initial_outputs,
input_rngs,
output_fns,
dag,
weighted,
significance,
pb));
#endregion INFERENCE
}
public void Analyze(long max_duration_in_ms)
{
var sw = new System.Diagnostics.Stopwatch();
sw.Start();
using (var pb = new ProgBar())
{
// Disable screen updating during analysis to speed things up
_app.ScreenUpdating = false;
// Build dependency graph (modifies data)
try
{
_dag = new DAG(_app.ActiveWorkbook, _app, IGNORE_PARSE_ERRORS);
var num_input_cells = _dag.numberOfInputCells();
}
catch (ExcelParserUtility.ParseException e)
{
// cleanup UI and then rethrow
_app.ScreenUpdating = true;
throw e;
}
if (_dag.terminalInputVectors().Length == 0)
{
System.Windows.Forms.MessageBox.Show("This spreadsheet contains no vector-input functions.");
_app.ScreenUpdating = true;
_flaggable = new KeyValuePair<AST.Address, int>[0];
return;
}
// Get bootstraps
var scores = Analysis.DataDebug(NBOOTS,
_dag,
_app,
weighted: USE_WEIGHTS,
all_outputs: CONSIDER_ALL_OUTPUTS,
max_duration_in_ms: max_duration_in_ms,
sw: sw,
significance: _tool_significance,
pb: pb)
.OrderByDescending(pair => pair.Value).ToArray();
if (_debug_mode)
{
var score_str = String.Join("\n", scores.Take(10).Select(score => score.Key.A1FullyQualified() + " -> " + score.Value.ToString()));
System.Windows.Forms.MessageBox.Show(score_str);
System.Windows.Forms.Clipboard.SetText(score_str);
}
List<KeyValuePair<AST.Address, int>> high_scores = new List<KeyValuePair<AST.Address, int>>();
// calculate cutoff idnex
int thresh = scores.Length - Convert.ToInt32(scores.Length * _tool_significance);
// filter out cells that are...
_flaggable = scores.Where(pair => pair.Value >= scores[thresh].Value) // below threshold
.Where(pair => !_known_good.Contains(pair.Key)) // known to be good
.Where(pair => pair.Value != 0).ToArray(); // score == 0
// Enable screen updating when we're done
_app.ScreenUpdating = true;
sw.Stop();
}
}
private static void RunSubletyExperiment(Excel.Application app, Excel.Workbook wb, Random rng, UserSimulation.Classification c, string output_dir, double thresh, ProgBar pb)
{
// number of bootstraps
var NBOOTS = 2700;
// the full path of this workbook
var filename = app.ActiveWorkbook.Name;
// the default output filename
var r = new System.Text.RegularExpressions.Regex(@"(.+)\.xls|xlsx", System.Text.RegularExpressions.RegexOptions.Compiled);
var default_output_file = "simulation_results.csv";
var default_log_file = r.Match(filename).Groups[1].Value + ".iterlog.csv";
// save file location (will append for additional runs)
var savefile = System.IO.Path.Combine(output_dir, default_output_file);
// log file location (new file for each new workbook)
var logfile = System.IO.Path.Combine(output_dir, default_log_file);
// disable screen updating
app.ScreenUpdating = false;
// run simulations
if (!UserSimulation.Config.RunSubletyExperiment(app, wb, NBOOTS, 0.95, thresh, c, rng, savefile, MAX_DURATION_IN_MS, logfile, pb, IGNORE_PARSE_ERRORS))
{
System.Windows.Forms.MessageBox.Show("This spreadsheet contains no numeric inputs.");
}
// enable screen updating
app.ScreenUpdating = true;
}
// remove errors until none remain
private UserResults SimulateUser(int nboots,
double significance,
CutoffKind ck,
DAG dag,
CellDict original_inputs,
CellDict errord,
CellDict correct_outputs,
Excel.Workbook wb,
Excel.Application app,
AnalysisType analysis_type,
bool weighted,
bool all_outputs,
long max_duration_in_ms,
Stopwatch sw,
String logfile,
ProgBar pb
)
{
// init user results data structure
var o = new UserResults();
HashSet<AST.Address> known_good = new HashSet<AST.Address>();
// initialize procedure
var errors_remain = true;
var max_errors = new ErrorDict();
var incorrect_outputs = Utility.SaveOutputs(dag.terminalFormulaNodes(all_outputs), dag);
var errors_found = 0;
var number_of_true_errors = errord.Count;
