add test files
This commit is contained in:
parent
26ea272114
commit
9c5a1b18f0
@ -19,6 +19,7 @@ from io import StringIO
|
||||
class Test_Persistence(unittest.TestCase):
|
||||
|
||||
def setUp(self):
|
||||
|
||||
if not os.path.exists("test_output"):
|
||||
os.mkdir("test_output")
|
||||
|
||||
@ -40,7 +41,7 @@ class Test_Persistence(unittest.TestCase):
|
||||
os.remove(test_file)
|
||||
|
||||
call = self.base_call.format(self.input_file, self.test_output_dir)
|
||||
call = call + " --url-encode --persistence sequence --collapse-user"
|
||||
call = call + " --url-encode --persistence sequence"
|
||||
print(os.path.abspath('.'))
|
||||
print(call)
|
||||
proc = subprocess.Popen(call,stdout=subprocess.PIPE,shell=True)
|
||||
@ -51,17 +52,17 @@ class Test_Persistence(unittest.TestCase):
|
||||
|
||||
# as a test let's make sure that we get equal data frames
|
||||
test = pd.read_table(test_file)
|
||||
self.assertEqual(test['tokens_added'][0],4)
|
||||
self.assertEqual(test['tokens_added'][1],5)
|
||||
self.assertEqual(test['tokens_added'][0],7)
|
||||
self.assertEqual(test['tokens_added'][1],10)
|
||||
self.assertEqual(test['tokens_added'][2],0)
|
||||
self.assertEqual(test['tokens_added'][3],6)
|
||||
self.assertEqual(test['tokens_added'][4],4)
|
||||
self.assertEqual(test['tokens_added'][3],8)
|
||||
self.assertEqual(test['tokens_added'][4],0)
|
||||
self.assertEqual(test['tokens_removed'][0],0)
|
||||
self.assertEqual(test['tokens_removed'][1],0)
|
||||
self.assertEqual(test['tokens_removed'][2],5)
|
||||
self.assertEqual(test['tokens_removed'][2],0)
|
||||
self.assertEqual(test['tokens_removed'][3],4)
|
||||
self.assertEqual(test['tokens_removed'][4],6)
|
||||
self.assertEqual(test['token_revs'][0],4*3)
|
||||
self.assertEqual(test['tokens_removed'][4],0)
|
||||
self.assertEqual(test['token_revs'][0],8*3)
|
||||
self.assertEqual(test['token_revs'][1],0)
|
||||
self.assertEqual(test['token_revs'][2],0)
|
||||
self.assertEqual(test['token_revs'][3],0)
|
||||
@ -69,7 +70,44 @@ class Test_Persistence(unittest.TestCase):
|
||||
|
||||
baseline = pd.read_table(baseline_file)
|
||||
assert_frame_equal(test,baseline)
|
||||
|
||||
def test_legacy_persistence(self):
|
||||
test_filename = "legacy-" + self.wikiq_out_name
|
||||
test_file = os.path.join(self.test_output_dir, test_filename)
|
||||
if os.path.exists(test_file):
|
||||
os.remove(test_file)
|
||||
|
||||
call = self.base_call.format(self.input_file, self.test_output_dir)
|
||||
call = call + " --url-encode --persistence legacy"
|
||||
print(os.path.abspath('.'))
|
||||
print(call)
|
||||
proc = subprocess.Popen(call,stdout=subprocess.PIPE,shell=True)
|
||||
proc.wait()
|
||||
|
||||
copyfile(self.call_output, test_file)
|
||||
baseline_file = os.path.join(".", self.baseline_output_dir, test_filename)
|
||||
|
||||
# as a test let's make sure that we get equal data frames
|
||||
test = pd.read_table(test_file)
|
||||
self.assertEqual(test['tokens_added'][0],7)
|
||||
self.assertEqual(test['tokens_added'][1],10)
|
||||
self.assertEqual(test['tokens_added'][2],0)
|
||||
self.assertEqual(test['tokens_added'][3],11)
|
||||
self.assertEqual(test['tokens_added'][4],0)
|
||||
self.assertEqual(test['tokens_removed'][0],0)
|
||||
self.assertEqual(test['tokens_removed'][1],0)
|
||||
self.assertEqual(test['tokens_removed'][2],0)
|
||||
self.assertEqual(test['tokens_removed'][3],7)
|
||||
self.assertEqual(test['tokens_removed'][4],0)
|
||||
self.assertEqual(test['token_revs'][0],7*3)
|
||||
self.assertEqual(test['token_revs'][1],0)
|
||||
self.assertEqual(test['token_revs'][2],0)
|
||||
self.assertEqual(test['token_revs'][3],0)
|
||||
self.assertEqual(test['token_revs'][4],0)
|
||||
|
||||
baseline = pd.read_table(baseline_file)
|
||||
assert_frame_equal(test,baseline)
|
||||
|
||||
|
||||
|
||||
class Test_Persistence_Bug(unittest.TestCase):
|
||||
@ -135,6 +173,7 @@ class Test_Wikipedia(unittest.TestCase):
|
||||
|
||||
call = self.base_call.format(self.input_file, self.test_output_dir)
|
||||
call = call + " --url-encode"
|
||||
print(call)
|
||||
proc = subprocess.Popen(call,stdout=subprocess.PIPE,shell=True)
|
||||
proc.wait()
|
||||
|
||||
@ -195,6 +234,7 @@ class Test_Basic(unittest.TestCase):
|
||||
os.remove(test_file)
|
||||
|
||||
call = self.base_call.format(self.input_file, self.test_output_dir)
|
||||
print(call)
|
||||
proc = subprocess.Popen(call,stdout=subprocess.PIPE,shell=True)
|
||||
proc.wait()
|
||||
|
||||
@ -216,6 +256,7 @@ class Test_Basic(unittest.TestCase):
