new/additional GWB CDX filter scripts

7 files changed, 142 insertions, 0 deletions

diff --git a/pig/filter-cdx-pdfs.pig b/pig/filter-cdx-pdfs.pig
new file mode 100644
index 0000000..a2882ac
--- /dev/null
+++ b/pig/filter-cdx-pdfs.pig
@@ -0,0 +1,24 @@
+
+-- Tries to filter down a large CDX file (GWB index) to a subset of PDFs, by mimetype.
+--
+-- Author: Bryan Newbold <bnewbold@archive.org>
+-- Date: May 2018
+
+%default INPUT ''
+%default OUTPUT ''
+
+set mapreduce.job.queuename default
+
+cdx = LOAD '$INPUT' AS cdxline:chararray;
+cdx = FILTER cdx BY not STARTSWITH (cdxline, 'filedesc');
+cdx = FILTER cdx BY not STARTSWITH (cdxline, ' ');
+
+cdx = FOREACH cdx GENERATE STRSPLIT(cdxline,'\\s+') as cols, cdxline;
+cdx = FOREACH cdx GENERATE (chararray)cols.$0 as url, (chararray)cols.$1 as timestamp, (chararray)cols.$3 as mimetype, (chararray)cols.$4 as httpstatus, cdxline;
+cdx = FILTER cdx BY not url matches '-';
+cdx = FILTER cdx BY httpstatus matches '200';
+cdx = FILTER cdx BY mimetype matches '.*pdf.*';
+cdx = ORDER cdx by url, timestamp PARALLEL 50;
+cdx = FOREACH cdx GENERATE cdxline;
+STORE cdx INTO '$OUTPUT' USING PigStorage(' ');
+
diff --git a/pig/filter-cdx-ps.pig b/pig/filter-cdx-ps.pig
index 6e80acc..b27a547 100644
--- a/pig/filter-cdx-ps.pig
+++ b/pig/filter-cdx-ps.pig
@@ -1,3 +1,9 @@
+-- Tries to filter down a large CDX file (GWB index) to a subset of postscript
+-- files, by mimetype.
+--
+-- Author: Bryan Newbold <bnewbold@archive.org>
+-- Date: May 2018
+
 %default INPUT ''
 %default OUTPUT ''
 
