Пример #1
0
def test():
    patns = mockdb.get_mock().patns
    def test_func(doc):
        return {'$set': {'dummy_field': 'yay!'}}
    n_right = parallelMap(test_func, patns, 'pno')
    print "num for which alleged successfull: {}".format(n_right)
    print "num with dummy_field: {}".format(patns.find({'dummy_field': 'yay!'}).count())
    return patns
Пример #2
0
def test():
    db = get_mock()
    in_deg_counts, out_deg_counts = in_and_out_counts(db, 100)
    f, (ax1, ax2) = plt.subplots(1, 2, sharey=True)
    f.set_size_inches(18.5, 10.5)
    ax1.hist(in_deg_counts.keys(), weights=in_deg_counts.values(), bins=20)
    ax1.set_xlabel('In-Degree')
    ax1.set_ylabel('Count')
    ax2.hist(out_deg_counts.keys(), weights=out_deg_counts.values(), bins=20)
    ax2.set_xlabel('Out-Degree')
    ax2.set_ylabel('Count')
    plt.suptitle('Degree Distributions')
    plt.savefig('degree_distributions_test.png')
Пример #3
0
def test():
    db = get_mock()
    in_deg_counts, out_deg_counts = in_and_out_counts(db, 100)
    f,(ax1, ax2) = plt.subplots(1,2,sharey=True)
    f.set_size_inches(18.5, 10.5)
    ax1.hist(in_deg_counts.keys(), weights=in_deg_counts.values(), bins=20)
    ax1.set_xlabel('In-Degree')
    ax1.set_ylabel('Count')
    ax2.hist(out_deg_counts.keys(), weights=out_deg_counts.values(), bins=20)
    ax2.set_xlabel('Out-Degree')
    ax2.set_ylabel('Count')
    plt.suptitle('Degree Distributions')
    plt.savefig('degree_distributions_test.png')
Пример #4
0
def test():
    db = get_mock()
    inctr, outctr = cites_over_time(db,limit=100)
    in_dates, in_cites = zip(*inctr.items())
    out_dates, out_cites = zip(*outctr.items())
    f,(ax1, ax2) = plt.subplots(1,2,sharey=True)
    f.set_size_inches(18.5, 10.5)
    ax1.scatter(in_dates, in_cites)
    ax1.set_xlabel('Date')
    ax1.set_title('In-degrees over Time.')
    ax2.scatter(out_dates, out_cites)
    ax2.set_xlabel('Date')
    ax2.set_title('Out-Degrees over Time.')
    plt.show()
    return inctr, outctr
Пример #5
0
 def setUp(self):
     self.db = get_mock()
     self.pat_coll = self.db.pat_text
     self.K = 5
     self.n_docs = 50
     data_dir = '/'.join([_this_dir, 'data'])
     self.out_dir = '/'.join([_this_dir, 'test_output'])
     self.stored_vocab_fn = '/'.join([data_dir, 'test_vocab.dict'])
     self.stored_corpus_fn = '/'.join([data_dir, 'test_corpus.svmlight'])
     self.visualize_fn = '/'.join([data_dir, 'test_vis.png'])
     fields = ['_id', 'patText']
     nulls = [None, '']
     self.pnos, self.texts = get_fields_unordered(self.pat_coll, fields,
                                                  nulls, self.n_docs)
     self.model = lda.MyLda(self.K, 'tester')
Пример #6
0
def test():
    enforce_func = lambda x: len(x.get('citedby', [])) > 100
    realdb = MongoClient().patents
    patent_family = [4061724, 4064521, 4340563, 4405829, 4655771, 4683202, 4723129, 5103459, 5143854, 5572643]
    print "Getting docs from real db..."
    patent_family_docs = [realdb.traits.find_one({'_id': pno}) for pno in patent_family]
#    pprint(patent_family_docs)
    mockdb = get_mock()
    print "computing 2 generation tf-idf reach..."
    compute_reach(trait='tf-idf', n_gens=2, db=mockdb, family=patent_family_docs, enforce_func = enforce_func)
    print "computing 2 generation w2v reach..."
    compute_reach(trait='w2v', n_gens=2, db=mockdb, family=patent_family_docs, enforce_func = enforce_func)
    print "computing 2 generation w2v trait variance..."
    compute_trait_variance(trait='w2v', n_gens=2, db=mockdb, family=patent_family_docs, enforce_func = enforce_func)
    print "computing 5 generation tf-idf reach..."
    compute_reach(trait='tf-idf', n_gens=5, db=mockdb, family=patent_family_docs, enforce_func = enforce_func)
    print "computing 5 generation w2v reach..."
    compute_reach(trait='w2v', n_gens=5, db=mockdb, family=patent_family_docs, enforce_func = enforce_func)
    print "computing 5 generation w2v trait variance..."
    compute_trait_variance(trait='w2v', n_gens=5, db=mockdb, family=patent_family_docs, enforce_func = enforce_func)
    print "Done testing."
Пример #7
0
 def setUp(self):
     self.db = get_mock()
     self.n_test = 10
Пример #8
0
 def setUp(self):
     self.db = get_mock()
     self.n_test = 2
     self.pnos_test = dbutil.get_fields_unordered(self.db.patns, ['pno'], [0], 
                                                  self.n_test)[0]
Пример #9
0
 def setUp(self):
     self.db = get_mock()
Пример #10
0
 def setUp(self):
     self.db = get_mock()
     self.n_test = 10
     self.family_docs = self.db.traits.find().limit(self.n_test)
Пример #11
0
 def setUp(self):
     self.enforcer = lambda x: len(x.get('citedby', [])) > 100
     self.db = get_mock()
     self.n_test = 5
     self.family_docs = self.db.traits.find().limit(self.n_test)
Пример #12
0
 def setUp(self):
     self.db = get_mock()
     self.n_test = 10
     self.family_docs = self.db.traits.find().limit(self.n_test)
Пример #13
0
 def setUp(self):
     self.enforcer = lambda x: len(x.get("citedby", [])) > 100
     self.db = get_mock()
     self.n_test = 5
     self.family_docs = self.db.traits.find().limit(self.n_test)