|
||||
call = self.base_call.format(self.input_file, self.test_output_dir)
|
||||
call = call + " --collapse-user"
|
||||
|
||||
print(call)
|
||||
proc = subprocess.Popen(call,stdout=subprocess.PIPE,shell=True)
|
||||
proc.wait()
|
||||
|
||||
@ -256,6 +297,7 @@ class Test_Basic(unittest.TestCase):
|
||||
|
||||
call = self.base_call.format(self.input_file, self.test_output_dir)
|
||||
call = call + " --persistence legacy"
|
||||
print(call)
|
||||
proc = subprocess.Popen(call,stdout=subprocess.PIPE,shell=True)
|
||||
proc.wait()
|
||||
|
||||
@ -275,6 +317,7 @@ class Test_Basic(unittest.TestCase):
|
||||
|
||||
call = self.base_call.format(self.input_file, self.test_output_dir)
|
||||
call = call + " --persistence"
|
||||
print(call)
|
||||
proc = subprocess.Popen(call,stdout=subprocess.PIPE,shell=True)
|
||||
proc.wait()
|
||||
|
||||
@ -297,7 +340,9 @@ class Test_Basic(unittest.TestCase):
|
||||
|
||||
call = self.base_call.format(self.input_file, self.test_output_dir)
|
||||
call = call + " --url-encode"
|
||||
print(call)
|
||||
proc = subprocess.Popen(call,stdout=subprocess.PIPE,shell=True)
|
||||
|
||||
proc.wait()
|
||||
|
||||
copyfile(self.call_output, test_file)
|
||||
@ -326,6 +371,7 @@ class Test_Malformed(unittest.TestCase):
|
||||
def test_malformed_noargs(self):
|
||||
|
||||
call = self.base_call.format(self.input_file, self.test_output_dir)
|
||||
print(call)
|
||||
proc = subprocess.Popen(call,stdout=subprocess.PIPE,stderr=subprocess.PIPE, shell=True)
|
||||
proc.wait()
|
||||
outs, errs = proc.communicate()
|
||||
@ -347,6 +393,7 @@ class Test_Stdout(unittest.TestCase):
|
||||
def test_noargs(self):
|
||||
|
||||
call = self.base_call.format(self.input_file)
|
||||
print(call)
|
||||
proc = subprocess.run(call,stdout=subprocess.PIPE,shell=True)
|
||||
outs = proc.stdout.decode("utf8")
|
||||
|
||||
|
||||
15
test/baseline_output/enwiki-test.tsv
Normal file
15
test/baseline_output/enwiki-test.tsv
Normal file
@ -0,0 +1,15 @@
|
||||
anon articleid collapsed_revs date_time deleted editor editor_id minor namespace revert reverteds revid sha1 text_chars title token_revs tokens_added tokens_removed tokens_window
|
||||
TRUE 23397386 3 2010-12-13 02:52:57 FALSE %2276.102.106.52%22 FALSE 0 FALSE 402074089 bdoleau92eed59p01lwu3hx71jfplyv 54928 %22The%20Amazing%20Race%2016%22 118658 19778 0 6
|
||||
FALSE 23397386 1 2010-12-13 04:44:25 FALSE %22Nmk829%22 12595469 FALSE 0 FALSE 402088763 7zt0ygstumo5fe0fj2xj32mcn1p4p1j 54926 %22The%20Amazing%20Race%2016%22 0 0 0 6
|
||||
FALSE 23397386 1 2010-12-14 18:31:01 FALSE %22Ryulong%22 889851 FALSE 0 FALSE 402376185 2itumo8i1amz1anlmnvnqcaauk0c6ti 54912 %22The%20Amazing%20Race%2016%22 0 0 6 6
|
||||
TRUE 23397386 1 2010-12-14 23:12:50 FALSE %22174.1.48.24%22 FALSE 0 FALSE 402421155 27izchr7l6uj5z0p9jc9w8mf5h45w8p 54973 %22The%20Amazing%20Race%2016%22 114 19 0 6
|
||||
FALSE 23397386 1 2010-12-14 23:18:49 FALSE %22Ryulong%22 889851 FALSE 0 FALSE 402422015 60nmgy7reem6isznmtnjr37qwj3akkb 55013 %22The%20Amazing%20Race%2016%22 102 17 0 6
|
||||
TRUE 23397386 1 2010-12-15 01:05:24 FALSE %22173.30.1.184%22 FALSE 0 FALSE 402436936 2rf7juf6sx1mi76rgnagt0v1h2vqkka 55033 %22The%20Amazing%20Race%2016%22 0 10 0 6
|
||||
FALSE 23397386 1 2010-12-15 04:03:25 FALSE %22Kuyabribri%22 1904358 FALSE 0 TRUE "402436936" 402459131 60nmgy7reem6isznmtnjr37qwj3akkb 55013 %22The%20Amazing%20Race%2016%22 0 0 0 6
|
||||
TRUE 23397386 1 2010-12-16 15:24:08 FALSE %22220.255.1.79%22 FALSE 0 FALSE 402702013 p2u5jcoov0c3396cdih5h2v8azrmb5z 55010 %22The%20Amazing%20Race%2016%22 3320 1030 7 6
|
||||
TRUE 23397386 1 2010-12-18 10:45:33 FALSE %2258.168.198.160%22 FALSE 0 FALSE 402999028 h9l6spwatmurwxcwcukky37iiozb8dq 55014 %22The%20Amazing%20Race%2016%22 0 2 0 5
|
||||
FALSE 23397386 1 2010-12-19 01:05:33 FALSE %22Dravecky%22 4588674 TRUE 0 TRUE "402999028" 403103973 p2u5jcoov0c3396cdih5h2v8azrmb5z 55010 %22The%20Amazing%20Race%2016%22 0 0 0 4
|
||||
FALSE 23397386 1 2010-12-22 12:21:27 FALSE %22Xqbot%22 8066546 TRUE 0 FALSE 403690886 lar0epx20k4qvdj8otqna8v9c1ibtbt 55008 %22The%20Amazing%20Race%2016%22 3162 1054 26 3
|
||||
TRUE 23397386 1 2011-01-02 10:37:54 FALSE %22118.137.242.186%22 FALSE 0 FALSE 405475313 ipyttx2y54xlilzeihwuaoatcmbcotz 54974 %22The%20Amazing%20Race%2016%22 0 0 0 2
|
||||
TRUE 23397386 2 2011-01-02 11:08:22 FALSE %2271.133.162.137%22 FALSE 0 FALSE 405478231 lsxi138j1q7qmim71u33kx0qjb4ema6 54976 %22The%20Amazing%20Race%2016%22 0 0 0 1
|
||||
