def similarity():
"""
Handles the similarity query page functionality. Returns ranked list of files and their cosine similarities to a comparison document.
"""
fileManager = session_functions.loadFileManager()
labels = fileManager.getActiveLabels()
if 'analyoption' not in session:
session['analyoption'] = constants.DEFAULT_ANALIZE_OPTIONS
if 'uploadname' not in session:
session['similarities'] = constants.DEFAULT_SIM_OPTIONS
if request.method == 'GET':
# 'GET' request occurs when the page is first loaded
similaritiesgenerated = False
return render_template('similarity.html', labels=labels, docsListScore="", docsListName="",
similaritiesgenerated=similaritiesgenerated)
if request.method == "POST":
# 'POST' request occur when html form is submitted (i.e. 'Get Graphs', 'Download...')
docsListScore, docsListName = utility.generateSimilarities(fileManager)
similaritiesgenerated = True
session_functions.cacheAnalysisOption()
session_functions.cacheSimOptions()
return render_template('similarity.html', labels=labels, docsListScore=docsListScore, docsListName=docsListName,
similaritiesgenerated=similaritiesgenerated)
def statistics():
"""
Handles the functionality on the Statistics page ...
Note: Returns a response object (often a render_template call) to flask and eventually
to the browser.
"""
fileManager = session_functions.loadFileManager()
if request.method == "GET":
# "GET" request occurs when the page is first loaded.
labels = fileManager.getActiveLabels()
# if len(labels) >= 1:
#FileInfoDict, corpusInfoDict = fileManager.generateStatistics()
# return render_template('statistics.html', labels=labels, FileInfoDict=FileInfoDict,
# corpusInfoDict=corpusInfoDict)
if 'analyoption' not in session:
session['analyoption'] = constants.DEFAULT_ANALIZE_OPTIONS
return render_template('statistics.html', labels=labels)
if request.method == "POST":
normalize = request.form['normalizeType']
labels = fileManager.getActiveLabels()
if len(labels) >= 1:
FileInfoDict, corpusInfoDict= utility.generateStatistics(fileManager)
session_functions.cacheAnalysisOption()
return render_template('statistics.html', labels=labels, FileInfoDict=FileInfoDict,
corpusInfoDict=corpusInfoDict, normalize=normalize)
def kmeans():
"""
Handles the functionality on the kmeans page. It analyzes the various texts and
displays the class label of the files.
Note: Returns a response object (often a render_template call) to flask and eventually
to the browser.
"""
fileManager = session_functions.loadFileManager()
labels = fileManager.getActiveLabels()
defaultK = int(len(labels) / 2)
if 'analyoption' not in session:
session['analyoption'] = constants.DEFAULT_ANALIZE_OPTIONS
if 'kmeanoption' not in session:
session['kmeanoption'] = constants.DEFAULT_KMEAN_OPTIONS
if request.method == 'GET':
# 'GET' request occurs when the page is first loaded
return render_template('kmeans.html', labels=labels, silhouettescore='', kmeansIndex=[], fileNameStr='',
fileNumber=len(labels), KValue=0, defaultK=defaultK,
colorChartStr='', kmeansdatagenerated=False)
if request.method == "POST":
