merge backfill into extraction directory

7 files changed, 308 insertions, 1 deletions

diff --git a/extraction/README.md b/extraction/README.md
index 1da1454..b063fba 100644
--- a/extraction/README.md
+++ b/extraction/README.md
@@ -1,5 +1,4 @@
 
-
 ## Development and Testing
 
 Requires (eg, via `apt`):
@@ -17,3 +16,26 @@ Run the tests with:
 
 TODO: GROBID and HBase during development?
 
+## Backfill
+
+An example actually connecting to HBase from a local machine, with thrift
+running on a devbox:
+
+    ./backfill_hbase_from_cdx.py tests/files/example.cdx \
+        --hbase-table wbgrp-journal-extract-0-qa \
+        --hbase-host bnewbold-dev.us.archive.org
+
+Actual invocation to run on Hadoop cluster (running on an IA devbox, where
+hadoop environment is configured):
+
+    # Create tarball of virtualenv
+    pipenv shell
+    tar -czf backfill-4OmRI0zZ.tar.gz -C /home/bnewbold/.local/share/virtualenvs/backfill-4OmRI0zZ .
+
+    ./backfill_hbase_from_cdx.py \
+        -r hadoop \
+        --hbase-host bnewbold-dev.us.archive.org \
+        --hbase-table wbgrp-journal-extract-0-qa \
+        -c mrjob.conf \
+        --archive backfill-4OmRI0zZ.tar.gz#venv \
+        hdfs:///user/bnewbold/journal_crawl_cdx/citeseerx_crawl_2017.cdx
diff --git a/extraction/TODO b/extraction/TODO
index ed10834..3459752 100644
--- a/extraction/TODO
+++ b/extraction/TODO
@@ -1,2 +1,6 @@
+- better test coverage (actually check coverage!)
+- use pre-mapper command to filter down, eg, by status type?
+- automation/docs for bundling virtualenv along
+- think about speedups
 - abstract CDX line reading and HBase stuff out into a common library
 - actual GROBID_SERVER="http://wbgrp-svc096.us.archive.org:8070"
diff --git a/extraction/backfill_hbase_from_cdx.py b/extraction/backfill_hbase_from_cdx.py
new file mode 100755
index 0000000..fe37bd5
--- /dev/null
+++ b/extraction/backfill_hbase_from_cdx.py
@@ -0,0 +1,195 @@
+#!/usr/bin/env python3
+"""
+Streaming Hadoop script to import CDX metadata into the HBase fulltext table,
+primarily for URL-agnostic crawl de-duplication. Takes only "fulltext" file
+formats.
+
+Requires:
+- happybase
+- mrjob
+
+TODO:
+- argparse
+- refactor into an object
+- tests in separate file
+- nose tests
+- sentry integration for error reporting
+"""
+
+import sys
+import json
+from datetime import datetime
+import happybase
+import mrjob
+from mrjob.job import MRJob
+
+NORMAL_MIME = (
+    'application/pdf',
+    'application/postscript',
+    'text/html',
+    'text/xml',
+)
+
+def normalize_mime(raw):
+    raw = raw.lower()
+    for norm in NORMAL_MIME:
+        if raw.startswith(norm):
+            return norm
+
+    # Special cases 
+    if raw.startswith('application/xml'):
+        return 'text/xml'
+    if raw.startswith('application/x-pdf'):
+        return 'application/pdf'
+    return None
+
+def test_normalize_mime():
+    assert normalize_mime("asdf") == None
+    assert normalize_mime("application/pdf") == "application/pdf"
+    assert normalize_mime("application/pdf+journal") == "application/pdf"
+    assert normalize_mime("Application/PDF") == "application/pdf"
+    assert normalize_mime("application/p") == None
+    assert normalize_mime("application/xml+stuff") == "text/xml"
+
+def transform_line(raw_cdx):
+
+    cdx = raw_cdx.split()
+    if len(cdx) < 11:
+        return None
+
+    surt = cdx[0]
+    dt = cdx[1]
+    url = cdx[2]
+    mime = normalize_mime(cdx[3])
+    http_status = cdx[4]
+    key = cdx[5]
+    c_size = cdx[8]
+    offset = cdx[9]
+    warc = cdx[10]
+
+    if not (key.isalnum() and c_size.isdigit() and offset.isdigit()
+            and http_status == "200" and len(key) == 32 and dt.isdigit()):
