Lz4 support for pt.io.autoopen() (#323, addresses #317)

* added lz4, addresses #317
* added test cases

pyterrier/io.py

@@ -17,7 +17,7 @@ def coerce_dataframe(obj):
 
 def autoopen(filename, mode='rb'):
     """
-    A drop-in for open() that applies automatic compression for .gz and .bz2 file extensions
+    A drop-in for open() that applies automatic compression for .gz, .bz2 and .lz4 file extensions
     """
 
     if filename.endswith(".gz"):
@@ -26,6 +26,9 @@ def autoopen(filename, mode='rb'):
     elif filename.endswith(".bz2"):
         import bz2
         return bz2.open(filename, mode)
+    elif filename.endswith(".lz4"):
+        import lz4.frame
+        return lz4.frame.open(filename, mode)
     return open(filename, mode)
 
 def find_files(dir):

requirements-test.txt

@@ -3,3 +3,4 @@ xgboost
 ray
 fastrank>=0.7.0
 torch
+lz4

tests/test_io.py

@@ -8,6 +8,15 @@
 
 class TestUtils(TempDirTestCase):
 
+    def test_compression(self):
+        with pt.io.autoopen(os.path.join(self.here, 'fixtures/testfile.txt.lz4'), 'rt') as f:
+            lines = f.readlines()
+            self.assertEqual('hello world', lines[0])
+        with pt.io.autoopen(os.path.join(self.test_dir, "bla.txt.gz"), 'wt') as f:
+            f.write('bla')
+        with pt.io.autoopen(os.path.join(self.test_dir, "bla.txt.gz"), 'rt') as f:
+            lines = f.readlines()
+            self.assertEqual('bla', lines[0])
 
     def test_save_trec(self):
         res = pd.DataFrame([["1", "d1", 5.3, 1]], columns=['qid', 'docno', 'score', 'rank'])