sys.path.append("../utils")
sys.path.append("../ringo-engine-python")
import ringo
import snap
import testutils
if __name__ == '__main__':
if len(sys.argv) < 3:
print "Usage: " + sys.argv[0] + " <srcfile1> <srcfile2>"
sys.exit(1)
srcfile1 = sys.argv[1]
srcfile2 = sys.argv[2]
ringo = ringo.Ringo()
t = testutils.Timer()
r = testutils.Resource()
t1 = ringo.Load(srcfile1)
t.show("load bin")
r.show("__loadbin__")
S = [("Index", "int")]
t2 = ringo.LoadTableTSV(S, srcfile2)
t.show("load text")
r.show("__loadtext__")
t3 = ringo.Join(t1, t2, "Src", "Index")
t.show("join")
r.show("__join__")
dstfile = sys.argv[2]
ringo = ringo.Ringo()
t = testutils.Timer()
r = testutils.Resource()
S = [("Id", "int"), ("OwnerUserId", "int"), ("AcceptedAnswerId", "int"), ("CreationDate", "string"), ("Score", "int"), ("Tag", "string")]
table = ringo.LoadTableTSV(S, srcfile)
t.show("load posts text")
.show("__loadpoststext__")
questions = ringo.Select(table, "Tag = 'python'", False)
t.show("selected tag = 'python'")
r.show("__selectedtagpython__")
qa = ringo.Join(questions, table, "AcceptedAnswerId", "Id")
graph = ringo.ToGraph(qa, "OwnerUserId-2", "OwnerUserId-1")
t.show("join")
r.show("__join__")
t.show("graph")
r.show("__graph__")
prtable = ringo.PageRank(graph, 'PageRank', False, 0.85, 1e-4, 100)
t.show("pagerank")
r.show("__pagerank__")
ringo.SaveTableBinary(prtable)
t.show("save bin")
r.show("__savebin__")
tagsfile = sys.argv[2]
ringo = ringo.Ringo()
t = testutils.Timer()
r = testutils.Resource()
S = [("Id", "int"), ("OwnerUserId", "int"), ("AcceptedAnswerId", "int"),
("CreationDate", "int"), ("Score", "int")]
posts = ringo.LoadTableTSV(S, postsfile)
t.show("load posts text")
r.show("__loadpoststext__")
S = [("PostId", "int"), ("Tag", "string")]
tags = ringo.LoadTableTSV(S, tagsfile)
t.show("load tags text")
r.show("__loadtagstext__")
table = ringo.Join(posts, tags, "Id", "PostId")
t.show("join")
r.show("__join__")
table2 = ringo.Join(table, posts, "AcceptedAnswerId", "Id")
t.show("join")
r.show("__join__")
table2.SetSrcCol("OwnerUserId-1")
table2.SetDstCol("OwnerUserId-2")
gseq = table2.ToGraphPerGroup("Tag", snap.aaFirst)
t.show("graphseq", gseq)
r.show("__graphseq__")
S = [('PostId', 'int'), ('UserId', 'int'), ('AnswerId', 'int'),
('CreationDate', 'string')]
t1 = ringo.LoadTableTSV(S, os.path.join(srcdir, POSTS_FILE))
t.show("load posts", t1)
# Load tags
S = [('PostId', 'int'), ('Tag', 'string')]
t2 = ringo.LoadTableTSV(S, os.path.join(srcdir, TAGS_FILE))
t.show("load tags", t2)
# Select
ringo.Select(t2, 'Tag = python', CompConstant=True)
t.show("select", t2)
# Join
t3 = ringo.Join(t1, t2, 'PostId', 'PostId')
t.show("join", t3)
# Join
t4 = ringo.Join(t3, t1, "1.AnswerId", "PostId")
t.show("join", t4)
# Graph
graph = ringo.ToGraph(t4, "1_2.1.UserId", "1.UserId")
t.show("graph", graph)
ringo.ShowMetadata(graph)
# Get authority scores
(HTHub, HTAuth) = ringo.GetHits(graph)
t.show("hits", graph)
ringo.ShowMetadata(HTAuth)
t = testutils.Timer(ENABLE_TIMER)
# Load posts
S = [('PostId', 'int'), ('UserId', 'int'), ('AnswerId', 'int'),
('CreationDate', 'string')]
POSTS = ringo.LoadTableTSV(S, os.path.join(srcdir, POSTS_FILE))
t.show("load posts", POSTS)
# Load tags
S = [('PostId', 'int'), ('Tag', 'string')]
T = ringo.LoadTableTSV(S, os.path.join(srcdir, TAGS_FILE))
t.show("load tags", T)
# Join
P = ringo.Join(POSTS, T, "PostId", "PostId")
t.show("join", P)
# Select Java posts
print ringo.GetSchema(P)