+        return None
+
+    if '-' in (surt, dt, url, mime, http_status, key, c_size, offset, warc):
+        return None
+
+    key = "sha1:{}".format(key)
+
+    info = dict(surt=surt, dt=dt, url=url, c_size=int(c_size),
+        offset=int(offset), warc=warc)
+
+    warc_file = warc.split('/')[-1]
+    dt_iso = datetime.strptime(dt, "%Y%m%d%H%M%S").isoformat()
+    try:
+        dt_iso = datetime.strptime(dt, "%Y%m%d%H%M%S").isoformat()
+    except:
+        return None
+
+    # 'i' intentionally not set
+    heritrix = dict(u=url, d=dt_iso, f=warc_file, o=int(offset), c=1)
+    return {'key': key, 'file:mime': mime, 'file:cdx': info, 'f:c': heritrix}
+
+def test_transform_line():
+
+    raw = "edu,upenn,ldc)/sites/www.ldc.upenn.edu/files/medar2009-large-arabic-broadcast-collection.pdf 20170828233154 https://www.ldc.upenn.edu/sites/www.ldc.upenn.edu/files/medar2009-large-arabic-broadcast-collection.pdf application/pdf 200 WL3FEA62TEU4F52Y5DOVQ62VET4QJW7G - - 210251 931661233 SEMSCHOLAR-PDF-CRAWL-2017-08-04-20170828231135742-00000-00009-wbgrp-svc284/SEMSCHOLAR-PDF-CRAWL-2017-08-04-20170828232253025-00005-3480~wbgrp-svc284.us.archive.org~8443.warc.gz"
+    correct = {
+        'key': "sha1:WL3FEA62TEU4F52Y5DOVQ62VET4QJW7G",
+        'file:mime': "application/pdf",
+        'file:cdx': {
+            'surt': "edu,upenn,ldc)/sites/www.ldc.upenn.edu/files/medar2009-large-arabic-broadcast-collection.pdf",
+            'url': "https://www.ldc.upenn.edu/sites/www.ldc.upenn.edu/files/medar2009-large-arabic-broadcast-collection.pdf",
+            'dt': "20170828233154",
+            'warc': "SEMSCHOLAR-PDF-CRAWL-2017-08-04-20170828231135742-00000-00009-wbgrp-svc284/SEMSCHOLAR-PDF-CRAWL-2017-08-04-20170828232253025-00005-3480~wbgrp-svc284.us.archive.org~8443.warc.gz",
+            'offset': 931661233,
+            'c_size': 210251,
+        },
+        'f:c': {
+            'u': "https://www.ldc.upenn.edu/sites/www.ldc.upenn.edu/files/medar2009-large-arabic-broadcast-collection.pdf",
+            'd': "2017-08-28T23:31:54",
+            'f': "SEMSCHOLAR-PDF-CRAWL-2017-08-04-20170828232253025-00005-3480~wbgrp-svc284.us.archive.org~8443.warc.gz",
+            'o': 931661233,
+            'c': 1,
+        }
+    }
+
+    assert transform_line(raw) == correct
+    assert transform_line(raw + "\n") == correct
+    assert transform_line(raw + " extra_field") == correct
+
+
+class MRCDXBackfillHBase(MRJob):
+
+    # CDX lines in; JSON status out
+    INPUT_PROTOCOL = mrjob.protocol.RawValueProtocol
+    OUTPUT_PROTOCOL = mrjob.protocol.JSONValueProtocol
+
+    def configure_args(self):
+        super(MRCDXBackfillHBase, self).configure_args()
+
+        self.add_passthru_arg('--hbase-table',
+                              type=str,
+                              default='wbgrp-journal-extract-0-qa',
+                              help='HBase table to backfill into (must exist)')
+        self.add_passthru_arg('--hbase-host',
+                              type=str,
+                              default='localhost',
+                              help='HBase thrift API host to connect to')
+
+    def __init__(self, *args, **kwargs):
+
+        # Allow passthrough for tests
+        if 'hb_table' in kwargs:
+            self.hb_table = kwargs.pop('hb_table')
+        else:
+            self.hb_table = None
+
+        super(MRCDXBackfillHBase, self).__init__(*args, **kwargs)
+        self.mime_filter = ['application/pdf']
+
+    def mapper_init(self):
+
+        if self.hb_table is None:
+            try:
+                host = self.options.hbase_host
+                # TODO: make these configs accessible from... mrconf.cfg?
