From f7ecd1c9d2f621ae86d8a1ea5088f51d6a9148a2 Mon Sep 17 00:00:00 2001
From: John Chilton <jmchilton@gmail.com>
Date: Wed, 25 Mar 2015 09:30:25 -0400
Subject: [PATCH] Allow testing tool outputs based on MD5 hashes.

This should not be considered the best practice for most text based files, but for binary files we cannot produce diffs anyway so this is probably better than storing the pointless binary blobs in repositories for instance. In particular Dan and I thought this would be useful for testing index files for data managers.

Includes test tool demonstrating this functionality, run test with

    ./run_tests.sh -framework -id md5sum
---
 lib/galaxy/tools/parser/xml.py              |  6 ++++--
 lib/galaxy/tools/test.py                    |  1 +
 test/base/twilltestcase.py                  | 20 ++++++++++++++++++++
 test/functional/tools/md5sum.xml            | 17 +++++++++++++++++
 test/functional/tools/samples_tool_conf.xml |  1 +
 5 files changed, 43 insertions(+), 2 deletions(-)
 create mode 100644 test/functional/tools/md5sum.xml

diff --git a/lib/galaxy/tools/parser/xml.py b/lib/galaxy/tools/parser/xml.py
index 2afe09f80db..c7e8b3595d7 100644
--- a/lib/galaxy/tools/parser/xml.py
+++ b/lib/galaxy/tools/parser/xml.py
@@ -362,11 +362,13 @@ def __parse_test_attributes( output_elem, attrib ):
     metadata = {}
     for metadata_elem in output_elem.findall( 'metadata' ):
         metadata[ metadata_elem.get('name') ] = metadata_elem.get( 'value' )
-    if not (assert_list or file or extra_files or metadata):
-        raise Exception( "Test output defines nothing to check (e.g. must have a 'file' check against, assertions to check, etc...)")
+    md5sum = attrib.get("md5", None)
+    if not (assert_list or file or extra_files or metadata or md5sum):
+        raise Exception( "Test output defines nothing to check (e.g. must have a 'file' check against, assertions to check, metadata or md5 tests, etc...)")
     attributes['assert_list'] = assert_list
     attributes['extra_files'] = extra_files
     attributes['metadata'] = metadata
+    attributes['md5'] = md5sum
     return file, attributes
 
 
diff --git a/lib/galaxy/tools/test.py b/lib/galaxy/tools/test.py
index 03f48106e6a..babdcb84fe2 100644
--- a/lib/galaxy/tools/test.py
+++ b/lib/galaxy/tools/test.py
@@ -133,6 +133,7 @@ class ToolTestBuilder( object ):
             self.stderr = test_dict.get("stderr", None)
             self.expect_exit_code = test_dict.get("expect_exit_code", None)
             self.expect_failure = test_dict.get("expect_failure", False)
+            self.md5 = test_dict.get("md5", None)
         except Exception, e:
             self.error = True
             self.exception = e
diff --git a/test/base/twilltestcase.py b/test/base/twilltestcase.py
index df8bb5305f8..2575b03e515 100644
--- a/test/base/twilltestcase.py
+++ b/test/base/twilltestcase.py
@@ -2271,6 +2271,14 @@ class TwillTestCase( unittest.TestCase ):
                 errmsg = 'History item %s different than expected\n' % (hid)
                 errmsg += str( err )
                 raise AssertionError( errmsg )
+        if attributes is not None and attributes.get("md5", None) is not None:
+            md5 = attributes.get("md5")
+            try:
+                self._verify_md5(data, md5)
+            except AssertionError, err:
+                errmsg = 'History item %s different than expected\n' % (hid)
+                errmsg += str( err )
+                raise AssertionError( errmsg )
         if filename is not None:
             local_name = self.get_filename( filename, shed_tool_id=shed_tool_id )
             temp_name = self.makeTfname(fname=filename)
@@ -2323,6 +2331,18 @@ class TwillTestCase( unittest.TestCase ):
                 if 'GALAXY_TEST_NO_CLEANUP' not in os.environ:
                     os.remove( temp_name )
 
+    def _verify_md5( self, data, expected_md5 ):
+        import md5
+        m = md5.new()
+        m.update( data )
+        actual_md5 = m.hexdigest()
+        if expected_md5 != actual_md5:
+            template = "Output md5sum [%s] does not match expected [%s]."
+            message = template % (actual_md5, expected_md5)
+            assert False, message
+
+
+
     def view_external_service( self, external_service_id, strings_displayed=[] ):
         self.visit_url( '%s/external_service/view_external_service?id=%s' % ( self.url, external_service_id ) )
         for check_str in strings_displayed:
diff --git a/test/functional/tools/md5sum.xml b/test/functional/tools/md5sum.xml
new file mode 100644
index 00000000000..737057142b5
--- /dev/null
+++ b/test/functional/tools/md5sum.xml
@@ -0,0 +1,17 @@
+<tool id="md5sum" name="md5sum" version="1.0.0">
+    <command>
+        cp $input $output
+    </command>
+    <inputs>
+        <param name="input" type="data" format="txt" />
+    </inputs>
+    <outputs>
+        <data name="output" format="txt" />
+    </outputs>
+    <tests>
+        <test>
+            <param name="input" value="simple_line.txt" />
+            <output name="out_file1" md5="ac94233685713ce99d1e712b0023c381" />
+        </test>
+    </tests>
+</tool>
diff --git a/test/functional/tools/samples_tool_conf.xml b/test/functional/tools/samples_tool_conf.xml
index 40780b3d9aa..22fbd8290b7 100644
--- a/test/functional/tools/samples_tool_conf.xml
+++ b/test/functional/tools/samples_tool_conf.xml
@@ -17,6 +17,7 @@
   <tool file="metadata.xml" />
   <tool file="metadata_bam.xml" />
   <tool file="detect_errors_aggressive.xml" />
+  <tool file="md5sum.xml" />
   <tool file="job_properties.xml" />
   <tool file="gzipped_inputs.xml" />
   <tool file="output_order.xml" />