def analyze(self, unknownDocument):
# unknownDocument is a single doc, type Document.
results = dict()
unknownDocHistogram: dict = histograms.normalizeHistogram(
histograms.generateAbsoluteHistogram(unknownDocument))
#unknownDocHistogramNp = np.asarray(list(unknownDocHistogram.items()))
results = dict()
if self.mode == "author":
for author in self._histograms:
authorResult = 0 # numerial result for an author (mean histogram)
for item in self._histograms[author]:
if unknownDocHistogram.get(item) != None:
authorResult -= self._histograms[author][
item] * math.log(unknownDocHistogram[item])
results[author] = authorResult
elif self.mode == "document":
for author in self._histograms:
for doc in self._histograms[author]:
docResult = 0 # numerical result for a single document
for item in doc:
if unknownDocHistogram.get(item) != None:
docResult -= self._histograms[author][doc][
item] * math.log(unknownDocHistogram[item])
results[doc] = docResult
return results
def analyze(self, unknownDocument):
'''Compare a normalized histogram of unknownDocument against the normalized known document histograms and return a dictionary of distances.'''
results = dict()
for author, knownHist in self._authorHistograms.items():
results[author] = self.distance.distance(
histograms.normalizeHistogram(
histograms.generateAbsoluteHistogram(unknownDocument)),
knownHist)
return results
def train(self, knownDocuments):
if self.mode == "author":
# authors -> mean histograms
self._histograms = histograms.generateKnownDocsMeanHistograms(
histograms.generateKnownDocsNormalizedHistogramSet(
knownDocuments))
#self._histogramsNp = {author:np.asarray(list(docHistogram.items())) for (author,docHistogram) in self._histograms.items()}
# ^^ goes into the histogram list and change mean histograms into numpy arrays
elif self.mode == 'document':
# authors -> list of histograms
self._histograms = histograms.generateKnownDocsNormalizedHistogramSet(
knownDocuments)
def train(self, knownDocuments): '''Get a mean normalized histogram for each known author.''' self.authorHistograms = histograms.generateKnownDocsMeanHistograms(histograms.generateKnownDocsNormalizedHistogramSet(knownDocuments))