FALSE 23397386 2 2011-01-02 12:19:03 FALSE %22Dravecky%22 4588674 FALSE 0 TRUE "405478231,405475313" 405484921 lar0epx20k4qvdj8otqna8v9c1ibtbt 55008 %22The%20Amazing%20Race%2016%22 0 0 0 0
|
||||
|
6
test/baseline_output/pwr-test.tsv
Normal file
6
test/baseline_output/pwr-test.tsv
Normal file
@ -0,0 +1,6 @@
|
||||
anon articleid date_time deleted editor editor_id minor namespace revert reverteds revid sha1 text_chars title token_revs tokens_added tokens_removed tokens_window
|
||||
TRUE 1 2010-12-13 02:51:55 FALSE %2276.102.106.52%22 FALSE 0 FALSE None e867e7cd0e53b58428b23ee346c6fa523756e7d1 21 %22Test%20Title%22 24 7 0 4
|
||||
TRUE 1 2010-12-13 02:51:55 FALSE %2276.102.106.52%22 FALSE 0 FALSE None cb3aeff37c151ab439ee4a7a76cea85679272d5e 44 %22Test%20Title%22 0 10 0 3
|
||||
TRUE 1 2010-12-13 02:51:55 FALSE %2276.102.106.52%22 FALSE 0 FALSE None cp0mfp4o90fbpyxooys1rw95zn3ddff 21 %22Test%20Title%22 0 0 0 2
|
||||
TRUE 1 2010-12-13 02:51:55 FALSE %2276.102.106.53%22 FALSE 0 FALSE None 9e63d8f3d87be0ed4cc3fab3f750a4429997bead 31 %22Test%20Title%22 0 8 4 1
|
||||
TRUE 1 2010-12-13 02:51:55 FALSE %2276.102.106.60%22 FALSE 0 TRUE "None,None,None" None e867e7cd0e53b58428b23ee346c6fa523756e7d1 21 %22Test%20Title%22 0 0 0 0
|
||||
|
15
test/baseline_output/sequence-enwiki-test.tsv
Normal file
15
test/baseline_output/sequence-enwiki-test.tsv
Normal file
@ -0,0 +1,15 @@
|
||||
anon articleid collapsed_revs date_time deleted editor editor_id minor namespace revert reverteds revid sha1 text_chars title token_revs tokens_added tokens_removed tokens_window
|
||||
TRUE 23397386 3 2010-12-13 02:52:57 FALSE %2276.102.106.52%22 FALSE 0 FALSE 402074089 bdoleau92eed59p01lwu3hx71jfplyv 54928 %22The%20Amazing%20Race%2016%22 118658 19778 0 6
|
||||
FALSE 23397386 1 2010-12-13 04:44:25 FALSE %22Nmk829%22 12595469 FALSE 0 FALSE 402088763 7zt0ygstumo5fe0fj2xj32mcn1p4p1j 54926 %22The%20Amazing%20Race%2016%22 0 0 0 6
|
||||
FALSE 23397386 1 2010-12-14 18:31:01 FALSE %22Ryulong%22 889851 FALSE 0 FALSE 402376185 2itumo8i1amz1anlmnvnqcaauk0c6ti 54912 %22The%20Amazing%20Race%2016%22 0 0 6 6
|
||||
TRUE 23397386 1 2010-12-14 23:12:50 FALSE %22174.1.48.24%22 FALSE 0 FALSE 402421155 27izchr7l6uj5z0p9jc9w8mf5h45w8p 54973 %22The%20Amazing%20Race%2016%22 114 19 0 6
|
||||
FALSE 23397386 1 2010-12-14 23:18:49 FALSE %22Ryulong%22 889851 FALSE 0 FALSE 402422015 60nmgy7reem6isznmtnjr37qwj3akkb 55013 %22The%20Amazing%20Race%2016%22 102 17 0 6
|
||||
TRUE 23397386 1 2010-12-15 01:05:24 FALSE %22173.30.1.184%22 FALSE 0 FALSE 402436936 2rf7juf6sx1mi76rgnagt0v1h2vqkka 55033 %22The%20Amazing%20Race%2016%22 0 10 0 6
|
||||
FALSE 23397386 1 2010-12-15 04:03:25 FALSE %22Kuyabribri%22 1904358 FALSE 0 TRUE "402436936" 402459131 60nmgy7reem6isznmtnjr37qwj3akkb 55013 %22The%20Amazing%20Race%2016%22 0 0 0 6
|
||||
TRUE 23397386 1 2010-12-16 15:24:08 FALSE %22220.255.1.79%22 FALSE 0 FALSE 402702013 p2u5jcoov0c3396cdih5h2v8azrmb5z 55010 %22The%20Amazing%20Race%2016%22 3320 1030 7 6
|
||||
TRUE 23397386 1 2010-12-18 10:45:33 FALSE %2258.168.198.160%22 FALSE 0 FALSE 402999028 h9l6spwatmurwxcwcukky37iiozb8dq 55014 %22The%20Amazing%20Race%2016%22 0 2 0 5
|
||||
FALSE 23397386 1 2010-12-19 01:05:33 FALSE %22Dravecky%22 4588674 TRUE 0 TRUE "402999028" 403103973 p2u5jcoov0c3396cdih5h2v8azrmb5z 55010 %22The%20Amazing%20Race%2016%22 0 0 0 4
|
||||
FALSE 23397386 1 2010-12-22 12:21:27 FALSE %22Xqbot%22 8066546 TRUE 0 FALSE 403690886 lar0epx20k4qvdj8otqna8v9c1ibtbt 55008 %22The%20Amazing%20Race%2016%22 3162 1054 26 3
|
||||
TRUE 23397386 1 2011-01-02 10:37:54 FALSE %22118.137.242.186%22 FALSE 0 FALSE 405475313 ipyttx2y54xlilzeihwuaoatcmbcotz 54974 %22The%20Amazing%20Race%2016%22 0 0 0 2
|
||||
TRUE 23397386 2 2011-01-02 11:08:22 FALSE %2271.133.162.137%22 FALSE 0 FALSE 405478231 lsxi138j1q7qmim71u33kx0qjb4ema6 54976 %22The%20Amazing%20Race%2016%22 0 0 0 1
|
||||
FALSE 23397386 2 2011-01-02 12:19:03 FALSE %22Dravecky%22 4588674 FALSE 0 TRUE "405478231,405475313" 405484921 lar0epx20k4qvdj8otqna8v9c1ibtbt 55008 %22The%20Amazing%20Race%2016%22 0 0 0 0
|
||||
|
6
test/baseline_output/sequence-pwr-test.tsv
Normal file
6
test/baseline_output/sequence-pwr-test.tsv
Normal file
@ -0,0 +1,6 @@
|
||||
anon articleid date_time deleted editor editor_id minor namespace revert reverteds revid sha1 text_chars title token_revs tokens_added tokens_removed tokens_window
|
||||
TRUE 1 2010-12-13 02:51:55 FALSE %2276.102.106.52%22 FALSE 0 FALSE None e867e7cd0e53b58428b23ee346c6fa523756e7d1 21 %22Test%20Title%22 12 4 0 4