ringo.Select(P, '2.Tag = java', CompConstant=True)
t.show("select", P)
# Select Questions
Q = ringo.Select(P, '1.AnswerId != 0', InPlace=False, CompConstant=True)
t.show("select", Q)
# Select Answers
A = ringo.Select(P, '1.AnswerId = 0', InPlace=False, CompConstant=True)
t.show("select", A)
dstdir = sys.argv[2] if len(sys.argv) >= 3 else None
if not dstdir is None:
try:
os.makedirs(dstdir)
except OSError:
pass
ringo = ringo.Ringo()
t = testutils.Timer(ENABLE_TIMER)
t1 = ringo.LoadTableBinary('t1')
t3 = ringo.LoadTableBinary('t3')
# Join
t4 = ringo.Join(t3, t1, "AnswerId", "PostId")
t.show("join", t4)
# Graph
graph = ringo.ToGraph(t4, "1_2.1.UserId", "1.UserId")
t.show("graph", graph)
ringo.ShowMetadata(graph)
# Get authority scores
(HTHub, HTAuth) = ringo.GetHits(graph)
t.show("hits", graph)
ringo.ShowMetadata(HTAuth)
t5 = ringo.TableFromHashMap(HTAuth, "UserId", "Authority")
t.show("authority score", t5)
import sys
import ringo
posts_file = sys.argv[1]
tags_file = sys.argv[2]
ringo = ringo.Ringo()
S_posts = [('Id','int'), ('PostTypeId','int'), ('AcceptedAnswerId','int'), ('OwnerUserId','int'), ('Body','string')]
S_tags = [('Id','int'), ('Tag','string')]
posts = ringo.LoadTableTSV(S_posts, posts_file)
ringo.Select(posts, 'OwnerUserId > 0')
#print ringo.ringo.DumpTableContent(posts) # buggy...
q_tags = ringo.LoadTableTSV(S_tags, tags_file)
QT = ringo.Join(posts, q_tags, "Id", "Id")
QT = ringo.Project(QT, ['Id', 'PostTypeId', 'AcceptedAnswerId', 'OwnerUserId', 'Body', 'Tag'])
a_tags = ringo.Project(QT, ['AcceptedAnswerId', 'Tag'], False)
AT = ringo.Join(posts, a_tags, 'Id', 'AcceptedAnswerId')
AT = ringo.Project(AT, ['Id', 'PostTypeId', 'AcceptedAnswerId', 'OwnerUserId', 'Body', 'Tag'])
T = ringo.UnionAll(QT, AT)
ringo.SaveTableTSV(T, 'so_posts.tsv')
import ringo
import sys
src_file = sys.argv[1]
Schema = [('Id', 'int'), ('PostTypeId', 'int'), ('AcceptedAnswerId', 'int'),
('OwnerUserId', 'int'), ('Body', 'string'), ('Tag', 'string')]
ringo = ringo.Ringo()
P = ringo.LoadTableTSV(Schema, src_file, '\t', True)
ringo.Project(P,
['Id', 'PostTypeId', 'AcceptedAnswerId', 'OwnerUserId', 'Tag'])
JP = ringo.Select(P, "Tag = 'java'", False)
Q = ringo.Select(JP, 'PostTypeId = 1', False)
A = ringo.Select(JP, 'PostTypeId = 2', False)
QA = ringo.Join(Q, A, 'AcceptedAnswerId', 'Id')
G = ringo.ToGraph(QA, 'OwnerUserId-1', 'OwnerUserId-2')
PR_MAP = ringo.PageRank(G) # a hash map object: node/user id -> PageRank score
PR = ringo.TableFromHashMap(PR_MAP, 'user', 'score')
PR = ringo.Order(PR, ['score'])
ringo.SaveTableTSV(PR, 'scores.tsv')
#ringo.SaveTableBinary(PR, 'scores')
ringo.GenerateProvenance(G, 'G.py')
import testutils
WINDOW_SIZE = 2592000 # 1 month
if __name__ == '__main__':
if len(sys.argv) < 2:
print """Usage: """ + sys.argv[0] + """ <srcfile>
srcfile: posts.tsv file from StackOverflow dataset"""
sys.exit(1)
srcfile = sys.argv[1]
ringo = ringo.Ringo()
t = testutils.Timer()
r = testutils.Resource()
S = [("Id", "int"), ("OwnerUserId", "int"), ("AcceptedAnswerId", "int"), ("CreationDate", "int"), ("Score", "int")]
table = ringo.LoadTableTSV(S, srcfile)
t.show("load text")
r.show("__loadtext__")
table2 = ringo.Join(table, table, "AcceptedAnswerId", "Id")
t.show("join")
r.show("__join__")
table2.SetSrcCol("OwnerUserId-1")
table2.SetDstCol("OwnerUserId-2")
gseq = table2.ToGraphSequence("CreationDate", snap.aaFirst, WINDOW_SIZE, WINDOW_SIZE)
t.show("graphseq", gseq)
r.show("__graphseq__")