+                hb_conn = happybase.Connection(host=host, transport="framed",
+                    protocol="compact")
+            except Exception as err:
+                raise Exception("Couldn't connect to HBase using host: {}".format(host))
+            self.hb_table = hb_conn.table(self.options.hbase_table)
+
+    def mapper(self, _, raw_cdx):
+
+        self.increment_counter('lines', 'total')
+
+        if (raw_cdx.startswith(' ') or raw_cdx.startswith('filedesc') or
+                raw_cdx.startswith('#')):
+
+            # Skip line
+            # XXX: tests don't cover this path; need coverage!
+            self.increment_counter('lines', 'invalid')
+            return _, dict(status="invalid")
+
+        info = transform_line(raw_cdx)
+        if info is None:
+            self.increment_counter('lines', 'invalid')
+            return _, dict(status="invalid")
+
+        if info['file:mime'] not in self.mime_filter:
+            self.increment_counter('lines', 'skip')
+            return _, dict(status="skip")
+
+        key = info.pop('key')
+        info['f:c'] = json.dumps(info['f:c'], sort_keys=True, indent=None)
+        info['file:cdx'] = json.dumps(info['file:cdx'], sort_keys=True,
+            indent=None)
+
+        self.hb_table.put(key, info)
+        self.increment_counter('lines', 'success')
+
+        yield _, dict(status="success")
+
+if __name__ == '__main__':
+    MRCDXBackfillHBase.run()
+
diff --git a/extraction/mrjob.conf b/extraction/mrjob.conf
new file mode 100644
index 0000000..cb286f1
--- /dev/null
+++ b/extraction/mrjob.conf
@@ -0,0 +1,4 @@
+runners:
+  hadoop:
+    setup:
+      - export PYTHONPATH=$PYTHONPATH:venv/lib/python3.5/site-packages/
diff --git a/extraction/pytest.ini b/extraction/pytest.ini
new file mode 100644
index 0000000..0a5e921
--- /dev/null
+++ b/extraction/pytest.ini
@@ -0,0 +1,8 @@
+
+[pytest]
+
+# allow imports from files in current directory
+python_paths = .
+
+# search for 'test_*' functions in all python files, not just under tests
+python_files = *.py
diff --git a/extraction/tests/files/example.cdx b/extraction/tests/files/example.cdx
new file mode 100644
index 0000000..84e3271
--- /dev/null
+++ b/extraction/tests/files/example.cdx
@@ -0,0 +1,20 @@
+edu,cmu,cs,adm,reports-archive)/anon/usr0/ftp/usr0/anon/2002/cmu-cs-02-119.pdf 20170706005950 http://reports-archive.adm.cs.cmu.edu/anon/usr0/ftp/usr0/anon/2002/CMU-CS-02-119.pdf application/pdf 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
+fi,tkk,lib)/diss/2001/isbn951225459x/isbn951225459x.pdf 20170705074926 http://lib.tkk.fi/Diss/2001/isbn951225459X/isbn951225459X.pdf application/pdf 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
+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
+org,ifaamas)/proceedings/aamas09/pdf/01_full%20papers/02_08_fp_0272.pdf 20170706081902 http://www.ifaamas.org/Proceedings/aamas09/pdf/01_Full%20Papers/02_08_FP_0272.pdf application/pdf 200 GYHX35QJWRJELWJ5GDQZPTPOUUZOCTKF - - 251180 34635154 CITESEERX-CRAWL-2017-06-20-20170706081825105-00419-00428-wbgrp-svc285/CITESEERX-CRAWL-2017-06-20-20170706081838210-00420-3671~wbgrp-svc285.us.archive.org~8443.warc.gz