|
||||
TRUE 1 2010-12-13 02:51:55 FALSE %2276.102.106.52%22 FALSE 0 FALSE None cb3aeff37c151ab439ee4a7a76cea85679272d5e 44 %22Test%20Title%22 0 5 0 3
|
||||
TRUE 1 2010-12-13 02:51:55 FALSE %2276.102.106.52%22 FALSE 0 FALSE None cp0mfp4o90fbpyxooys1rw95zn3ddff 21 %22Test%20Title%22 0 0 0 2
|
||||
TRUE 1 2010-12-13 02:51:55 FALSE %2276.102.106.53%22 FALSE 0 FALSE None 9e63d8f3d87be0ed4cc3fab3f750a4429997bead 31 %22Test%20Title%22 0 6 4 1
|
||||
TRUE 1 2010-12-13 02:51:55 FALSE %2276.102.106.60%22 FALSE 0 TRUE "None,None,None" None e867e7cd0e53b58428b23ee346c6fa523756e7d1 21 %22Test%20Title%22 0 0 0 0
|
||||
|
1597
test/dumps/arwiki-20180301-pages-meta-history.head.xml
Normal file
1597
test/dumps/arwiki-20180301-pages-meta-history.head.xml
Normal file
File diff suppressed because it is too large
Load Diff
100001
test/dumps/arwiki-20180301-pages-meta-history.head.xml~
Normal file
100001
test/dumps/arwiki-20180301-pages-meta-history.head.xml~
Normal file
File diff suppressed because it is too large
Load Diff
11997
test/dumps/enwiki-test.xml
Normal file
11997
test/dumps/enwiki-test.xml
Normal file
File diff suppressed because it is too large
Load Diff
117
test/dumps/enwiki-test.xml~
Normal file
117
test/dumps/enwiki-test.xml~
Normal file
@ -0,0 +1,117 @@
|
||||
<mediawiki xmlns="http://www.mediawiki.org/xml/export-0.10/" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xsi:schemaLocation="http://www.mediawiki.org/xml/export-0.10/ http://www.mediawiki.org/xml/export-0.10.xsd" version="0.10" xml:lang="en">
|
||||
<siteinfo>
|
||||
<sitename>Wikipedia</sitename>
|
||||
<dbname>enwiki</dbname>
|
||||
<base>https://en.wikipedia.org/wiki/Main_Page</base>
|
||||
<generator>MediaWiki 1.31.0-wmf.22</generator>
|
||||
<case>first-letter</case>
|
||||
<namespaces>
|
||||
<namespace key="-2" case="first-letter">Media</namespace>
|
||||
<namespace key="-1" case="first-letter">Special</namespace>
|
||||
<namespace key="0" case="first-letter" />
|
||||
<namespace key="1" case="first-letter">Talk</namespace>
|
||||
<namespace key="2" case="first-letter">User</namespace>
|
||||
<namespace key="3" case="first-letter">User talk</namespace>
|
||||
<namespace key="4" case="first-letter">Wikipedia</namespace>
|
||||
<namespace key="5" case="first-letter">Wikipedia talk</namespace>
|
||||
<namespace key="6" case="first-letter">File</namespace>
|
||||
<namespace key="7" case="first-letter">File talk</namespace>
|
||||
<namespace key="8" case="first-letter">MediaWiki</namespace>
|
||||
<namespace key="9" case="first-letter">MediaWiki talk</namespace>
|
||||
<namespace key="10" case="first-letter">Template</namespace>
|
||||
<namespace key="11" case="first-letter">Template talk</namespace>
|
||||
<namespace key="12" case="first-letter">Help</namespace>
|
||||
<namespace key="13" case="first-letter">Help talk</namespace>
|
||||
<namespace key="14" case="first-letter">Category</namespace>
|
||||
<namespace key="15" case="first-letter">Category talk</namespace>
|
||||
<namespace key="100" case="first-letter">Portal</namespace>
|
||||
<namespace key="101" case="first-letter">Portal talk</namespace>
|
||||
<namespace key="108" case="first-letter">Book</namespace>
|
||||
<namespace key="109" case="first-letter">Book talk</namespace>
|
||||
<namespace key="118" case="first-letter">Draft</namespace>
|
||||
<namespace key="119" case="first-letter">Draft talk</namespace>
|
||||
<namespace key="446" case="first-letter">Education Program</namespace>
|
||||
<namespace key="447" case="first-letter">Education Program talk</namespace>
|
||||
<namespace key="710" case="first-letter">TimedText</namespace>
|
||||
<namespace key="711" case="first-letter">TimedText talk</namespace>
|
||||
<namespace key="828" case="first-letter">Module</namespace>
|
||||
<namespace key="829" case="first-letter">Module talk</namespace>
|
||||
<namespace key="2300" case="first-letter">Gadget</namespace>
|
||||
<namespace key="2301" case="first-letter">Gadget talk</namespace>
|
||||
<namespace key="2302" case="case-sensitive">Gadget definition</namespace>
|
||||
<namespace key="2303" case="case-sensitive">Gadget definition talk</namespace>
|
||||
</namespaces>
|
||||
</siteinfo>
|
||||
|
||||
<page>
|
||||
|
||||
<title>Test Title</title>
|
||||
<ns>0</ns>
|
||||
<id>1</id>
|
||||
<revision>
|
||||
<timestamp>2010-12-13T02:51:55Z</timestamp>
|
||||
<contributor>
|
||||
<ip>76.102.106.52</ip>
|
||||
</contributor>
|
||||
<comment>baseline</comment>
