Merge pull request #91 from pitthsls/turnaways

Turnaways support

CHANGELOG.md

@@ -1,4 +1,7 @@
 # Changelog
+## 2.1.0 (unreleased)
+* add support for COUNTER 4 title-level turnaways reports (JR2 and BR3).
+
 ## 2.0.0 (2019-09-13)
 
 * COUNTER 5 support (still a work in progress; see README).

README.rst

@@ -79,7 +79,7 @@ Usage
 -----
 
 Parsing COUNTER reports (currently supports COUNTER 3 and 4, in .csv, .tsv, 
-or .xlsx files, reports JR1, DB1, DB2, BR1, and BR2)::
+or .xlsx files, reports JR1, JR2, DB1, DB2, PR1, BR1, BR2 and BR3)::
 
     >>> import pycounter.report
     >>> report = pycounter.report.parse("COUNTER4_2015.tsv")  # filename or path to file

pycounter/report.py

@@ -772,7 +772,7 @@ def parse_generic(report_reader):
         # these reports do not have line with totals
         six.next(report_reader)
 
-    if report.report_type == "DB2":
+    if report.report_type in ("DB2", "BR3", "JR3"):
         # this report has two lines of totals
         six.next(report_reader)
 
@@ -792,6 +792,7 @@ def _parse_line(line, report, last_col):
     :param last_col: last column number containing data
     :return: an appropriate CounterResource subclass instance
     """
+    # pylint: disable=too-many-locals,too-many-branches,too-many-statements
     issn = None
     eissn = None
     isbn = None
@@ -800,6 +801,7 @@ def _parse_line(line, report, last_col):
     doi = ""
     prop_id = ""
 
+    metric = report.metric
     if report.report_version >= 4:
         if report.report_type.startswith("JR1") or report.report_type == "TR_J1":
             old_line = line
@@ -816,6 +818,16 @@ def _parse_line(line, report, last_col):
             isbn = line[3].strip()
             issn = line[4].strip()
 
+        elif report.report_type in ("BR3", "JR2"):
+            metric = line[7]
+            doi = line[3]
+            prop_id = line[4]
+            line = line[0:3] + line[5:7] + line[9:last_col]
+            eissn = line[4].strip()
+            if report.report_type == "BR3":
+                isbn = line[3].strip()
+            else:
+                issn = line[3].strip()
         # For DB1 and DB2, nothing additional to do here
 
     else:
@@ -841,7 +853,7 @@ def _parse_line(line, report, last_col):
         curr_month = next_month(curr_month)
     if report.report_type.startswith("JR") or report.report_type == "TR_J1":
         return CounterJournal(
-            metric=report.metric,
+            metric=metric,
             month_data=month_data,
             doi=doi,
             issn=issn,
@@ -853,7 +865,7 @@ def _parse_line(line, report, last_col):
         )
     elif report.report_type.startswith("BR"):
         return CounterBook(
-            metric=report.metric,
+            metric=metric,
             month_data=month_data,
             doi=doi,
             issn=issn,
@@ -923,6 +935,8 @@ def _year_from_header(header, report):
         first_date_col = 5
     elif report.report_type == "PR1" and report.report_version == 4:
         first_date_col = 4
+    elif report.report_type == "JR2":
+        first_date_col = 11
     year = int(header[first_date_col].split("-")[1])
     if year < 100:
         year += 2000

pycounter/test/conftest.py

@@ -143,3 +143,15 @@ def report_file_output(request):
 def db_report(request):
     """All C4 database reports."""
     return parsedata(request.param)
+
+
+@pytest.fixture
+def br3_report():
+    """Book report 3 (turnaways)."""
+    return parsedata("C4BR3.csv")
+
+
+@pytest.fixture
+def jr2_report():
+    """Journal report 2 (turnaways)."""
+    return parsedata("C4JR2.csv")