+de,fau,cs)/publications/2014/lukas_14_masterthesis.pdf 20170705101722 http://www4.cs.fau.de/Publications/2014/lukas_14_masterthesis.pdf application/pdf 200 GIUQT7SXZ33TWEFBM2MWURJI2M3QE3IW - - 1290532 71068435 CITESEERX-CRAWL-2017-06-20-20170705101605019-00279-00288-wbgrp-svc284/CITESEERX-CRAWL-2017-06-20-20170705101714659-00281-31209~wbgrp-svc284.us.archive.org~8443.warc.gz
+de,bund,jki,pub)/index.php/jabfq/article/download/3568/4462 20170706041152 http://pub.jki.bund.de/index.php/JABFQ/article/download/3568/4462/ text/html 301 XZBNO24W2ZPQQMJYE6YUUCSRUF7G3ZBT - - 552 417292708 CITESEERX-CRAWL-2017-06-20-20170706040506112-00160-00169-wbgrp-svc285/CITESEERX-CRAWL-2017-06-20-20170706041021844-00165-3671~wbgrp-svc285.us.archive.org~8443.warc.gz
+whois://whois.arin.net/z+%2B+132.177.133.114 20170713120653 whois://whois.arin.net/z+%2B+132.177.133.114 text/plain - IDEID4YQ6MVJSOE57NPVDLL53ZB3J4DX - - 876 30983517 CITESEERX-CRAWL-2017-06-20-20170707064626094-01007-01015-wbgrp-svc285/CITESEERX-CRAWL-2017-06-20-20170711214025652-01014-3671~wbgrp-svc285.us.archive.org~8443.warc.gz
+za,co,csir,researchspace)/dspace/bitstream/10204/4048/1/smith2_2010.pdf 20170706094159 http://researchspace.csir.co.za/dspace/bitstream/10204/4048/1/Smith2_2010.pdf unk 301 3I42H3S6NNFQ2MSVX7XZKYAYSCX5QBYJ - - 446 104830407 CITESEERX-CRAWL-2017-06-20-20170706093829986-00509-00518-wbgrp-svc285/CITESEERX-CRAWL-2017-06-20-20170706094137978-00512-3671~wbgrp-svc285.us.archive.org~8443.warc.gz
+org,annals)/article.aspx?articleid=705034 20170707013120 http://annals.org/article.aspx?articleid=705034 text/html 301 QQYKL57QSERLFM3LXSWMNOFXMOCN7C5G - - 22665 28113974 CITESEERX-CRAWL-2017-06-20-20170707013100780-00967-00976-wbgrp-svc285/CITESEERX-CRAWL-2017-06-20-20170707013100780-00967-3671~wbgrp-svc285.us.archive.org~8443.warc.gz
+org,annals)/pdfaccess.ashx?url=/data/journals/aim/20105/0000605-200512200-00013.pdf 20170707045304 http://annals.org/pdfaccess.ashx?url=/data/journals/aim/20105/0000605-200512200-00013.pdf text/html 302 423S7EMGLCVIZ3FLVD7TLAG75HWE4RGI - - 644 222908628 CITESEERX-CRAWL-2017-06-20-20170707042504366-00997-01006-wbgrp-svc285/CITESEERX-CRAWL-2017-06-20-20170707045044604-00999-3671~wbgrp-svc285.us.archive.org~8443.warc.gz
+com,sagepub,spi)/content/28/4/501.full.pdf 20170705092027 http://spi.sagepub.com/content/28/4/501.full.pdf unk 301 3I42H3S6NNFQ2MSVX7XZKYAYSCX5QBYJ - - 396 553180242 CITESEERX-CRAWL-2017-06-20-20170705091311851-00219-00228-wbgrp-svc284/CITESEERX-CRAWL-2017-06-20-20170705091759818-00223-31209~wbgrp-svc284.us.archive.org~8443.warc.gz
+ir,mediaj)/favicon.ico 20170705075240 http://mediaj.ir/favicon.ico text/html 404 E3WSNQ7JAFOW7N3ZJ6GLV27T52T25JDK - - 589 455827180 CITESEERX-CRAWL-2017-06-20-20170705074433815-00129-00138-wbgrp-svc284/CITESEERX-CRAWL-2017-06-20-20170705075051100-00135-31209~wbgrp-svc284.us.archive.org~8443.warc.gz
+com,sagepub,jpr)/content/8/3-4/239.full.pdf 20170705074931 http://jpr.sagepub.com/content/8/3-4/239.full.pdf unk 301 3I42H3S6NNFQ2MSVX7XZKYAYSCX5QBYJ - - 400 270368088 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
