Use URLs to specify data sources

ejacox · Nov 10, 2015 · 07c896c · 07c896c
1 parent 58b228b
commit 07c896c
Show file tree

Hide file tree

Showing 6 changed files with 20 additions and 15 deletions.
diff --git a/ga4gh/frontend.py b/ga4gh/frontend.py
@@ -172,12 +172,13 @@ def configure(configFile=None, baseConfig="ProductionConfig",
     cors.CORS(app, allow_headers='Content-Type')
     app.serverStatus = ServerStatus()
     # Allocate the backend
-    # TODO is this a good way to determine what type of backend we should
-    # instantiate? We should think carefully about this. The approach of
-    # using the special strings __SIMULATED__ and __EMPTY__ seems OK for
-    # now, but is certainly not ideal.
-    dataSource = app.config["DATA_SOURCE"]
-    if dataSource == "__SIMULATED__":
+    # We use URLs to specify the backend. Currently we have file:// URLs (or
+    # URLs with no scheme) for the FileSystemBackend, and special empty:// and
+    # simulated:// URLs for empty or simulated data sources.
+    dataSource = urlparse.urlparse(app.config["DATA_SOURCE"], "file")
+
+    if dataSource.scheme == "simulated":
+        # Ignore the query string
         randomSeed = app.config["SIMULATED_BACKEND_RANDOM_SEED"]
         numCalls = app.config["SIMULATED_BACKEND_NUM_CALLS"]
         variantDensity = app.config["SIMULATED_BACKEND_VARIANT_DENSITY"]
@@ -186,18 +187,22 @@ def configure(configFile=None, baseConfig="ProductionConfig",
             "SIMULATED_BACKEND_NUM_REFERENCE_SETS"]
         numReferencesPerReferenceSet = app.config[
             "SIMULATED_BACKEND_NUM_REFERENCES_PER_REFERENCE_SET"]
-        numAlignments = app.config[
+        numAlignmentsPerReadGroup = app.config[
             "SIMULATED_BACKEND_NUM_ALIGNMENTS_PER_READ_GROUP"]
         theBackend = backend.SimulatedBackend(
             randomSeed=randomSeed, numCalls=numCalls,
             variantDensity=variantDensity, numVariantSets=numVariantSets,
             numReferenceSets=numReferenceSets,
             numReferencesPerReferenceSet=numReferencesPerReferenceSet,
-            numAlignments=numAlignments)
-    elif dataSource == "__EMPTY__":
+            numAlignments=numAlignmentsPerReadGroup)
+    elif dataSource.scheme == "empty":
         theBackend = backend.EmptyBackend()
+    elif dataSource.scheme == "file":
+        theBackend = backend.FileSystemBackend(os.path.join(
+            dataSource.netLoc, dataSource.path))
     else:
-        theBackend = backend.FileSystemBackend(dataSource)
+        raise exceptions.ConfigurationException(
+            "Unsupported data source scheme: " + dataSource.scheme)
     theBackend.setRequestValidation(app.config["REQUEST_VALIDATION"])
     theBackend.setResponseValidation(app.config["RESPONSE_VALIDATION"])
     theBackend.setDefaultPageSize(app.config["DEFAULT_PAGE_SIZE"])

diff --git a/ga4gh/serverconfig.py b/ga4gh/serverconfig.py
@@ -20,7 +20,7 @@ class BaseConfig(object):
     REQUEST_VALIDATION = False
     RESPONSE_VALIDATION = False
     DEFAULT_PAGE_SIZE = 100
-    DATA_SOURCE = "__EMPTY__"
+    DATA_SOURCE = "empty://"
 
     # Options for the simulated backend.
     SIMULATED_BACKEND_RANDOM_SEED = 0

diff --git a/tests/end_to_end/server.py b/tests/end_to_end/server.py
@@ -172,7 +172,7 @@ def getConfig(self):
         config = """
 SIMULATED_BACKEND_NUM_VARIANT_SETS = 10
 SIMULATED_BACKEND_VARIANT_DENSITY = 1
-DATA_SOURCE = "__SIMULATED__"
+DATA_SOURCE = "simulated://"
 DEBUG = True
 """
         if self.useOidc:

diff --git a/tests/unit/test_oidc.py b/tests/unit/test_oidc.py
@@ -100,11 +100,11 @@ class TestFrontendOidc(unittest.TestCase):
     @classmethod
     def setUpClass(cls):
         config = {
-            "DATA_SOURCE": "__SIMULATED__",
             "SIMULATED_BACKEND_RANDOM_SEED": 1111,
             "SIMULATED_BACKEND_NUM_CALLS": 1,
             "SIMULATED_BACKEND_VARIANT_DENSITY": 1.0,
             "SIMULATED_BACKEND_NUM_VARIANT_SETS": 1,
+            "DATA_SOURCE": "simulated://",
             "OIDC_CLIENT_ID": "123",
             "OIDC_CLIENT_SECRET": RANDSTR,
             "OIDC_PROVIDER": "http://auth.com"

diff --git a/tests/unit/test_simulated_stack.py b/tests/unit/test_simulated_stack.py
@@ -29,7 +29,7 @@ def setUpClass(cls):
         # Set the random seed to make tests reproducible.
         random.seed(1)
         config = {
-            "DATA_SOURCE": "__SIMULATED__",
+            "DATA_SOURCE": "simulated://",
             "SIMULATED_BACKEND_RANDOM_SEED": 1111,
             "SIMULATED_BACKEND_NUM_CALLS": 5,
             "SIMULATED_BACKEND_VARIANT_DENSITY": 1.0,

diff --git a/tests/unit/test_views.py b/tests/unit/test_views.py
@@ -22,7 +22,7 @@ class TestFrontend(unittest.TestCase):
     @classmethod
     def setUpClass(cls):
         config = {
-            "DATA_SOURCE": "__SIMULATED__",
+            "DATA_SOURCE": "simulated://",
             "SIMULATED_BACKEND_RANDOM_SEED": 1111,
             "SIMULATED_BACKEND_NUM_CALLS": 1,
             "SIMULATED_BACKEND_VARIANT_DENSITY": 1.0,