|
||||
<model>wikitext</model>
|
||||
<format>text/x-wiki</format>
|
||||
<text xml:space="preserve">
|
||||
start with four words
|
||||
</text>
|
||||
<sha1>cp0mfp4o90fbpyxooys1rw95zn3ddff</sha1>
|
||||
</revision>
|
||||
<revision>
|
||||
<timestamp>2010-12-13T02:51:55Z</timestamp>
|
||||
<contributor>
|
||||
<ip>76.102.106.52</ip>
|
||||
</contributor>
|
||||
<comment>baseline</comment>
|
||||
<model>wikitext</model>
|
||||
<format>text/x-wiki</format>
|
||||
<text xml:space="preserve">
|
||||
start with four words and then add five more
|
||||
</text>
|
||||
<sha1>cp0mfp4o90fbpyxooys1rw95zn3ddff</sha1>
|
||||
</revision>
|
||||
<revision>
|
||||
<timestamp>2010-12-13T02:51:55Z</timestamp>
|
||||
<contributor>
|
||||
<ip>76.102.106.52</ip>
|
||||
</contributor>
|
||||
<comment>baseline</comment>
|
||||
<model>wikitext</model>
|
||||
<format>text/x-wiki</format>
|
||||
<text xml:space="preserve">
|
||||
start with four words
|
||||
</text>
|
||||
<sha1>cp0mfp4o90fbpyxooys1rw95zn3ddff</sha1>
|
||||
</revision>
|
||||
<revision>
|
||||
<timestamp>2010-12-13T02:51:55Z</timestamp>
|
||||
<contributor>
|
||||
<ip>76.102.106.53</ip>
|
||||
</contributor>
|
||||
<comment>baseline</comment>
|
||||
<model>wikitext</model>
|
||||
<format>text/x-wiki</format>
|
||||
<text xml:space="preserve">
|
||||
this edit will also be reverted
|
||||
</text>
|
||||
<sha1>cp0mfp4o90fbpyxooys1rw95zn3ddff</sha1>
|
||||
</revision>
|
||||
<revision>
|
||||
<timestamp>2010-12-13T02:51:55Z</timestamp>
|
||||
<contributor>
|
||||
<ip>76.102.106.60</ip>
|
||||
</contributor>
|
||||
<comment>baseline</comment>
|
||||
<model>wikitext</model>
|
||||
<format>text/x-wiki</format>
|
||||
<text xml:space="preserve">
|
||||
start with four words
|
||||
</text>
|
||||
<sha1>cp0mfp4o90fbpyxooys1rw95zn3ddff</sha1>
|
||||
</revision>
|
||||
</page>
|
||||
</mediawiki>
|
||||
BIN
test/dumps/ikwiki-small.xml.bz2
Normal file
BIN
test/dumps/ikwiki-small.xml.bz2
Normal file
Binary file not shown.
BIN
test/dumps/ikwiki-small.xml.bz2~
Normal file
BIN
test/dumps/ikwiki-small.xml.bz2~
Normal file
Binary file not shown.
102
test/dumps/pwr-test.xml
Normal file
102
test/dumps/pwr-test.xml
Normal file
@ -0,0 +1,102 @@
|
||||
<mediawiki xmlns="http://www.mediawiki.org/xml/export-0.10/" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xsi:schemaLocation="http://www.mediawiki.org/xml/export-0.10/ http://www.mediawiki.org/xml/export-0.10.xsd" version="0.10" xml:lang="en">
|
||||
<siteinfo>
|
||||
<sitename>Wikipedia</sitename>
|
||||
<dbname>enwiki</dbname>
|
||||
<base>https://en.wikipedia.org/wiki/Main_Page</base>
|
||||
<generator>MediaWiki 1.31.0-wmf.22</generator>
|
||||
<case>first-letter</case>
|
||||
<namespaces>
|
||||
<namespace key="-2" case="first-letter">Media</namespace>
|
||||
<namespace key="-1" case="first-letter">Special</namespace>
|
||||
<namespace key="0" case="first-letter" />
|
||||
<namespace key="1" case="first-letter">Talk</namespace>
|
||||
<namespace key="2" case="first-letter">User</namespace>
|
||||
<namespace key="3" case="first-letter">User talk</namespace>
|
||||
<namespace key="4" case="first-letter">Wikipedia</namespace>
|
||||
<namespace key="5" case="first-letter">Wikipedia talk</namespace>
|
||||
<namespace key="6" case="first-letter">File</namespace>
|
||||
<namespace key="7" case="first-letter">File talk</namespace>
|
||||
<namespace key="8" case="first-letter">MediaWiki</namespace>
|
||||
<namespace key="9" case="first-letter">MediaWiki talk</namespace>
|
||||
<namespace key="10" case="first-letter">Template</namespace>
|
||||
<namespace key="11" case="first-letter">Template talk</namespace>
|
||||
<namespace key="12" case="first-letter">Help</namespace>
|
||||
<namespace key="13" case="first-letter">Help talk</namespace>
|
||||
<namespace key="14" case="first-letter">Category</namespace>
|
||||
<namespace key="15" case="first-letter">Category talk</namespace>
|
||||
<namespace key="100" case="first-letter">Portal</namespace>
|
||||
<namespace key="101" case="first-letter">Portal talk</namespace>
|
||||
<namespace key="108" case="first-letter">Book</namespace>
|
||||
<namespace key="109" case="first-letter">Book talk</namespace>
|
||||
<namespace key="118" case="first-letter">Draft</namespace>
|
||||
<namespace key="119" case="first-letter">Draft talk</namespace>
|
||||
<namespace key="446" case="first-letter">Education Program</namespace>
|
||||
<namespace key="447" case="first-letter">Education Program talk</namespace>
