tests compare against git hash, eliminate cached .h5

This updates the test infrastructure to store a git hash of the reference code, instead of generated .h5 files. The test infrastructure creates a cache of the reference .h5 files from the specified git hash if it doesn't already exist, and then compares against that cache.

This gets rid of the need for git-lfs and the cached .h5 files, requiring instead the code is checked out from git.

gwinc/test/__main__.py

 import os
 import sys
 import glob
+import shutil
 import signal
 import logging
 import tempfile
@@ -28,13 +29,64 @@ logging.basicConfig(
 
 FREQ = np.logspace(np.log10(5), np.log10(6000), 3000)
 TOLERANCE = 1e-6
+CACHE_LIMIT = 5
 
 
 def test_path(*args):
+    """Return path to package file."""
     return os.path.join(os.path.dirname(__file__), *args)
 
 
-def gen_cache(ref_hash, path):
+def git_ref_resolve_hash(git_ref):
+    """Resolve a git reference into its hash string."""
+    try:
+        return subprocess.run(
+            ['git', 'show', '-s', '--format=format:%H', git_ref],
+            capture_output=True, universal_newlines=True,
+        ).stdout
+    except subprocess.CalledProcessError:
+        return None
+
+
+def write_ref_hash(ref_hash):
+    """Write ref hash to reference file
+
+    """
+    with open(test_path('ref_hash'), 'w') as f:
+        f.write('{}\n'.format(ref_hash))
+
+
+def load_ref_hash():
+    """Load the current reference git hash.
+
+    """
+    try:
+        with open(test_path('ref_hash')) as f:
+            return f.read().strip()
+    except IOError:
+        return None
+
+
+def prune_cache_dir():
+    """Prune all but the N most recently accessed caches.
+
+    """
+    cache_dir = test_path('cache')
+    if not os.path.exists(cache_dir):
+        return
+    expired_paths = sorted(
+        [os.path.join(cache_dir, path) for path in os.listdir(cache_dir)],
+        key=lambda path: os.stat(path).st_atime,
+    )[CACHE_LIMIT:]
+    if not expired_paths:
+        return
+    logging.info("pruning {} old caches...".format(len(expired_paths)))
+    for path in expired_paths:
+        logging.debug("pruning {}...".format(path))
+        shutil.rmtree(path)
+
+
+def gen_cache_for_ref(ref_hash, path):
     """generate cache from git reference
 
     The ref_hash should be a git hash, and path should be the location
@@ -47,7 +99,8 @@ def gen_cache(ref_hash, path):
     """
     logging.info("creating new cache from reference {}...".format(ref_hash))
     subprocess.run(
-        [test_path('gen_cache.sh'), ref_hash, path]
+        [test_path('gen_cache.sh'), ref_hash, path],
+        check=True,
     )
 
 
@@ -65,7 +118,7 @@ def load_cache(path):
             ref_hash = f.read().strip()
     else:
         ref_hash = None
-    logging.info("cache git hash: {}".format(ref_hash))
+    logging.debug("cache hash: {}".format(ref_hash))
     cache['ref_hash'] = ref_hash
     cache['ifos'] = {}
     for f in sorted(os.listdir(path)):
@@ -213,31 +266,52 @@ def main():
         help='specific ifos to test (default all)')
     args = parser.parse_args()
 
+    # get the current hash of HEAD
+    head_hash = git_ref_resolve_hash('HEAD')
+    if not head_hash:
+        logging.warning("could not determine git HEAD hash.")
+
+    # update the reference if specified
     if args.update_ref:
         if args.update_ref == 'HEAD':
-            ref_hash = subprocess.run(
-                ['git', 'show', '-s', '--format=format:%H', 'HEAD'],
-                capture_output=True, universal_newlines=True,
-            ).stdout
+            if not head_hash:
+                sys.exit("Could not update reference to head.")
+            logging.info("updating reference to HEAD...")
+            ref_hash = head_hash
         else:
-            ref_hash = args.update_ref
-        logging.info("updating reference git hash to {}...".format(ref_hash))
-        with open(test_path('ref_hash'), 'w') as f:
-            f.write('{}\n'.format(ref_hash))
+            ref_hash = git_ref_resolve_hash(args.update_ref)
+        logging.info("updating reference git hash: {}".format(ref_hash))
+        write_ref_hash(ref_hash)
         sys.exit()
 
+    # get the reference hash
     if args.git_ref:
-        ref_hash = args.git_ref
-    elif os.path.exists(test_path('ref_hash')):
-        with open(test_path('ref_hash')) as f:
-            ref_hash = f.read().strip()
+        ref_hash = git_ref_resolve_hash(args.git_ref)
     else:
-        sys.exit("Unspecified reference git hash, could not run test.")
+        ref_hash = load_ref_hash()
+        if not ref_hash:
+            pass
+        try:
+            with open(test_path('ref_hash')) as f:
+                ref_hash = f.read().strip()
+        except IOError:
+            logging.warning("could not open reference")
+            sys.exit("Unspecified reference git hash, could not run test.")
+
+    logging.info("head hash: {}".format(head_hash))
+    logging.info("ref  hash: {}".format(ref_hash))
+
+    # don't bother test if hashes match
+    if ref_hash == head_hash:
+        logging.info("HEAD matches reference, not bothering to calculate.")
+        logging.info("Use --git-ref to compare against an arbitrary git commit.")
+        sys.exit()
 
     # load the cache
     cache_path = test_path('cache', ref_hash)
     if not os.path.exists(cache_path):
-        gen_cache(ref_hash, cache_path)
+        prune_cache_dir()
+        gen_cache_for_ref(ref_hash, cache_path)
     cache = load_cache(cache_path)
 
     if args.report: