Write some missing tests. Closes wireservice#112 wireservice#113 wireservice#114 wireservice#115 wireservice#116.

Author: onyxfish

.gitignore

@@ -9,3 +9,4 @@ dist
 docs/_build
 .coverage
 .tox
+cover

csvkit/utilities/csvclean.py

@@ -11,7 +11,7 @@ class CSVClean(CSVKitUtility):
 
     def add_arguments(self):
         self.argparser.add_argument('-n', '--dry-run', dest='dryrun', action='store_true',
-                                    help='If this argument is present, no output will be created. Information about what would have been done will be printed to STDERR.''')
+            help='Do not create output files. Information about what would have been done will be printed to STDERR.')
 
     def main(self):
         reader = CSVKitReader(self.args.file, **self.reader_kwargs)
@@ -22,15 +22,15 @@ def main(self):
                 pass
             if checker.errs:
                 for e in checker.errs:
-                    self.output_file.write("Line %i: %s\n" % (e.line_number,e.msg))
+                    self.output_file.write('Line %i: %s\n' % (e.line_number,e.msg))
             else:
-                self.output_file.write("No errors.\n")
+                self.output_file.write('No errors.\n')
             if checker.joins:
-                self.output_file.write("%i rows would have been joined/reduced to %i rows after eliminating expected internal line breaks.\n" % (checker.rows_joined, checker.joins))
+                self.output_file.write('%i rows would have been joined/reduced to %i rows after eliminating expected internal line breaks.\n' % (checker.rows_joined, checker.joins))
         else:
             base,ext = splitext(self.args.file.name)
             # should we preserve delimiters and other dialect args from CLI?
-            cleaned_file = CSVKitWriter(open("%s_out.csv" % base,"w"), **self.writer_kwargs)
+            cleaned_file = CSVKitWriter(open('%s_out.csv' % base,'w'), **self.writer_kwargs)
 
             checker = RowChecker(reader)
             cleaned_file.writerow(checker.column_names)
@@ -39,23 +39,22 @@ def main(self):
 
             if checker.errs:
                 # should we preserve delimiters and other dialect args from CLI?
-                err_filename = "%s_err.csv" % base
-                err_file = CSVKitWriter(open(err_filename, "w"), **self.writer_kwargs)
+                err_filename = '%s_err.csv' % base
+                err_file = CSVKitWriter(open(err_filename, 'w'), **self.writer_kwargs)
                 err_header = ['line_number','msg']
                 err_header.extend(checker.column_names)
                 err_file.writerow(err_header)
                 for e in checker.errs:
                     err_file.writerow(self._format_error_row(e))
                     err_count = len(checker.errs)
-                self.output_file.write("%i error%s logged to %s\n" % (err_count,"" if err_count == 1 else "s", err_filename))
+                self.output_file.write('%i error%s logged to %s\n' % (err_count,'' if err_count == 1 else 's', err_filename))
             else:
-                self.output_file.write("No errors.\n")
+                self.output_file.write('No errors.\n')
 
             if checker.joins:
-                self.output_file.write("%i rows were joined/reduced to %i rows after eliminating expected internal line breaks.\n" % (checker.rows_joined, checker.joins))
+                self.output_file.write('%i rows were joined/reduced to %i rows after eliminating expected internal line breaks.\n' % (checker.rows_joined, checker.joins))
 
     def _format_error_row(self, e):
-        """Format a row for """
         err_row = [e.line_number, e.msg]
         err_row.extend(e.row)
         return err_row
@@ -64,6 +63,6 @@ def launch_new_instance():
     utility = CSVClean()
     utility.main()
 
-if __name__ == "__main__":
+if __name__ == '__main__':
     launch_new_instance()
 

examples/join_a.csv

@@ -0,0 +1,4 @@
+a,b,c
+1,b,c
+2,b,c
+3,b,c

examples/join_b.csv

@@ -0,0 +1,4 @@
+a,b,c
+1,b,c
+1,b,c
+4,b,c

tests/test_utilities/test_csvclean.py

@@ -1,9 +1,34 @@
 #!/usr/bin/env python
 
+from cStringIO import StringIO
+import os
 import unittest
 
 from csvkit.utilities.csvclean import CSVClean
 
 class TestCSVClean(unittest.TestCase):
-    pass
+    def test_simple(self):
+        args = ['examples/bad.csv']
+        output_file = StringIO()
+
+        utility = CSVClean(args, output_file)
+        utility.main()
+
+        self.assertTrue(os.path.exists('examples/bad_err.csv'))
+        self.assertTrue(os.path.exists('examples/bad_out.csv'))
+
+        with open('examples/bad_err.csv') as f:
+            f.next()
+            self.assertEqual(f.next()[0], '1')
+            self.assertEqual(f.next()[0], '2')
+            self.assertRaises(StopIteration, f.next)
+            
+        with open('examples/bad_out.csv') as f:
+            f.next()
+            self.assertEqual(f.next()[0], '0')
+            self.assertRaises(StopIteration, f.next)
+
+        # Cleanup
+        os.remove('examples/bad_err.csv')
+        os.remove('examples/bad_out.csv')
 

tests/test_utilities/test_csvjoin.py

@@ -1,9 +1,63 @@
 #!/usr/bin/env python
 
+from cStringIO import StringIO
 import unittest
 
-from csvkit.utilities.csvjoin import CSVJoin 
+from csvkit.utilities.csvjoin import CSVJoin
 
 class TestCSVJoin(unittest.TestCase):
-    pass
+    def test_sequential(self):
+        args = ['examples/join_a.csv', 'examples/join_b.csv']
+        output_file = StringIO()
+
+        utility = CSVJoin(args, output_file)
+        utility.main()
+
+        output = StringIO(output_file.getvalue())
+
+        self.assertEqual(len(output.readlines()), 4)
+
+    def test_inner(self):
+        args = ['-c', 'a', 'examples/join_a.csv', 'examples/join_b.csv']
+        output_file = StringIO()
+
+        utility = CSVJoin(args, output_file)
+        utility.main()
+
+        output = StringIO(output_file.getvalue())
+
+        self.assertEqual(len(output.readlines()), 3)
+
+    def test_left(self):
+        args = ['-c', 'a', '--left', 'examples/join_a.csv', 'examples/join_b.csv']
+        output_file = StringIO()
+
+        utility = CSVJoin(args, output_file)
+        utility.main()
+
+        output = StringIO(output_file.getvalue())
+
+        self.assertEqual(len(output.readlines()), 5)
+
+    def test_right(self):
+        args = ['-c', 'a', '--right', 'examples/join_a.csv', 'examples/join_b.csv']
+        output_file = StringIO()
+
+        utility = CSVJoin(args, output_file)
+        utility.main()
+
+        output = StringIO(output_file.getvalue())
+
+        self.assertEqual(len(output.readlines()), 4)
+
+    def test_outer(self):
+        args = ['-c', 'a', '--outer', 'examples/join_a.csv', 'examples/join_b.csv']
+        output_file = StringIO()
+
+        utility = CSVJoin(args, output_file)
+        utility.main()
+
+        output = StringIO(output_file.getvalue())
+
+        self.assertEqual(len(output.readlines()), 6)
 

tests/test_utilities/test_csvstat.py

@@ -1,9 +1,15 @@
 #!/usr/bin/env python
 
+from cStringIO import StringIO
 import unittest
 
 from csvkit.utilities.csvstat import CSVStat
 
 class TestCSVStat(unittest.TestCase):
-    pass
+    def test_runs(self):
+        args = ['examples/dummy.csv']
+        output_file = StringIO()
+
+        utility = CSVStat(args, output_file)
+        utility.main()