|
||||
<namespace key="710" case="first-letter">TimedText</namespace>
|
||||
<namespace key="711" case="first-letter">TimedText talk</namespace>
|
||||
<namespace key="828" case="first-letter">Module</namespace>
|
||||
<namespace key="829" case="first-letter">Module talk</namespace>
|
||||
<namespace key="2300" case="first-letter">Gadget</namespace>
|
||||
<namespace key="2301" case="first-letter">Gadget talk</namespace>
|
||||
<namespace key="2302" case="case-sensitive">Gadget definition</namespace>
|
||||
<namespace key="2303" case="case-sensitive">Gadget definition talk</namespace>
|
||||
</namespaces>
|
||||
</siteinfo>
|
||||
|
||||
<page>
|
||||
<title>Test Title</title>
|
||||
<ns>0</ns>
|
||||
<id>1</id>
|
||||
<revision>
|
||||
<timestamp>2010-12-13T02:51:55Z</timestamp>
|
||||
<contributor>
|
||||
<ip>76.102.106.52</ip>
|
||||
</contributor>
|
||||
<comment>baseline</comment>
|
||||
<model>wikitext</model>
|
||||
<format>text/x-wiki</format>
|
||||
<text xml:space="preserve">start with four words</text>
|
||||
</revision>
|
||||
<revision>
|
||||
<timestamp>2010-12-13T02:51:55Z</timestamp>
|
||||
<contributor>
|
||||
<ip>76.102.106.52</ip>
|
||||
</contributor>
|
||||
<comment>baseline</comment>
|
||||
<model>wikitext</model>
|
||||
<format>text/x-wiki</format>
|
||||
<text xml:space="preserve">start with four words and then add five more</text>
|
||||
</revision>
|
||||
<revision>
|
||||
<timestamp>2010-12-13T02:51:55Z</timestamp>
|
||||
<contributor>
|
||||
<ip>76.102.106.52</ip>
|
||||
</contributor>
|
||||
<comment>baseline</comment>
|
||||
<model>wikitext</model>
|
||||
<format>text/x-wiki</format>
|
||||
<text xml:space="preserve">start with four words</text>
|
||||
<sha1>cp0mfp4o90fbpyxooys1rw95zn3ddff</sha1>
|
||||
</revision>
|
||||
<revision>
|
||||
<timestamp>2010-12-13T02:51:55Z</timestamp>
|
||||
<contributor>
|
||||
<ip>76.102.106.53</ip>
|
||||
</contributor>
|
||||
<comment>baseline</comment>
|
||||
<model>wikitext</model>
|
||||
<format>text/x-wiki</format>
|
||||
<text xml:space="preserve">this edit will also be reverted</text>
|
||||
</revision>
|
||||
<revision>
|
||||
<timestamp>2010-12-13T02:51:55Z</timestamp>
|
||||
<contributor>
|
||||
<ip>76.102.106.60</ip>
|
||||
</contributor>
|
||||
<comment>baseline</comment>
|
||||
<model>wikitext</model>
|
||||
<format>text/x-wiki</format>
|
||||
<text xml:space="preserve">start with four words</text>
|
||||
</revision>
|
||||
</page>
|
||||
</mediawiki>
|
||||
102
test/dumps/pwr-test.xml~
Normal file
102
test/dumps/pwr-test.xml~
Normal file
@ -0,0 +1,102 @@
|
||||
<mediawiki xmlns="http://www.mediawiki.org/xml/export-0.10/" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xsi:schemaLocation="http://www.mediawiki.org/xml/export-0.10/ http://www.mediawiki.org/xml/export-0.10.xsd" version="0.10" xml:lang="en">
|
||||
<siteinfo>
|
||||
<sitename>Wikipedia</sitename>
|
||||
<dbname>enwiki</dbname>
|
||||
<base>https://en.wikipedia.org/wiki/Main_Page</base>
|
||||
<generator>MediaWiki 1.31.0-wmf.22</generator>
|
||||
<case>first-letter</case>
|
||||
<namespaces>
|
||||
<namespace key="-2" case="first-letter">Media</namespace>
|
||||
<namespace key="-1" case="first-letter">Special</namespace>
|
||||
<namespace key="0" case="first-letter" />
|
||||
<namespace key="1" case="first-letter">Talk</namespace>
|
||||
<namespace key="2" case="first-letter">User</namespace>
|
||||
<namespace key="3" case="first-letter">User talk</namespace>
|
||||
<namespace key="4" case="first-letter">Wikipedia</namespace>
|
||||
<namespace key="5" case="first-letter">Wikipedia talk</namespace>
|
||||
<namespace key="6" case="first-letter">File</namespace>
|
||||
<namespace key="7" case="first-letter">File talk</namespace>
|
||||
<namespace key="8" case="first-letter">MediaWiki</namespace>
|
||||
<namespace key="9" case="first-letter">MediaWiki talk</namespace>
|
||||
<namespace key="10" case="first-letter">Template</namespace>
|
||||
<namespace key="11" case="first-letter">Template talk</namespace>
|
||||
<namespace key="12" case="first-letter">Help</namespace>
|
||||
<namespace key="13" case="first-letter">Help talk</namespace>
|
||||
<namespace key="14" case="first-letter">Category</namespace>
|
||||
<namespace key="15" case="first-letter">Category talk</namespace>
|
||||
<namespace key="100" case="first-letter">Portal</namespace>
|
||||
<namespace key="101" case="first-letter">Portal talk</namespace>
|
||||
<namespace key="108" case="first-letter">Book</namespace>