# 'POST' request occur when html form is submitted (i.e. 'Get Graphs', 'Download...')
if request.form['viz'] == 'PCA':
kmeansIndex, silhouetteScore, fileNameStr, KValue, colorChartStr = utility.generateKMeansPCA(fileManager)
session_functions.cacheAnalysisOption()
session_functions.cacheKmeanOption()
session_functions.saveFileManager(fileManager)
return render_template('kmeans.html', labels=labels, silhouettescore=silhouetteScore,
kmeansIndex=kmeansIndex,
fileNameStr=fileNameStr, fileNumber=len(labels), KValue=KValue, defaultK=defaultK,
colorChartStr=colorChartStr, kmeansdatagenerated=True)
elif request.form['viz'] == 'Voronoi':
kmeansIndex, silhouetteScore, fileNameStr, KValue, colorChartStr, finalPointsList, finalCentroidsList, textData, maxVal = utility.generateKMeansVoronoi(fileManager)
session_functions.cacheAnalysisOption()
session_functions.cacheKmeanOption()
session_functions.saveFileManager(fileManager)
return render_template('kmeans.html', labels=labels, silhouettescore=silhouetteScore,
kmeansIndex=kmeansIndex, fileNameStr=fileNameStr, fileNumber=len(labels),
KValue=KValue, defaultK=defaultK, colorChartStr=colorChartStr,
finalPointsList=finalPointsList, finalCentroidsList=finalCentroidsList,
textData=textData, maxVal=maxVal, kmeansdatagenerated=True)
def tokenizer():
"""
Handles the functionality on the tokenizer page. It analyzes the texts to produce
and send various frequency matrices.
Note: Returns a response object (often a render_template call) to flask and eventually
to the browser.
"""
fileManager = session_functions.loadFileManager()
if 'analyoption' not in session:
session['analyoption'] = constants.DEFAULT_ANALIZE_OPTIONS
if 'csvoptions' not in session:
session['csvoptions'] = constants.DEFAULT_CSV_OPTIONS
if request.method == "GET":
# "GET" request occurs when the page is first loaded.
labels = fileManager.getActiveLabels()
return render_template('tokenizer.html', labels=labels, matrixExist=False)
if 'gen-csv' in request.form:
# The 'Generate and Visualize Matrix' button is clicked on tokenizer.html.
session_functions.cacheAnalysisOption()
session_functions.cacheCSVOptions()
labels = fileManager.getActiveLabels()
matrixTitle, tableStr = utility.generateTokenizeResults(fileManager)
session_functions.saveFileManager(fileManager)
return render_template('tokenizer.html', labels=labels, matrixTitle=matrixTitle,
tableStr=tableStr, matrixExist=True)
if 'get-csv' in request.form:
# The 'Download Matrix' button is clicked on tokenizer.html.
session_functions.cacheAnalysisOption()
session_functions.cacheCSVOptions()
savePath, fileExtension = utility.generateCSV(fileManager)
session_functions.saveFileManager(fileManager)
return send_file(savePath, attachment_filename="frequency_matrix" + fileExtension, as_attachment=True)
def topword():
"""
Handles the topword page functionality.
"""
fileManager = session_functions.loadFileManager()
labels = fileManager.getActiveLabels()
if 'topwordoption' not in session:
session['topwordoption'] = constants.DEFAULT_TOPWORD_OPTIONS
if 'analyoption' not in session:
session['analyoption'] = constants.DEFAULT_ANALIZE_OPTIONS
if request.method == 'GET':
# 'GET' request occurs when the page is first loaded
# get the class label and eliminate the id (this is not the unique id in filemanager)
ClassdivisionMap = fileManager.getClassDivisionMap()[1:]
# if there is no file active (ClassdivisionMap == []) just jump to the page
# notice python eval from right to left
# if there is only one chunk then make the default test prop-z for all
if ClassdivisionMap != [] and len(ClassdivisionMap[0]) == 1:
session['topwordoption']['testMethodType'] = 'pz'
session['topwordoption']['testInput'] = 'useAll'
return render_template('topword2.html', labels=labels, classmap=ClassdivisionMap, topwordsgenerated='class_div')
if request.method == "POST":