Utility.UpdatePerFunctionMaxError(correct_outputs, incorrect_outputs, max_errors);
// the corrected state of the spreadsheet
CellDict partially_corrected_outputs = correct_outputs.ToDictionary(p => p.Key, p => p.Value);
// remove errors loop
var cells_inspected = 0;
List<KeyValuePair<AST.Address, int>> filtered_high_scores = null;
bool correction_made = true;
while (errors_remain)
{
Console.Write(".");
AST.Address flagged_cell = null;
// choose the appropriate test
if (analysis_type == AnalysisType.CheckCell5 ||
analysis_type == AnalysisType.CheckCell10
)
{
flagged_cell = SimulationStep.CheckCell_Step(o,
significance,
ck,
nboots,
dag,
app,
weighted,
all_outputs,
correction_made,
known_good,
ref filtered_high_scores,
max_duration_in_ms,
sw,
pb);
} else if (analysis_type == AnalysisType.NormalPerRange)
{
flagged_cell = SimulationStep.NormalPerRange_Step(dag, wb, known_good, max_duration_in_ms, sw);
}
else if (analysis_type == AnalysisType.NormalAllInputs)
{
flagged_cell = SimulationStep.NormalAllOutputs_Step(dag, app, wb, known_good, max_duration_in_ms, sw);
}
// stop if the test no longer returns anything or if
// the test is simply done inspecting based on a fixed threshold
if (flagged_cell == null || (ck.isCountBased && ck.Threshold == cells_inspected))
{
errors_remain = false;
}
else // a cell was flagged
{
//cells_inspected should only be incremented when a cell is actually flagged. If nothing is flagged,
//then nothing is inspected, so cells_inspected doesn't increase.
cells_inspected += 1;
// check to see if the flagged value is actually an error
if (errord.ContainsKey(flagged_cell))
{
correction_made = true;
errors_found += 1;
// P(k) * rel(k)
o.PrecRel_at_k.Add(errors_found / (double)cells_inspected);
o.true_positives.Add(flagged_cell);
// correct flagged cell
flagged_cell.GetCOMObject(app).Value2 = original_inputs[flagged_cell];
Utility.UpdatePerFunctionMaxError(correct_outputs, partially_corrected_outputs, max_errors);
// compute total error after applying this correction
var current_total_error = Utility.CalculateTotalError(correct_outputs, partially_corrected_outputs);
o.current_total_error.Add(current_total_error);
// save outputs
partially_corrected_outputs = Utility.SaveOutputs(dag.terminalFormulaNodes(all_outputs), dag);
}
else
{
correction_made = false;
// numerator is 0 here because rel(k) = 0 when no error was found
o.PrecRel_at_k.Add(0.0);
o.false_positives.Add(flagged_cell);
}
// mark it as known good -- at this point the cell has been
// 'inspected' regardless of whether it was an error
// It was either corrected or marked as OK
known_good.Add(flagged_cell);
// compute output error magnitudes
var output_error_magnitude = Utility.MeanErrorMagnitude(partially_corrected_outputs, correct_outputs);
// compute input error magnitude
double num_input_error_magnitude;
double str_input_error_magnitude;
if (errord.ContainsKey(flagged_cell))
{
if (Utility.BothNumbers(errord[flagged_cell], original_inputs[flagged_cell]))
{
num_input_error_magnitude = Utility.NumericalMagnitudeChange(Double.Parse(errord[flagged_cell]), Double.Parse(original_inputs[flagged_cell]));
str_input_error_magnitude = 0;
}
else
{
num_input_error_magnitude = 0;
str_input_error_magnitude = Utility.StringMagnitudeChange(errord[flagged_cell], original_inputs[flagged_cell]);
}
}
else
{
num_input_error_magnitude = 0;
str_input_error_magnitude = 0;
}
// write error log
var logentry = new LogEntry(analysis_type,
wb.Name,
flagged_cell,
original_inputs[flagged_cell],
errord.ContainsKey(flagged_cell) ? errord[flagged_cell] : original_inputs[flagged_cell],
output_error_magnitude,
num_input_error_magnitude,
str_input_error_magnitude,
true,
correction_made,
significance,
ck.Threshold);
logentry.WriteLog(logfile);
_error_log.Add(logentry);
}
}
// find all of the false negatives