|
||||
<namespace key="109" case="first-letter">Book talk</namespace>
|
||||
<namespace key="118" case="first-letter">Draft</namespace>
|
||||
<namespace key="119" case="first-letter">Draft talk</namespace>
|
||||
<namespace key="446" case="first-letter">Education Program</namespace>
|
||||
<namespace key="447" case="first-letter">Education Program talk</namespace>
|
||||
<namespace key="710" case="first-letter">TimedText</namespace>
|
||||
<namespace key="711" case="first-letter">TimedText talk</namespace>
|
||||
<namespace key="828" case="first-letter">Module</namespace>
|
||||
<namespace key="829" case="first-letter">Module talk</namespace>
|
||||
<namespace key="2300" case="first-letter">Gadget</namespace>
|
||||
<namespace key="2301" case="first-letter">Gadget talk</namespace>
|
||||
<namespace key="2302" case="case-sensitive">Gadget definition</namespace>
|
||||
<namespace key="2303" case="case-sensitive">Gadget definition talk</namespace>
|
||||
</namespaces>
|
||||
</siteinfo>
|
||||
|
||||
<page>
|
||||
<title>Test Title</title>
|
||||
<ns>0</ns>
|
||||
<id>1</id>
|
||||
<revision>
|
||||
<timestamp>2010-12-13T02:51:55Z</timestamp>
|
||||
<contributor>
|
||||
<ip>76.102.106.52</ip>
|
||||
</contributor>
|
||||
<comment>baseline</comment>
|
||||
<model>wikitext</model>
|
||||
<format>text/x-wiki</format>
|
||||
<text xml:space="preserve">start with four words</text>
|
||||
</revision>
|
||||
<revision>
|
||||
<timestamp>2010-12-13T02:51:55Z</timestamp>
|
||||
<contributor>
|
||||
<ip>76.102.106.52</ip>
|
||||
</contributor>
|
||||
<comment>baseline</comment>
|
||||
<model>wikitext</model>
|
||||
<format>text/x-wiki</format>
|
||||
<text xml:space="preserve">start with four words and then add five more</text>
|
||||
</revision>
|
||||
<revision>
|
||||
<timestamp>2010-12-13T02:51:55Z</timestamp>
|
||||
<contributor>
|
||||
<ip>76.102.106.52</ip>
|
||||
</contributor>
|
||||
<comment>baseline</comment>
|
||||
<model>wikitext</model>
|
||||
<format>text/x-wiki</format>
|
||||
<text xml:space="preserve">start with four words</text>
|
||||
<sha1>cp0mfp4o90fbpyxooys1rw95zn3ddff</sha1>
|
||||
</revision>
|
||||
<revision>
|
||||
<timestamp>2010-12-13T02:51:55Z</timestamp>
|
||||
<contributor>
|
||||
<ip>76.102.106.53</ip>
|
||||
</contributor>
|
||||
<comment>baseline</comment>
|
||||
<model>wikitext</model>
|
||||
<format>text/x-wiki</format>
|
||||
<text xml:space="preserve">this edit will also be reverted</text>
|
||||
</revision>
|
||||
<revision>
|
||||
<timestamp>2010-12-13T02:51:55Z</timestamp>
|
||||
<contributor>
|
||||
<ip>76.102.106.60</ip>
|
||||
</contributor>
|
||||
<comment>baseline</comment>
|
||||
<model>wikitext</model>
|
||||
<format>text/x-wiki</format>
|
||||
<text xml:space="preserve">start with four words</text>
|
||||
</revision>
|
||||
</page>
|
||||
</mediawiki>
|
||||
94
wikiq
94
wikiq
@ -17,7 +17,7 @@ from mwxml import Dump, Page
|
||||
from deltas.tokenizers import wikitext_split
|
||||
from mwdiffs.utilities import dump2diffs
|
||||
import mwpersistence
|
||||
from mwpersistence.state import Version, apply_opdocs, apply_operations, persist_revision_once
|
||||
from mwpersistence.state import DiffState
|
||||
|
||||
from mwpersistence import Token
|
||||
from mwpersistence.utilities import diffs2persistence
|
||||
@ -29,46 +29,8 @@ from deltas import SegmentMatcher
|
||||
TO_ENCODE = ('title', 'editor')
|
||||
PERSISTENCE_RADIUS=7
|
||||
|
||||
# this is a simple override of mwpersistence.DiffState that doesn't do anything special for reverts.
|
||||
class WikiqDiffState(mwpersistence.DiffState):
|
||||
def _update(self, text=None, checksum=None, opdocs=None, revision=None):
|
||||
if checksum is None:
|
||||
if text is None:
|
||||
raise TypeError("Either 'text' or 'checksum' must be " +
|
||||
"specified.")
|
||||
else:
|
||||
checksum = sha1(bytes(text, 'utf8')).hexdigest()
|
||||
|
||||
current_version = Version()
|
||||
|
||||
# the main difference we have is that we don't do anything special for reverts
|
||||
if opdocs is not None:
|
||||
transition = apply_opdocs(opdocs, self.last.tokens or [])
|
||||
current_version.tokens, _, _ = transition
|
||||
else:
|
||||
# NOTICE: HEAVY COMPUTATION HERE!!!
|
||||
#
|
||||
# Diffs usually run in O(n^2) -- O(n^3) time and most
|
||||
# tokenizers produce a lot of tokens.
|
||||
if self.diff_processor is None:
|
||||
raise RuntimeError("DiffState cannot process raw text " +
|
||||
"without a diff_engine specified.")