# 'POST' request occur when html form is submitted (i.e. 'Get Graphs', 'Download...')
if request.form['testMethodType'] == 'pz':
if request.form['testInput'] == 'useclass':
result = utility.GenerateZTestTopWord(fileManager)
# only give the user a preview of the topWord
for key in result.keys():
if len(result[key]) > 20:
result.update({key: result[key][:20]})
session_functions.cacheAnalysisOption()
session_functions.cacheTopwordOptions()
return render_template('topword2.html', result=result, labels=labels, topwordsgenerated='pz_class')
else:
result = utility.GenerateZTestTopWord(fileManager)
print result[0]
# only give the user a preview of the topWord
for i in range(len(result)):
if len(result[i][1]) > 20:
result[i][1] = result[i][1][:20]
print result[1]
session_functions.cacheAnalysisOption()
session_functions.cacheTopwordOptions()
return render_template('topword2.html', result=result, labels=labels, topwordsgenerated='pz_all')
else:
result = utility.generateKWTopwords(fileManager)
print result
# only give the user a preview of the topWord
if len(result) > 50:
result = result[:50]
session_functions.cacheAnalysisOption()
session_functions.cacheTopwordOptions()
return render_template('topword2.html', result=result, labels=labels, topwordsgenerated='KW')
def hierarchy():
"""
Handles the functionality on the hierarchy page. It analyzes the various texts and
displays a dendrogram.
Note: Returns a response object (often a render_template call) to flask and eventually
to the browser.
"""
fileManager = session_functions.loadFileManager()
leq = '≤'.decode('utf-8')
if 'analyoption' not in session:
session['analyoption'] = constants.DEFAULT_ANALIZE_OPTIONS
if 'hierarchyoption' not in session:
session['hierarchyoption'] = constants.DEFAULT_HIERARCHICAL_OPTIONS
if request.method == "GET":
# "GET" request occurs when the page is first loaded.
labels = fileManager.getActiveLabels()
thresholdOps = {}
return render_template('hierarchy.html', labels=labels, thresholdOps=thresholdOps)
if 'dendro_download' in request.form:
# The 'Download Dendrogram' button is clicked on hierarchy.html.
# sends pdf file to downloads folder.
utility.generateDendrogram(fileManager)
attachmentname = "den_" + request.form['title'] + ".pdf" if request.form['title'] != '' else 'dendrogram.pdf'
session_functions.cacheAnalysisOption()
session_functions.cacheHierarchyOption()
return send_file(pathjoin(session_functions.session_folder(), constants.RESULTS_FOLDER + "dendrogram.pdf"),
attachment_filename=attachmentname, as_attachment=True)
if 'dendroSVG_download' in request.form:
utility.generateDendrogram(fileManager)
attachmentname = "den_" + request.form['title'] + ".svg" if request.form['title'] != '' else 'dendrogram.svg'
session_functions.cacheAnalysisOption()
session_functions.cacheHierarchyOption()
return send_file(pathjoin(session_functions.session_folder(), constants.RESULTS_FOLDER + "dendrogram.svg"),
attachment_filename=attachmentname, as_attachment=True)
if 'getdendro' in request.form:
# The 'Get Dendrogram' button is clicked on hierarchy.html.
pdfPageNumber, score, inconsistentMax, maxclustMax, distanceMax, distanceMin, monocritMax, monocritMin, threshold = utility.generateDendrogram(fileManager)
session['dengenerated'] = True
labels = fileManager.getActiveLabels()
inconsistentOp = "0 " + leq + " t " + leq + " " + str(inconsistentMax)
maxclustOp = "2 " + leq + " t " + leq + " " + str(maxclustMax)
distanceOp = str(distanceMin) + " " + leq + " t " + leq + " " + str(distanceMax)
monocritOp = str(monocritMin) + " " + leq + " t " + leq + " " + str(monocritMax)
thresholdOps = {"inconsistent": inconsistentOp, "maxclust": maxclustOp, "distance": distanceOp,
"monocrit": monocritOp}
session_functions.saveFileManager(fileManager)
session_functions.cacheAnalysisOption()
session_functions.cacheHierarchyOption()
return render_template('hierarchy.html', labels=labels, pdfPageNumber=pdfPageNumber, score=score,
inconsistentMax=inconsistentMax, maxclustMax=maxclustMax, distanceMax=distanceMax,
distanceMin=distanceMin, monocritMax=monocritMax, monocritMin=monocritMin,
threshold=threshold, thresholdOps=thresholdOps)