o.false_negatives = Utility.GetFalseNegatives(o.true_positives, o.false_positives, errord);
o.max_errors = max_errors;
var last_out_err_mag = Utility.MeanErrorMagnitude(partially_corrected_outputs, correct_outputs);
// write out all false negative information
foreach (AST.Address fn in o.false_negatives)
{
double num_input_error_magnitude;
double str_input_error_magnitude;
if (Utility.BothNumbers(errord[fn], original_inputs[fn]))
{
num_input_error_magnitude = Utility.NumericalMagnitudeChange(Double.Parse(errord[fn]), Double.Parse(original_inputs[fn]));
str_input_error_magnitude = 0;
}
else
{
num_input_error_magnitude = 0;
str_input_error_magnitude = Utility.StringMagnitudeChange(errord[fn], original_inputs[fn]);
}
// write error log
_error_log.Add(new LogEntry(analysis_type,
wb.Name,
fn,
original_inputs[fn],
errord[fn],
last_out_err_mag,
num_input_error_magnitude,
str_input_error_magnitude,
false,
true,
significance,
ck.Threshold));
}
return o;
}
// returns the number of cells inspected
public int Run(int nboots, // number of bootstraps
string xlfile, // name of the workbook
double significance, // significance threshold for test
CutoffKind ck, // kind of threshold function to use
Excel.Application app, // reference to Excel app
Classification c, // data from which to generate errors
Random r, // a random number generator
AnalysisType analysisType, // the type of analysis to run
bool weighted, // should we weigh things?
bool all_outputs, // if !all_outputs, we only consider terminal outputs
DAG dag,
Excel.Workbook wb,
AST.Address[] terminal_formula_cells,
AST.Range[] terminal_input_vectors,
CellDict original_inputs,
CellDict correct_outputs,
long max_duration_in_ms,
String logfile, //filename for the output log
ProgBar pb
)
{
//set wbname and path
_wb_name = xlfile;
_wb_path = wb.Path;
_analysis_type = analysisType;
_significance = significance;
_all_outputs = all_outputs;
_weighted = weighted;
//Now we want to inject the errors from _errors
Utility.InjectValues(app, wb, _errors);
// save function outputs
CellDict incorrect_outputs = Utility.SaveOutputs(terminal_formula_cells, dag);
//Time the removal of errors
Stopwatch sw = new Stopwatch();
sw.Start();
// remove errors until none remain; MODIFIES WORKBOOK
_user = SimulateUser(nboots, significance, ck, dag, original_inputs, _errors, correct_outputs, wb, app, analysisType, weighted, all_outputs, max_duration_in_ms, sw, logfile, pb);
sw.Stop();
TimeSpan elapsed = sw.Elapsed;
_analysis_time = elapsed.TotalSeconds;
// save partially-corrected outputs
var partially_corrected_outputs = Utility.SaveOutputs(terminal_formula_cells, dag);
// compute total relative error
_error = Utility.CalculateNormalizedError(correct_outputs, partially_corrected_outputs, _user.max_errors);
_total_relative_error = Utility.TotalRelativeError(_error);
// compute starting total relative error (normalized by max_errors)
ErrorDict starting_error = Utility.CalculateNormalizedError(correct_outputs, incorrect_outputs, _user.max_errors);
_initial_total_relative_error = Utility.TotalRelativeError(starting_error);
// effort
_max_effort = dag.allCells().Length;
_effort = (_user.true_positives.Count + _user.false_positives.Count);
_expended_effort = (double)_effort / (double)_max_effort;
// compute average precision
// AveP = (\sum_{k=1}^n (P(k) * rel(k))) / |total positives|
// where P(k) is the precision at threshold k,
// rel(k) = \{ 1 if item at k is a true positive, 0 otherwise
_average_precision = _user.PrecRel_at_k.Sum() / (double)_errors.Count;
// restore original values
Utility.InjectValues(app, wb, original_inputs);
_tree_construct_time = dag.AnalysisMilliseconds / 1000.0;
// flag that we're done; safe to print output results
_simulation_run = true;
// return the number of cells inspected
return _effort;
}