Merge pull request #4 from bento-platform/features/service-registry-v…

…ersion

Features/service registry version

.gitignore

@@ -6,3 +6,4 @@ htmlcov/
 .idea/workspace.xml
 .idea/dataSources*
 /chord_services.json
+.vscode

bento_service_registry/app.py

@@ -2,6 +2,7 @@
 import os
 import requests
 import sys
+import subprocess
 
 from bento_lib.responses.flask_errors import (
     flask_error_wrap,
@@ -22,6 +23,7 @@
 
 
 application = Flask(__name__)
+
 application.config.from_mapping(
     BENTO_DEBUG=os.environ.get("CHORD_DEBUG", os.environ.get("FLASK_ENV", "production")).strip().lower() in (
         "true", "1", "development"),
@@ -32,6 +34,8 @@
     URL_PATH_FORMAT=os.environ.get("URL_PATH_FORMAT", "api/{artifact}"),
 )
 
+path_for_git = os.path.abspath(os.path.join(os.path.dirname(__file__), ".."))
+
 # Generic catch-all
 application.register_error_handler(Exception, flask_error_wrap_with_traceback(flask_internal_server_error,
                                                                               sr_compat=True,
@@ -44,6 +48,20 @@ def get_service_url(artifact: str):
     return urljoin(current_app.config["CHORD_URL"], current_app.config["URL_PATH_FORMAT"].format(artifact=artifact))
 
 
+def get_chord_services():
+    """
+    Reads the list of services from the chord_services.json file
+    """
+    services = []
+    try:
+        with open(current_app.config["CHORD_SERVICES"], "r") as f:
+            services = [s for s in json.load(f) if not s.get("disabled")]  # Skip disabled services
+    except Exception as e:
+        except_name = type(e).__name__
+        print("Error in retrieving services information from json file.", except_name)
+    return services
+
+
 with application.app_context():
     SERVICE_ID = current_app.config["SERVICE_ID"]
     SERVICE_INFO = {
@@ -56,77 +74,93 @@ def get_service_url(artifact: str):
             "url": "http://www.computationalgenomics.ca"
         },
         "contactUrl": "mailto:[email protected]",
-        "version": bento_service_registry.__version__
+        "version": bento_service_registry.__version__,
+        "url": get_service_url(SERVICE_ARTIFACT),
+        "environment": "prod"
     }
 
-    with open(current_app.config["CHORD_SERVICES"], "r") as f:
-        CHORD_SERVICES = [s for s in json.load(f) if not s.get("disabled")]  # Skip disabled services
-
-    service_info_cache = {
-        # Pre-populate service-info cache with data for the current service
-        SERVICE_ARTIFACT: {**SERVICE_INFO, "url": get_service_url(SERVICE_ARTIFACT)},
-    }
+    chord_services_content = get_chord_services()   # Not a constant: can change when a service is updated
 
 
 def get_service(service_artifact):
-    s_url = get_service_url(service_artifact)
+    # special case: requesting info about the current service. Avoids request timeout
+    # when running gunicorn on a single worker
+    if service_artifact == SERVICE_ARTIFACT:
+        return _service_info()
 
-    if service_artifact not in service_info_cache:
-        service_info_url = urljoin(f"{s_url}/", "service-info")
+    s_url = get_service_url(service_artifact)
+    service_info_url = urljoin(f"{s_url}/", "service-info")
 
-        print(f"[{SERVICE_NAME}] Contacting {service_info_url}", flush=True)
+    print(f"[{SERVICE_NAME}] Contacting {service_info_url}", flush=True)
 
-        # Optional Authorization HTTP header to forward to nested requests
-        # TODO: Move X-Auth... constant to bento_lib
-        auth_header = request.headers.get("X-Authorization", request.headers.get("Authorization"))
+    # Optional Authorization HTTP header to forward to nested requests
+    # TODO: Move X-Auth... constant to bento_lib
+    auth_header = request.headers.get("X-Authorization", request.headers.get("Authorization"))
 
-        try:
-            r = requests.get(
-                service_info_url,
-                headers={"Authorization": auth_header} if auth_header else {},
-                timeout=current_app.config["CONTACT_TIMEOUT"],
-                verify=not current_app.config["BENTO_DEBUG"],
-            )
+    service_resp = {}
 
-            if r.status_code != 200:
-                print(f"[{SERVICE_NAME}] Non-200 status code on {service_artifact}: {r.status_code}\n"
-                      f"                 Content: {r.text}", file=sys.stderr, flush=True)
+    try:
+        r = requests.get(
+            service_info_url,
+            headers={"Authorization": auth_header} if auth_header else {},
+            timeout=current_app.config["CONTACT_TIMEOUT"],
+            verify=not current_app.config["BENTO_DEBUG"],
+        )
 
-                # If we have the special case where we got a JWT error from the proxy script, we can safely print out
-                # headers for debugging, since the JWT leaked isn't valid anyway.
-                if "invalid jwt" in r.text:
-                    print(f"                 Encountered auth error, tried to use header: {auth_header}",
-                          file=sys.stderr, flush=True)
+        if r.status_code != 200:
+            print(f"[{SERVICE_NAME}] Non-200 status code on {service_artifact}: {r.status_code}\n"
+                  f"                 Content: {r.text}", file=sys.stderr, flush=True)
 
-                return None
+            # If we have the special case where we got a JWT error from the proxy script, we can safely print out
+            # headers for debugging, since the JWT leaked isn't valid anyway.
+            if "invalid jwt" in r.text:
+                print(f"                 Encountered auth error, tried to use header: {auth_header}",
+                      file=sys.stderr, flush=True)
 