|
||||
operations, _, current_tokens = \
|
||||
self.diff_processor.process(text, token_class=Token)
|
||||
|
||||
transition = apply_operations(operations,
|
||||
self.last.tokens or [],
|
||||
current_tokens)
|
||||
current_version.tokens, _, _ = transition
|
||||
|
||||
# Record persistence
|
||||
persist_revision_once(current_version.tokens, revision)
|
||||
|
||||
# Update last version
|
||||
self.last = current_version
|
||||
|
||||
# Return the tranisitoned state
|
||||
return transition
|
||||
ws_lex = ['break','whitespace']
|
||||
punct_lex = ['period','qmark','epoint','comma','colon','scolon','paren_open','paren_close','brack_open','brack_close','dbrack_close','dbrack_open','tab_close','tab_open','dcurly_close','dcurly_open','equals','bar','etc','bold','italic','tag','comment_end','comment_start']
|
||||
|
||||
class PersistMethod:
|
||||
none = 0
|
||||
@ -76,20 +38,14 @@ class PersistMethod:
|
||||
segment = 2
|
||||
legacy = 3
|
||||
|
||||
def calculate_persistence(tokens_added, tokens_removed, exclude_ws = True, exclude_punct = False):
|
||||
cond = True
|
||||
ws_lex = ['break','whitespace']
|
||||
punct_lex = ['period','qmark','epoint','comma','colon','scolon','paren_open','paren_close','brack_open','brack_close','dbrack_close','dbrack_open','tab_close','tab_open','dcurly_close','dcurly_open','equals','bar','etc','bold','italic','tag','comment_end','comment_start']
|
||||
def calculate_persistence(tokens_added, tokens_removed, exclude_ws = False, exclude_punct = False, legacy = False):
|
||||
|
||||
if not legacy:
|
||||
cond = lambda t: not (exclude_punct and (t.type in punct_lex)) \
|
||||
and not(exclude_ws and (t.type in ws_lex))
|
||||
|
||||
if exclude_ws:
|
||||
cond = lambda t: cond and not t.type in ws_lex
|
||||
|
||||
if exclude_punct:
|
||||
cond = lambda t: cond and not t.type in punct_lex
|
||||
|
||||
tokens_added = [t for t in tokens_added if cond(t)]
|
||||
tokens_removed = [t for t in tokens_removed if cond(t)]
|
||||
tokens_added = [t for t in tokens_added if cond(t)]
|
||||
tokens_removed = [t for t in tokens_removed if cond(t)]
|
||||
|
||||
return(sum([(len(x.revisions)-1) for x in tokens_added]),
|
||||
len(tokens_added),
|
||||
@ -100,17 +56,18 @@ class WikiqIterator(Dump):
|
||||
|
||||
@classmethod
|
||||
def from_file(cls, fh, collapse_user = False):
|
||||
cls = super(WikiqIterator, cls).from_file(fh)
|
||||
cls.fh = fh
|
||||
cls.collapse_user = collapse_user
|
||||
cls.namespace_map = { ns.id : ns.name for ns in
|
||||
cls.site_info.namespaces }
|
||||
cls = super(WikiqIterator, cls).from_file(fh)
|
||||
return cls
|
||||
|
||||
@classmethod
|
||||
def process_item(cls, item_element, namespace_map, collapse_user = False):
|
||||
def process_item(cls, item_element, namespace_map):
|
||||
if not hasattr(cls,'inv_namespace_map'):
|
||||
cls.inv_namespace_map = {ns.id:name for name, ns in namespace_map.items()}
|
||||
|
||||
if item_element.tag == "page":
|
||||
return WikiqPage.from_element(item_element, namespace_map, collapse_user)
|
||||
return WikiqPage.from_element(item_element, namespace_map, cls.inv_namespace_map, cls.collapse_user)
|
||||
elif item_element.tag == "logitem":
|
||||
return LogItem.from_element(item_element, namespace_map)
|
||||
else:
|
||||
@ -122,11 +79,9 @@ class WikiqPage(Page):
|
||||
'restrictions','collapse_user')
|
||||
|
||||
@classmethod
|
||||
def from_element(cls, item_element, namespace_map, collapse_user = False):
|
||||
def from_element(cls, item_element, namespace_map, inv_namespace_map, collapse_user = False):
|
||||
cls.prev_rev = None
|
||||
|
||||
inv_namespace_map = {ns.id:name for name,ns in namespace_map.items()}
|
||||
|
||||
cls = super(WikiqPage, cls).from_element(item_element, namespace_map)
|
||||
|
||||
# following mwxml, we assume namespace 0 in cases where
|
||||
@ -186,7 +141,8 @@ class WikiqPage(Page):
|
||||
|
||||
if self.collapse_user:
|
||||
collapsed_revs = 1
|
||||
rev.collapsed_revs = collapsed_revs
|
||||
self.prev_rev.collapsed_revs = collapsed_revs
|
||||
prev_rev = self.prev_rev
|
||||
|
||||
for rev in self.revisions:
|
||||
rev = WikiqPage._correct_sha(rev)
|
||||
@ -235,12 +191,6 @@ class WikiqPage(Page):
|
||||
revision.page = self
|
||||
yield revision
|
||||
|
||||
# def __iter__(self):
|
||||
# return self.__revisions
|
||||
|
||||
# def __next__(self):
|
||||
# return next(self.__revisions)
|
||||
|
||||
class WikiqParser():
|
||||
|
||||
def __init__(self, input_file, output_file, collapse_user=False, persist=None, urlencode=False, namespaces = None):
|
||||
@ -315,11 +265,11 @@ class WikiqParser():
|
||||
window = deque(maxlen=PERSISTENCE_RADIUS)
|
||||
|
||||
if self.persist == PersistMethod.sequence:
|
||||
state = WikiqDiffState(SequenceMatcher(tokenizer = wikitext_split),
|
||||
state = DiffState(SequenceMatcher(tokenizer = wikitext_split),
|
||||
revert_radius=PERSISTENCE_RADIUS)
|
||||
|
||||
elif self.persist == PersistMethod.segment:
|
||||
state = WikiqDiffState(SegmentMatcher(tokenizer = wikitext_split),
|
||||
state = DiffState(SegmentMatcher(tokenizer = wikitext_split),
|
||||
revert_radius=PERSISTENCE_RADIUS)
|
||||
|
||||
else:
|
||||
@ -410,7 +360,7 @@ class WikiqParser():
|
||||
if len(window) == PERSISTENCE_RADIUS:
|
||||
old_rev_id, old_rev_data, old_tokens_added, old_tokens_removed = window[0]
|
||||
|
||||
num_token_revs, num_tokens_added, num_tokens_removed = calculate_persistence(old_tokens_added, old_tokens_removed)
|
||||
num_token_revs, num_tokens_added, num_tokens_removed = calculate_persistence(old_tokens_added, old_tokens_removed, legacy = self.persist == PersistMethod.legacy)
|
||||
|
||||
old_rev_data["token_revs"] = num_token_revs
|
||||
old_rev_data["tokens_added"] = num_tokens_added
|
||||
@ -433,7 +383,7 @@ class WikiqParser():
|
||||
|
||||
rev_id, rev_data, tokens_added, tokens_removed = item
|
||||
|
||||
num_token_revs, num_tokens_added, num_tokens_removed = calculate_persistence(tokens_added, tokens_removed)
|
||||
num_token_revs, num_tokens_added, num_tokens_removed = calculate_persistence(tokens_added, tokens_removed, legacy = self.persist == PersistMethod.legacy)
|
||||
|
||||
rev_data["token_revs"] = num_token_revs
|
||||
rev_data["tokens_added"] = num_tokens_added
|
||||
|
||||
Loading…
Reference in New Issue
Block a user