+jp,co,nittuden)/business/pdf/transparent_thermoplastic_resin_with_electron_beam_cross-linking.pdf 20170706083459 http://www.nittuden.co.jp/business/pdf/Transparent_Thermoplastic_Resin_with_Electron_Beam_Cross-Linking.pdf application/pdf 200 V32E3CCO7NMI2M4OHLKG73DXD72LR4B2 - - 715081 761088410 CITESEERX-CRAWL-2017-06-20-20170706082646066-00429-00438-wbgrp-svc285/CITESEERX-CRAWL-2017-06-20-20170706083257353-00436-3671~wbgrp-svc285.us.archive.org~8443.warc.gz
+lt,lms)/robots.txt 20170705122708 http://www.lms.lt/robots.txt text/plain 200 PF3HTQQT2ULYRWFLJGUWZKHTVZUVMZ2F - - 592 668333707 CITESEERX-CRAWL-2017-06-20-20170705121748408-00399-00408-wbgrp-svc284/CITESEERX-CRAWL-2017-06-20-20170705122352502-00406-31209~wbgrp-svc284.us.archive.org~8443.warc.gz
+hu,bme,phy)/~szalay/pub/multipartcriteriaposter.pdf 20170705124828 http://www.phy.bme.hu/%7Eszalay/pub/multipartcriteriaPoster.pdf application/pdf 200 L3TUEEZLBJTHAVH74B5N426FAIDBCCOE - - 187866 964760782 CITESEERX-CRAWL-2017-06-20-20170705123641979-00419-00428-wbgrp-svc284/CITESEERX-CRAWL-2017-06-20-20170705124315591-00426-31209~wbgrp-svc284.us.archive.org~8443.warc.gz
+org,adb,openaccess)/bitstream/handle/11540/1260/new-regime-sme-finance-asia.pdf;jsessionid=f966a3bdac9882ec5a7c326b130f6f81?sequence=1 20170705090940 https://openaccess.adb.org/bitstream/handle/11540/1260/new-regime-sme-finance-asia.pdf%3Bjsessionid%3DF966A3BDAC9882EC5A7C326B130F6F81?sequence%3D1 unk 301 3I42H3S6NNFQ2MSVX7XZKYAYSCX5QBYJ - - 515 634039376 CITESEERX-CRAWL-2017-06-20-20170705090333400-00209-00218-wbgrp-svc284/CITESEERX-CRAWL-2017-06-20-20170705090728803-00212-31209~wbgrp-svc284.us.archive.org~8443.warc.gz
+org,physiology,ajpregu)/content/272/4/r1084 20170706131006 http://ajpregu.physiology.org/content/272/4/R1084 text/html 200 3FOQSKT4WBYOUA6VKKJCEQCN6QF35ANT - - 27346 336293585 CITESEERX-CRAWL-2017-06-20-20170706130432396-00707-00716-wbgrp-svc285/CITESEERX-CRAWL-2017-06-20-20170706130850866-00711-3671~wbgrp-svc285.us.archive.org~8443.warc.gz
+de,desy,www-it)/common/documentation/cd-docs/sc2002/paperpdf/pap234.pdf 20170705121813 http://www-it.desy.de/common/documentation/cd-docs/SC2002/paperpdf/pap234.pdf application/pdf 200 BONCZ4NNGRNYR22ASFVU7VYTQ24RRNP4 - - 72421 381715704 CITESEERX-CRAWL-2017-06-20-20170705120827801-00389-00398-wbgrp-svc284/CITESEERX-CRAWL-2017-06-20-20170705121708700-00397-31209~wbgrp-svc284.us.archive.org~8443.warc.gz
+org,oxfordjournals,bmb)/content/28/3/247.full.pdf 20170706014948 http://bmb.oxfordjournals.org/content/28/3/247.full.pdf text/html 301 EJWYVOPONJRARK7SGG6COFRN7CSTHROY - - 643 119398161 CITESEERX-CRAWL-2017-06-20-20170706014800946-00020-00029-wbgrp-svc285/CITESEERX-CRAWL-2017-06-20-20170706014907678-00022-3671~wbgrp-svc285.us.archive.org~8443.warc.gz
diff --git a/extraction/tests/test_backfill_hbase_from_cdx.py b/extraction/tests/test_backfill_hbase_from_cdx.py
new file mode 100644
index 0000000..1a13e5b
--- /dev/null
+++ b/extraction/tests/test_backfill_hbase_from_cdx.py
@@ -0,0 +1,54 @@
+"""
+TODO: could probably refactor to use unittest.mock.patch('happybase')
+"""
+
+import io
+import json
+import pytest
+import mrjob
+import happybase_mock
+from backfill_hbase_from_cdx import MRCDXBackfillHBase
+
+@pytest.fixture
+def job():
+    """
+    Note: this mock only seems to work with job.run_mapper(), not job.run();
+    the later results in a separate instantiation without the mock?
