grp.py: added tests, improved docstrings

cmapPy/set_io/gmt.py

@@ -22,8 +22,15 @@
 
 
 def read(file_path):
-    """
-    Read a gmt file at the path specified by file_path.    
+    """ Read a gmt file at the path specified by file_path.
+
+    Args:
+        file_path (string): path to gmt file
+
+    Returns:
+        gmt (GMT object): list of dicts, where each dict corresponds to one
+            line of the GMT file
+
     """
     # Read in file
     actual_file_path = os.path.expanduser(file_path)
@@ -67,6 +74,15 @@ def read(file_path):
 
 
 def verify_gmt_integrity(gmt):
+    """ Make sure that set ids are unique.
+
+    Args:
+        gmt (GMT object): list of dicts
+
+    Returns:
+        None
+
+    """
 
     # Verify that set ids are unique
     set_ids = [d[SET_IDENTIFIER_FIELD] for d in gmt]
@@ -75,10 +91,16 @@ def verify_gmt_integrity(gmt):
 
 
 def write(gmt, out_path):
-    """
-    Write a GMT to a text file.
-    """
+    """ Write a GMT to a text file.
 
+    Args:
+        gmt (GMT object): list of dicts
+        out_path (string): output path
+
+    Returns:
+        None
+
+    """
     with open(out_path, 'w') as f:
         for _, each_dict in enumerate(gmt):
             f.write(each_dict[SET_IDENTIFIER_FIELD] + '\t')

cmapPy/set_io/grp.py

@@ -1,59 +1,49 @@
 """
-module with class definition and methods for reading and writing .grp files
-Created on Jun 20, 2012
-@author: David Wadden
+grp.py
+
+IO methods for handling GRP files.
+
+A GRP file is stored as a list. Lines beginning with # are ignored.
+
+AUTHOR: David Wadden, Broad Institute, 2012
+MODIFIED: Lev Litichevskiy, 2017
 """
 
 import os
 import re
 
 
-class GRP:
-    """
-    class to read .grp files and return a list
-    """
-    def __init__(self, src):
-        # if it"s a file string, check that it exists and read the file
-        if type(src) is str:
-            assert os.path.exists(src), "{0} is not a valid file path. Use a list to input plate names directly".format(src)
-            self.read(src)
-        # if it's a list, just read it in
-        elif type(src) is list:
-            self.grp = src
-
-    def read(self, in_path):
-        """
-        read a .grp file
-        """
-        with open(in_path, "r") as f:
-            lines = f.readlines()
-            # need the second conditional to ignore comment lines
-            self.grp = [line.strip() for line in lines if line and not re.match("^#", line)]
-
-    def write(self, out):
-        """
-        write a .grp file
-        """
-        with open(out, "w") as f:
-            for x in self.grp:
-                f.write(str(x) + "\n")
-
-
-def write_grp(in_list, out):
-    """
-    standalone methods to write .grp files
-    """
-    with open(out, "w") as f:
-        for x in in_list:
-            f.write(str(x) + "\n")
+def read(in_path):
+    """ Read a grp file at the path specified by in_path.
 
+    Args:
+        in_path (string): path to GRP file
+
+    Returns:
+        grp (list)
 
-def read_grp(in_path):
-    """
-    standalone method to read .grp files
     """
-    assert os.path.exists(in_path), "The following file can't be found. in_path: {}".format(in_path)
+    assert os.path.exists(in_path), "The following GRP file can't be found. in_path: {}".format(in_path)
+
     with open(in_path, "r") as f:
         lines = f.readlines()
-        # again, second conditional ignores comment lines
-        return [line.strip() for line in lines if line and not re.match("^#", line)]
+        # need the second conditional to ignore comment lines
+        grp = [line.strip() for line in lines if line and not re.match("^#", line)]
+
+    return grp
+
+
+def write(grp, out_path):
+    """ Write a GRP to a text file.
+
+    Args:
+        grp (list): GRP object to write to new-line delimited text file
+        out_path (string): output path
+
+    Returns:
+        None
+
+    """
+    with open(out_path, "w") as f:
+        for x in grp:
+            f.write(str(x) + "\n")

cmapPy/set_io/tests/functional_tests/test.grp

@@ -0,0 +1,5 @@
+#a
+r
+d
+e
+#f

cmapPy/set_io/tests/test_gmt.py

@@ -2,13 +2,13 @@
 import os
 import unittest
 from cmapPy.pandasGEXpress import setup_GCToo_logger as setup_logger
-import cmapPy.pandasGEXpress.gmt as gmt
+import cmapPy.set_io.gmt as gmt
 
 logger = logging.getLogger(setup_logger.LOGGER_NAME)
 FUNCTIONAL_TESTS_DIR = "functional_tests"
 
 
-class GMT(unittest.TestCase):
+class TestGMT(unittest.TestCase):
 
 	@classmethod
 	def setUpClass(cls):

cmapPy/set_io/tests/test_grp.py

@@ -0,0 +1,39 @@
+import logging
+import os
+import unittest
+from cmapPy.pandasGEXpress import setup_GCToo_logger as setup_logger
+import cmapPy.set_io.grp as grp
+
+logger = logging.getLogger(setup_logger.LOGGER_NAME)
+FUNCTIONAL_TESTS_DIR = "functional_tests"
+
+
+class TestGRP(unittest.TestCase):
+
+	def test_read(self):
+
+		in_grp = grp.read(os.path.join(FUNCTIONAL_TESTS_DIR, "test.grp"))
+		self.assertEqual(in_grp, ["r", "d", "e"])
+
+		with self.assertRaises(AssertionError) as e:
+			grp.read("testt.grp")
+		self.assertIn("The following GRP file", str(e.exception))
+
+	def test_write(self):
+
+		example_grp = ["x", "z", "w"]
+
+		out_path = os.path.join(FUNCTIONAL_TESTS_DIR, "test_write.grp")
+		grp.write(example_grp, out_path)
+		self.assertTrue(os.path.exists(out_path))
+
+		read_back_in = grp.read(out_path)
+		self.assertEqual(example_grp, read_back_in)
+
+		# Cleanup
+		os.remove(out_path)
+
+if __name__ == "__main__":
+	setup_logger.setup(verbose=True)
+
+	unittest.main()