print(get_usage())
sys.exit(1)
if not dstDir is None:
try:
os.makedirs(dstDir)
except OSError:
pass
t = testutils.Timer(ENABLE_TIMER)
ringo = ringo.Ringo()
# Load data
# >>> authors = ringo.load('authors.tsv')
S1 = [("userid1", "string"), ("userid2", "string"), ("created_at", "int")]
Tfollow = ringo.LoadTableTSV(S1, file_cache[TFOLLOW])
t.show("load follow")
# UserId, Owner of Repo, Name of repo, created_at => (owner, name) uniquely identies a repo
S2 = [("userid", "string"), ("owner", "string"), ("name", "string"),
("created_at", "int")]
Tcollab = ringo.LoadTableTSV(S2, file_cache[TCOLLAB])
t.show("load collab")
S3 = [("userid", "string"), ("owner", "string"), ("name", "string"),
("pullid", "int"), ("status", "string"), ("created_at", "int")]
Tpull = ringo.LoadTableTSV(S3, file_cache[TPULL])
t.show("load pull")
Tfork = ringo.LoadTableTSV(S2, file_cache[TFORK])
t.show("load fork")
sys.path.append("../utils")
sys.path.append("../ringo-engine-python")
import ringo
import snap
import testutils
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", "string"), ("Score", "int")]
table = ringo.LoadTableTSV(S, srcfile)
t.show("load text")
r.show("__loadtext__")
table = table.IsNextK("CreationDate", 1, "OwnerUserId")
t.show("isnextk")
r.show("__isnextk__")
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__")
srcdir = sys.argv[1]
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)
# Load posts
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)
if len(sys.argv) < 2:
print """Usage: """ + sys.argv[0] + """ <postsfile> <tagsfile>
postsfile: posts.tsv file from StackOverflow dataset
tagsfile: tags.tsv file from StackOverflow dataset"""
sys.exit(1)
postsfile = sys.argv[1]
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__")
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')