-        except requests.exceptions.Timeout:
-            print(f"[{SERVICE_NAME}] Encountered timeout with {service_info_url}", file=sys.stderr, flush=True)
             return None
 
-        try:
-            service_info_cache[service_artifact] = {**r.json(), "url": s_url}
-        except JSONDecodeError:
-            print(f"[{SERVICE_NAME}] Encountered invalid response from {service_info_url}: {r.text}")
-            return None
+    except requests.exceptions.Timeout:
+        print(f"[{SERVICE_NAME}] Encountered timeout with {service_info_url}", file=sys.stderr, flush=True)
+        return None
+
+    try:
+        service_resp[service_artifact] = {**r.json(), "url": s_url}
+    except JSONDecodeError:
+        print(f"[{SERVICE_NAME}] Encountered invalid response from {service_info_url}: {r.text}")
+        return None
 
-    return service_info_cache[service_artifact]
+    return service_resp[service_artifact]
+
+
+@application.before_first_request
+def before_first_request_func():
+    try:
+        subprocess.run(["git", "config", "--global", "--add", "safe.directory", str(path_for_git)])
+    except Exception as e:
+        except_name = type(e).__name__
+        print("Error in dev-mode retrieving git folder configuration", except_name)
 
 
 @application.route("/bento-services")
 @application.route("/chord-services")
 def chord_services():
-    return jsonify(CHORD_SERVICES)
+    return jsonify(get_chord_services())
+
+
+def _services():
+    return [s for s in (
+        get_service(s["type"]["artifact"]) for s in chord_services_content
+        ) if s is not None]
 
 
 @application.route("/services")
 def services():
-    return jsonify([s for s in (get_service(s["type"]["artifact"]) for s in CHORD_SERVICES) if s is not None])
+    return jsonify(_services())
 
 
 @application.route("/services/<string:service_id>")
 def service_by_id(service_id):
-    services_by_id = {s["id"]: s for s in service_info_cache.values()}
+    services_by_id = {s["id"]: s for s in _services()}
     if service_id not in services_by_id:
         return flask_not_found_error(f"Service with ID {service_id} was not found in registry")
 
@@ -136,10 +170,33 @@ def service_by_id(service_id):
 
 @application.route("/services/types")
 def service_types():
-    return jsonify(sorted(set(s["type"] for s in service_info_cache.values())))
+    return jsonify(sorted(set(s["type"] for s in _services())))
+
+
+def _service_info():
+    if not current_app.config["BENTO_DEBUG"]:
+        return SERVICE_INFO
+
+    info = {
+        **SERVICE_INFO,
+        "environment": "dev"
+    }
+    try:
+        res_tag = subprocess.check_output(["git", "describe", "--tags", "--abbrev=0"])
+        if res_tag:
+            info["git_tag"] = res_tag.decode().rstrip()
+        res_branch = subprocess.check_output(["git", "branch", "--show-current"])
+        if res_branch:
+            info["git_branch"] = res_branch.decode().rstrip()
+
+    except Exception as e:
+        except_name = type(e).__name__
+        print("Error in dev-mode retrieving git information", except_name)
+
+    return info  # updated service info with the git info
 
 
 @application.route("/service-info")
 def service_info():
     # Spec: https://github.com/ga4gh-discovery/ga4gh-service-info
-    return jsonify(SERVICE_INFO)
+    return jsonify(_service_info())

bento_service_registry/package.cfg

@@ -1,5 +1,5 @@
 [package]
 name = bento_service_registry
-version = 0.5.1
+version = 0.6.0
 authors = David Lougheed
 author_emails = [email protected]

setup.py

@@ -17,7 +17,7 @@
     python_requires=">=3.6",
     install_requires=[
         "bento_lib[flask]==2.2.1",
-        "Flask>=1.1.2,<2.0", 
+        "Flask>=1.1.2,<2.0",
         "requests>=2.25.1,<3.0",
     ],
 

tests/conftest.py

@@ -5,6 +5,7 @@
 def _setup_env():
     os.environ["CHORD_SERVICES"] = os.path.join(os.path.dirname(__file__), "chord_services.json")
     os.environ["URL_PATH_FORMAT"] = ""  # Mount off of root URL for testing
+    os.environ["CHORD_DEBUG"] = ""
 
 
 @pytest.fixture()

tests/test_api.py

@@ -1,12 +1,9 @@
-import json
-
-
 def test_service_info(client, service_info):
     r = client.get("/service-info")
     d = r.get_json()
     # TODO: Check against service-info schema
     assert r.status_code == 200
-    assert json.dumps(d, sort_keys=True) == json.dumps(service_info, sort_keys=True)
+    assert d == service_info
 
 
 def test_chord_service_list(client):
@@ -22,20 +19,14 @@ def test_service_list(client, service_info):
     d = r.get_json()
     assert r.status_code == 200
     assert len(d) == 1
-    assert json.dumps(d[0], sort_keys=True) == json.dumps({
-        **service_info,
-        "url": "http://127.0.0.1:5000/"
-    }, sort_keys=True)
+    assert d[0] == service_info
 
 
 def test_service_detail(client, service_info):
     r = client.get(f"/services/{service_info['id']}")
     d = r.get_json()
     assert r.status_code == 200
-    assert json.dumps(d, sort_keys=True) == json.dumps({
-        **service_info,
-        "url": "http://127.0.0.1:5000/"
-    }, sort_keys=True)
+    assert d == service_info
 
     r = client.get("/services/does-not-exist")
     assert r.status_code == 404