def it_should_search_test(self):
vectorSpace = VectorSpace(self.documents)
eq_(vectorSpace.search(["cat"]), [
0.14487566959813258, 0.1223402602604157, 0.07795622058966725,
0.05586504042763477
])
def run(data, queries, max_response=10):
all_documents = []
for entry in data:
all_documents.append(entry["raw_data"])
vector_space = VectorSpace(all_documents)
#Search for cat
indexed_result = {}
result = vector_space.search([queries])
index = 0
for entry in result:
indexed_result[index] = entry
index += 1
sorted_resp = sorted(indexed_result.items(),
key=operator.itemgetter(1),
reverse=True)
sorted_resp = sorted_resp[:int(max_response) + 1]
response = {}
rank = 1
for entry in sorted_resp:
data_index = entry[0]
response[rank] = data[data_index]
rank += 1
return response
def it_should_find_related_test(self):
vector_space = VectorSpace(self.documents)
eq_(vector_space.related(0), [1.0000000000000002, 0.9999999999999998, 0.0])
def it_should_search_test(self):
vector_space = VectorSpace(self.documents, transforms = [])
eq_(vector_space.search(["cat"]), [1.0, 0.7071067811865475, 0.0])
def it_should_find_return_similarity_rating_test(self):
vectorSpace = VectorSpace(self.documents)
eq_(vectorSpace.related(0), [1.0, 0.9922455760198575, 0.08122814162371816, 0.0762173599906487])
queries = pickle.load(open("QueryStrings.p", "rb"))
print "total queries"
print len(queries)
print "loaded queries"
documents = pickle.load(open("documentContentList2.p", "rb"))
print "loaded documents"
docIds = pickle.load(open("docIdList.p", "rb"))
print len(docIds)
print "loaded doc ids"
documents = list(documents)
print "loaded documents"
print len(documents)
#documents = documents[:100]
#docIds = docIds[:100]
#queries = queries[:10]
vector_space = VectorSpace(documents)
print "finished conversion"
print "load user click file"
userQueriesAndClicks = pickle.load(
open("user_specific_positive_negative_examples_dic_test", "rb"))
print "finished loading user click file"
#queryResults = dict( [ (x[0], (x[1], x[2])) for x in userQueriesAndClicks_strict[userID] ])
# given a query and a ranking, this function provides the relevanceJudgements list as
# required by averagePrecision
def turnIntoBinaryRelevanceThing(query, ranking, relevantDocuments):
#rel = self.relevantDocuments[query]
binarized = []
# Create the corpus
file_content_all=[]
corpus='AspectJ'
creator = sourceCorpusCreator()
sourcepath = "E:\PhD\LSI\Repo\\"+corpus+"\SourceAndBugData244\\"
keywordsfilepath='E:\PhD\LSI\Repo\\'+corpus+'\data\keyword-documents.txt'
#querypath="E:\PhD\LSI\Repo\\"+corpus+"\BugData\\"
source_content_all={}
source_content_all=creator.CorpusCreatorDict(sourcepath, '.java')
print ('Total files in corpus ')
print(len(source_content_all))
print (source_content_all)
vector_space = VectorSpace(source_content_all)
file_path_all=vector_space.get_file_path_all()
print (file_path_all)
document_ID_file_info_mapping=vector_space.get_document_ID_file_info_mapping()
print (document_ID_file_info_mapping)
keywords_docs_string=str(vector_space.vector_index_to_keyword_mapping)
file_read_write=FileReadWrite(sourcepath)
file_read_write.writeFiles(keywordsfilepath, keywords_docs_string)
print (len(vector_space.vector_index_to_keyword_mapping))
#import pdb
#pdb.set_trace()
print ("Keyords-document vector/matrix")
print ('length of vector_space.collection_of_document_term_vectors')
print len(vector_space.collection_of_document_term_vectors)
document_term_matrix=vector_space.collection_of_document_term_vectors
import pandas as pd
import pickle
from semanticpy.vector_space import VectorSpace
data = pd.read_json('cc_jokes_valid.json')
df = pd.DataFrame(data)
dfList = df['content'].tolist()
#builds vector space model and saves to picle (takes long time)
vector_space = VectorSpace(dfList)
filehandler = open('vsm.obj', 'w')
pickle.dump(vector_space, filehandler)
def vector_space_mapping(self):
v = VectorSpace(self.documents)
matrix = v.collection_of_document_term_vectors
return matrix