pycounter/test/data/C4BR3.csv

@@ -0,0 +1,14 @@
+Book Report 3 (R4),"Access Denied to Content Items by Month, Title and Category",,,,,,,,,,,,,
+University of Maximegalon,,,,,,,,,,,,,,
+BlahBlah,,,,,,,,,,,,,,
+Period covered by Report:,,,,,,,,,,,,,,
+2012-01-01 to 2012-06-30,,,,,,,,,,,,,,
+Date run:,,,,,,,,,,,,,,
+2012-07-09,,,,,,,,,,,,,,
+,Publisher,Platform,Book DOI,Propriatary Identifier,ISBN,ISSN,Access Denied Category,Reporting Period Total,Jan-2012,Feb-2012,Mar-2012,Apr-2012,May-2012,Jun-2012
+Total for all titles,Megadodo Publications,HHGTTG Online,,,,,Access denied: concurrent/simultaneous user license limit exceeded,14,4,0,2,2,3,3
+Total for all titles,Megadodo Publications,HHGTTG Online,,,,,Access denied: content item not licensed,30,10,3,9,1,4,3
+Hitchhikers Guide to the Galaxy,Megadodo Publications,HHGTTG Online,,,9787490833809,1900-1000,Access denied: concurrent/simultaneous user license limit exceeded,14,4,0,2,2,3,3
+Hitchhikers Guide to the Galaxy,Megadodo Publications,HHGTTG Online,,,9787490833809,1900-1000,Access denied: content item not licensed,0,0,0,0,0,0,0
+Öfake Book,Megadodo Publications,HHGTTG Online,,,9783572324031,,Access denied: concurrent/simultaneous user license limit exceeded,0,0,0,0,0,0,0
+Öfake Book,Megadodo Publications,HHGTTG Online,,,9783572324031,,Access denied: content item not licensed,30,10,3,9,1,4,3

pycounter/test/data/C4JR2.csv

@@ -0,0 +1,11 @@
+"Journal Report 2 (R4)","Access denied to Full-text Articles by Month, Journal and Category"
+"University of Maximegalon"
+""
+"Period covered by Report:"
+"2011-01-01 to 2011-12-31"
+"Date run:"
+2012-02-21
+Journal,Publisher,Platform,"Journal DOI","Proprietary Identifier","Print ISSN","Online ISSN","Access Denied Category","Reporting Period Total","Reporting Period HTML","Reporting Period PDF",Jan-2011,Feb-2011,Mar-2011,Apr-2011,May-2011,Jun-2011,Jul-2011,Aug-2011,Sep-2011,Oct-2011,Nov-2011,Dec-2011
+"Total for all journals","Maximegalon University Press",MJO,,,,,16,0,16,2,1,0,0,0,5,1,1,0,5,1,0
+"Abstracts of Working Papers in Economics","Maximegalon University Press",MJO,,,0951-0079,0951-0079,Access denied: concurrent/simultaneous user license limit exceeded,3,3,0,0,0,0,0,0,0,0,0,0,0,0,0
+"Acta Numerica","Maximegalon University Press",MJO,,,0962-4929,1474-0508,Access denied: concurrent/simultaneous user license limit exceeded,16,0,16,2,1,0,0,0,5,1,1,0,5,1,0

pycounter/test/test_br3.py

@@ -0,0 +1,22 @@
+"""Test parsing of COUNTER 4 BR3 report (turnaways)"""
+
+import datetime
+
+
+def test_metric(br3_report):
+    assert br3_report.metric is None  # Multiple metrics per report
+
+
+def test_type(br3_report):
+    assert br3_report.report_type == u"BR3"
+
+
+def test_data(br3_report):
+    i = iter(br3_report)
+    row = next(i)
+    item = next(iter(row))
+    assert item == (
+        datetime.date(2012, 1, 1),
+        u"Access denied: concurrent/simultaneous user license limit exceeded",
+        4,
+    )

pycounter/test/test_jr2.py

@@ -0,0 +1,22 @@
+"""Test parsing of COUNTER 4 JR2 report (turnaways)"""
+
+import datetime
+
+
+def test_metric(jr2_report):
+    assert jr2_report.metric is None  # Multiple metrics per report
+
+
+def test_type(jr2_report):
+    assert jr2_report.report_type == u"JR2"
+
+
+def test_data(jr2_report):
+    i = iter(jr2_report)
+    row = next(i)
+    item = next(iter(row))
+    assert item == (
+        datetime.date(2011, 1, 1),
+        u"Access denied: concurrent/simultaneous user license limit exceeded",
+        3,
+    )