public void Extract_Tests()
{
var featureExtraction = new FeatureExtraction();
var validValues = "Button Fly,Button-End,D Ring,Double D Ring,Drawstring,Elastic,Flat Solid Buckle,Hook & Eye,J-Clip,Pull On,Round Classic Ring,Self Tie,Snap On,Snaps,Square Classic Ring,Velcro,Zipper";
//Scenario 1: exact match
var source1 = "Double D Ring";
var scenario1 = featureExtraction.Extract(source1, validValues);
Assert.AreEqual("Double D Ring", scenario1);
//Scenario 2: source contains list of valid values
// 2.1 source contains single matching
// 2.2 source contains multiple matching
// 2.2 source contains multiple matching but get only one
// 2.3 source contains multiple matching but get only one
// with assistance of keyword
var source21 = "Versace Collection D Ring woven leather tote bag with golden hardware. Flat top handles with hanging logo medallion, 5.5\" drop. Removable chain shoulder strap, 18.5\" drop. Flap top with logo plaque; snap closure. Interior, cotton lining; one zip and two slip pockets. Metal feet protect bottom of bag. 10.5\"H x 14\"W x 4.5\"D; weighs 13 oz. Imported.";
var scenario21 = featureExtraction.Extract(source21, validValues);
Assert.AreEqual("D Ring", scenario21);
var source221 =
"Versace Collection D Ring woven leather tote bag with golden hardware. Flat top handles with hanging logo medallion, 5.5\" drop. Removable chain shoulder strap, 18.5\" drop. Flap top with logo plaque; snap closure. Interior, cotton lining; one zip and two slip pockets. Metal feet protect bottom of bag. 10.5\"H x 14\"W x 4.5\"D; weighs 13 oz. Imported. Elastic";
var scenario221 = featureExtraction.Extract(source221, validValues,
3, 0);
Assert.AreEqual("D Ring/Elastic", scenario221);
var scenario222 = featureExtraction.Extract(source221, validValues,
3, 1);
Assert.AreEqual("D Ring", scenario222);
//var scenario223 = featureExtraction.Extract(source221, validValues,
// 3, 1, "/", "closure");
//Assert.AreEqual("Snaps", scenario223);
//Scenario 3:
//Scenario 4: similarity checks
// return multiple result
var source4 = "Versace Collection woven D Rin leather tote bag with golden hardware. Flat top handles with hanging logo medallion, 5.5\" drop. Removable chain shoulder strap, 18.5\" drop. Flap top with logo plaque; snap closure. Interior, cotton lining; one zip and two slip pockets. Metal feet protect bottom of bag. 10.5\"H x 14\"W x 4.5\"D; weighs 13 oz. Imported.";
var scenario4 = featureExtraction.Extract(source4, validValues, 1, 0);
var source41 = "Versace Collection woven Elasti leather tote bag with golden hardware. Flat top handles with hanging logo medallion, 5.5\" drop. Removable chain shoulder strap, 18.5\" drop. Flap top with logo plaque; snap closure. Interior, cotton lining; one zip and two slip pockets. Metal feet protect bottom of bag. 10.5\"H x 14\"W x 4.5\"D; weighs 13 oz. Imported.";
var scenario41 = featureExtraction.Extract(source41, validValues, 1, 1);
//var source = "Versace Collection D Ring woven leather tote bag with golden hardware. Flat top handles with hanging logo medallion, 5.5\" drop. Removable chain shoulder strap, 18.5\" drop. Flap top with logo plaque; snap closure. Interior, cotton lining; one zip and two slip pockets. Metal feet protect bottom of bag. 10.5\"H x 14\"W x 4.5\"D; weighs 13 oz. Imported.";
//var s = new FeatureExtraction().Extract(source,validValues,3);
}
public static void ExtractAllFeatures()
{
OrderFeatures = new HashSet <Feature>();
SelectorFeatures = new HashSet <Feature>();
FeatureFrequencey = new Dictionary <string, int>();
foreach (String docName in trainingDocsNames)
{
HtmlNode src = null;
docsAndNames.TryGetValue(docName, out src);
HashSet <HtmlNode> docNodes = new HashSet <HtmlNode>(src.SelectNodes("//*"));
//extractFeatures
HashSet <Feature> extacted = new HashSet <Feature>();
if (docsFeatures.ContainsKey(docName))
{
docsFeatures.TryGetValue(docName, out extacted);
}
else
{
HashSet <HtmlNode> currNodes = new HashSet <HtmlNode>(docNodes.Intersect(TargetNodes));
foreach (HtmlNode curr in currNodes)
{
extacted.UnionWith(FeatureExtraction.Extract(curr));
}
docsFeatures.Add(docName, extacted);
}
foreach (Feature currFeature in extacted)
{
if (currFeature.type.Equals(Feature.FeatureType.Order))
{
OrderFeatures.Add(currFeature);
}
else
{
SelectorFeatures.Add(currFeature);
int val = 0;
if (FeatureFrequencey.TryGetValue(currFeature.feature.First().ToLower(), out val))
{
FeatureFrequencey.Remove(currFeature.feature.First().ToLower());
FeatureFrequencey.Add(currFeature.feature.First().ToLower(), val + 1);
}
else
{
FeatureFrequencey.Add(currFeature.feature.First().ToLower(), 1);
}
}
}
}
int thresholdNumber = 2;
if (trainingDocsNames.Count() < 2)
{
thresholdNumber = 1;
}
var sharedFeatures = new HashSet <String>(FeatureFrequencey.Where(x => (x.Value >= thresholdNumber /*(docs.Count())*/)).Select(x => x.Key));
HashSet <Feature> selectorFeaturesFiltered = new HashSet <Feature>();
foreach (Feature f in SelectorFeatures)
{
if (sharedFeatures.Contains(f.feature.First().ToLower()))
{
selectorFeaturesFiltered.Add(f);
}
}
SelectorFeatures = selectorFeaturesFiltered;
}