diff --git a/pig/filter-cdx-source-code-crude.pig b/pig/filter-cdx-source-code-crude.pig
new file mode 100644
index 0000000..589aebd
--- /dev/null
+++ b/pig/filter-cdx-source-code-crude.pig
@@ -0,0 +1,40 @@
+
+-- Tries to filter down a large CDX file (GWB index) to a subset of source code
+-- files by mimetype and file extension.
+-- This is pretty crude and requires the URL to end with the file extension.
+---
+-- Author: Bryan Newbold <bnewbold@archive.org>
+-- Date: October 2019
+
+
+%default INPUT ''
+%default OUTPUT ''
+
+set mapreduce.job.queuename default
+
+cdx = LOAD '$INPUT' AS cdxline:chararray;
+cdx = FILTER cdx BY not STARTSWITH (cdxline, 'filedesc');
+cdx = FILTER cdx BY not STARTSWITH (cdxline, ' ');
+
+cdx = FOREACH cdx GENERATE STRSPLIT(cdxline,'\\s+') as cols, cdxline;
+cdx = FOREACH cdx GENERATE (chararray)cols.$0 as surt, (chararray)cols.$1 as timestamp, (chararray)cols.$3 as mimetype, (chararray)cols.$4 as httpstatus, (chararray)cols.$5 as sha1sum, cdxline;
+cdx = FILTER cdx BY not surt matches '-';
+cdx = FILTER cdx BY httpstatus matches '200';
+cdx = FILTER cdx BY mimetype matches '.*text.*';
+
+-- This is the core regex
+cdx = FILTER cdx
+
+        -- file suffix
+        BY surt matches '.*\\).*\\.(c|h|py|java)';
+
+-- DISTINCT by sha1 column
+cdx_uniq = FOREACH (GROUP cdx BY sha1sum) {
+    r = TOP(1, 0, $1);
+    GENERATE FLATTEN(r);
+};
+
+cdx_uniq = ORDER cdx_uniq by surt, timestamp PARALLEL 50;
+cdx_uniq = FOREACH cdx_uniq GENERATE cdxline;
+STORE cdx_uniq INTO '$OUTPUT' USING PigStorage(' ');
+
diff --git a/pig/filter-cdx-tarball.pig b/pig/filter-cdx-tarball.pig
new file mode 100644
index 0000000..d0be0f7
--- /dev/null
+++ b/pig/filter-cdx-tarball.pig
@@ -0,0 +1,38 @@
+
+-- Tries to filter down a large CDX file (GWB index) to a subset of tarballs
+-- (.tar.gz). Intention is to find software code that isn't in, eg, git.
+--
+-- Author: Bryan Newbold <bnewbold@archive.org>
+-- Date: May 2018
+
+
+%default INPUT ''
+%default OUTPUT ''
+
+set mapreduce.job.queuename default
+
+cdx = LOAD '$INPUT' AS cdxline:chararray;
+cdx = FILTER cdx BY not STARTSWITH (cdxline, 'filedesc');
+cdx = FILTER cdx BY not STARTSWITH (cdxline, ' ');
+
+cdx = FOREACH cdx GENERATE STRSPLIT(cdxline,'\\s+') as cols, cdxline;
+cdx = FOREACH cdx GENERATE (chararray)cols.$0 as surt, (chararray)cols.$1 as timestamp, (chararray)cols.$3 as mimetype, (chararray)cols.$4 as httpstatus, (chararray)cols.$5 as sha1sum, cdxline;
+cdx = FILTER cdx BY not surt matches '-';
+cdx = FILTER cdx BY httpstatus matches '200';
+cdx = FILTER cdx BY mimetype matches '.*(octet|gzip|gtar|tgz).*';
+
+-- This is the core regex
+cdx = FILTER cdx
+        -- .tar.gz in URL
+        BY surt matches '(?i).+\\).*\\.tar\\.gz.*';
+
+-- DISTINCT by sha1 column
+cdx_uniq = FOREACH (GROUP cdx BY sha1sum) {
+    r = TOP(1, 0, $1);
+    GENERATE FLATTEN(r);
+};
+
+cdx_uniq = ORDER cdx_uniq by surt, timestamp PARALLEL 50;
+cdx_uniq = FOREACH cdx_uniq GENERATE cdxline;
+STORE cdx_uniq INTO '$OUTPUT' USING PigStorage(' ');
+
diff --git a/pig/tests/files/sourcecode.cdx b/pig/tests/files/sourcecode.cdx
new file mode 100644
index 0000000..eeb397c
--- /dev/null
+++ b/pig/tests/files/sourcecode.cdx
@@ -0,0 +1,6 @@
+# match
+edu,cmu,cs,adm,reports-archive)/anon/usr0/ftp/usr0/anon/2002/cmu-cs-02-119.java 20170706005950 http://reports-archive.adm.cs.cmu.edu/anon/usr0/ftp/usr0/anon/2002/CMU-CS-02-119.java text/plain 200 MQHD36X5MNZPWFNMD5LFOYZSFGCHUN3V - - 361006 17120058 CITESEERX-CRAWL-2017-06-20-20170706004100259-00924-00932-wbgrp-svc284/CITESEERX-CRAWL-2017-06-20-20170706005946792-00926-31209~wbgrp-svc284.us.archive.org~8443.warc.gz
+# no
+fi,tkk,lib)/diss/2001/isbn951225459x/isbn951225459x.pyc 20170705074926 http://lib.tkk.fi/Diss/2001/isbn951225459X/isbn951225459X.pyc text/plain 200 KJBCOT7LGBNIAVGEGPUELK5OK6RTFORR - - 344175 255650124 CITESEERX-CRAWL-2017-06-20-20170705074433815-00129-00138-wbgrp-svc284/CITESEERX-CRAWL-2017-06-20-20170705074843696-00134-31209~wbgrp-svc284.us.archive.org~8443.warc.gz
+# no
+org,oxfordjournals,nar)/cgi/reprint/gkl1060v1.pdf 20170706035441 http://nar.oxfordjournals.org/cgi/reprint/gkl1060v1.pdf text/html 301 OX6MLVDFURLT2KSYCXUYW2PZNOVFSEVF - - 697 49346051 CITESEERX-CRAWL-2017-06-20-20170706034741172-00140-00149-wbgrp-svc285/CITESEERX-CRAWL-2017-06-20-20170706035435634-00148-3671~wbgrp-svc285.us.archive.org~8443.warc.gz
diff --git a/pig/tests/files/tarballs.cdx b/pig/tests/files/tarballs.cdx
new file mode 100644
index 0000000..7a81b79
--- /dev/null
+++ b/pig/tests/files/tarballs.cdx
@@ -0,0 +1,10 @@
+#http://research.fit.edu/sealevelriselibrary/documents/doc_mgr/448/Florida_Keys_Low_Island_Biodiversity_&_SLR_-_Ross_et_al_2009.pdf
+#http://ijs.sgmjournals.org:80/cgi/reprint/54/6/2217.pdf
+#http://eprints.ecs.soton.ac.uk/12020/1/mind-the-semantic-gap.pdf
+#http://eprint.uq.edu.au/archive/00004120/01/R103_Forrester_pp.pdf
+
+# should match 2:
+
+edu,fit,research)/sealevelriselibrary/documents/doc_mgr/448/Florida_Keys_Low_Island_Biodiversity_&_SLR_-_Ross_et_al_2009.pdf 20170706005950 http://mit.edu/file.pdf application/pdf 200 MQHD36X5MNZPWFNMD5LFOYZSFGCHUN3I - - 123 456 CRAWL/CRAWL.warc.gz
+edu,fit,research)/sealevelriselibrary/documents/doc_mgr/448/Florida_Keys_Low_Island_Biodiversity_&_SLR_-_Ross_et_al_2009.tar.gz 20170706005950 http://mit.edu/file.tar.gz application/octet-stream 200 NQHD36X5MNZPWFNMD5LFOYZSFGCHUN3I - - 123 456 CRAWL/CRAWL.warc.gz
+org,sgmjournals,ijs)//cgi/reprint/54/6/2217.tar.gz 20170706005950 http://mit.edu/file.tar.gz application/gzip 200 TQHD36X5MNZPWFNMD5LFOYZSFGCHUN3V - - 123 456 CRAWL/CRAWL.warc.gz
diff --git a/pig/tests/test_filter_software.py b/pig/tests/test_filter_software.py
new file mode 100644
index 0000000..f0ea1b6
--- /dev/null
+++ b/pig/tests/test_filter_software.py
@@ -0,0 +1,18 @@
+
+import os
+import unittest
+from pighelper import PigTestHelper
+
+def count_lines(s):
+    return len([l for l in s.strip().split('\n') if len(l) > 0])
+
+class TestFilterCDXSoftware(PigTestHelper):
+
+    def test_tarballs(self):
+        r = self.run_pig("filter-cdx-tarball.pig", "tests/files/tarballs.cdx")
+        assert count_lines(r) == 2
+
+    def test_source_code(self):
+        r = self.run_pig("filter-cdx-source-code-crude.pig", "tests/files/sourcecode.cdx")
+        assert count_lines(r) == 1
+