+    """
+    conn = happybase_mock.Connection()
+    conn.create_table('wbgrp-journal-extract-test',
+        {'file': {}, 'grobid0': {}, 'f': {}})
+    table = conn.table('wbgrp-journal-extract-test')
+
+    job = MRCDXBackfillHBase(['--no-conf', '-'], hb_table=table)
+    return job
+
+
+def test_some_lines(job):
+
+    raw = io.BytesIO(b"""
+com,sagepub,cep)/content/28/9/960.full.pdf 20170705062200 http://cep.sagepub.com/content/28/9/960.full.pdf application/pdf 301 3I42H3S6NNFQ2MSVX7XZKYAYSCX5QBYJ - - 401 313356621 CITESEERX-CRAWL-2017-06-20-20170705061647307-00039-00048-wbgrp-svc284/CITESEERX-CRAWL-2017-06-20-20170705062052659-00043-31209~wbgrp-svc284.us.archive.org~8443.warc.gz
+eu,eui,cadmus)/bitstream/handle/1814/36635/rscas_2015_03.pdf;jsessionid=761393014319a39f40d32ae3eb3a853f?sequence=1 20170705062202 http://cadmus.eui.eu/bitstream/handle/1814/36635/RSCAS_2015_03.pdf%3Bjsessionid%3D761393014319A39F40D32AE3EB3A853F?sequence%3D1 application/PDF 200 MPCXVWMUTRUGFP36SLPHKDLY6NGU4S3J - - 854156 328850624 CITESEERX-CRAWL-2017-06-20-20170705061647307-00039-00048-wbgrp-svc284/CITESEERX-CRAWL-2017-06-20-20170705062052659-00043-31209~wbgrp-svc284.us.archive.org~8443.warc.gz
+com,pbworks,educ333b)/robots.txt 20170705063311 http://educ333b.pbworks.com/robots.txt text/plain 200 6VAUYENMOU2SK2OWNRPDD6WTQTECGZAD - - 638 398190140 CITESEERX-CRAWL-2017-06-20-20170705062707827-00049-00058-wbgrp-svc284/CITESEERX-CRAWL-2017-06-20-20170705063158203-00053-31209~wbgrp-svc284.us.archive.org~8443.warc.gz
+""")
+
+    job.sandbox(stdin=raw)
+    job.run_mapper()
+
+    assert job.hb_table.row(b'1') == {}
+    # HTTP 301
+    assert job.hb_table.row(b'sha1:3I42H3S6NNFQ2MSVX7XZKYAYSCX5QBYJ') == {}
+    # valid
+    assert job.hb_table.row(b'sha1:MPCXVWMUTRUGFP36SLPHKDLY6NGU4S3J') != {}
+    # text/plain
+    assert job.hb_table.row(b'sha1:6VAUYENMOU2SK2OWNRPDD6WTQTECGZAD') == {}
+
+    row = job.hb_table.row(b'sha1:MPCXVWMUTRUGFP36SLPHKDLY6NGU4S3J')
+    assert row[b'file:mime'] == b"application/pdf"
+
+    file_cdx = json.loads(row[b'file:cdx'].decode('utf-8'))
+    assert int(file_cdx['offset']) == 328850624
+
+    f_c = json.loads(row[b'f:c'].decode('utf-8'))
+    assert f_c['u'] == "http://cadmus.eui.eu/bitstream/handle/1814/36635/RSCAS_2015_03.pdf%3Bjsessionid%3D761393014319A39F40D32AE3EB3A853F?sequence%3D1"
+    assert b'i' not in f_c