refactor: existing python code

refactor: existing python code refactor: existing python code revert unintended change feat: rust code fix configuration loading feat: fix settings (rust) fix configuration loading feat: fix settings (rust) feat: rust code debugging feat: rust code debugging feat: rust code debugging feat: rust code debugging feat: rust code debugging fix configuration loading feat: fix settings (rust) fix configuration loading feat: fix settings (rust) feat: rust code debugging feat: rust code debugging feat: rust code debugging feat: rust code debugging feat: rust code debugging refactor: existing python code revert unintended change feat: rust code fix configuration loading feat: fix settings (rust) fix configuration loading feat: fix settings (rust) feat: rust code debugging feat: rust code debugging feat: rust code debugging feat: rust code debugging feat: rust code debugging fix configuration loading feat: fix settings (rust) fix configuration loading feat: fix settings (rust) feat: rust code debugging feat: rust code debugging feat: rust code debugging feat: rust code debugging feat: rust code debugging refactor: exiting python code fix configuration loading feat: fix settings (rust) fix configuration loading feat: fix settings (rust) feat: rust code debugging feat: rust code debugging feat: rust code debugging feat: rust code debugging feat: rust code debugging fix configuration loading feat: fix settings (rust) fix configuration loading feat: fix settings (rust) feat: rust code debugging feat: rust code debugging feat: rust code debugging feat: rust code debugging feat: rust code debugging refactor: existing python code revert unintended change feat: rust code fix configuration loading feat: fix settings (rust) fix configuration loading feat: fix settings (rust) feat: rust code debugging feat: rust code debugging feat: rust code debugging feat: rust code debugging feat: rust code debugging fix configuration loading feat: fix settings (rust) fix configuration loading feat: fix settings (rust) feat: rust code debugging feat: rust code debugging feat: rust code debugging feat: rust code debugging feat: rust code debugging refactor: existing python code refactor: existing python code refactor: existing python code refactor: existing python code
dasch-swiss · Feb 24, 2025 · 14184b7 · 14184b7
1 parent e2d9d12
commit 14184b7
Show file tree

Hide file tree

Showing 10 changed files with 296 additions and 80 deletions.
diff --git a/python/ark_resolver/__init__.py → python/src/ark_resolver/__init__.py b/python/ark_resolver/__init__.py → python/src/ark_resolver/__init__.py
diff --git a/python/ark_resolver/ark-config.ini → python/src/ark_resolver/ark-config.ini b/python/ark_resolver/ark-config.ini → python/src/ark_resolver/ark-config.ini
diff --git a/python/ark_resolver/ark-registry.ini → python/src/ark_resolver/ark-registry.ini b/python/ark_resolver/ark-registry.ini → python/src/ark_resolver/ark-registry.ini
@@ -41,7 +41,7 @@ PhpResourceVersionRedirectUrl : http://$host/resources/$resource_int_id?citdate=
 ProjectHost : meta.dasch.swiss
 
 ############################################################################
-# anything test project
+# anything tests project
 
 [0001]
 

diff --git a/python/ark_resolver/ark.py → python/src/ark_resolver/ark.py b/python/ark_resolver/ark.py → python/src/ark_resolver/ark.py
@@ -19,13 +19,41 @@
 from asyncio import sleep
 from io import StringIO
 
+from opentelemetry import trace
+from opentelemetry.sdk.trace import TracerProvider
+from opentelemetry.sdk.trace.export import (
+    BatchSpanProcessor,
+    ConsoleSpanExporter,
+)
+
+import sentry_sdk
+from sentry_sdk.integrations.asyncio import AsyncioIntegration
+from sentry_sdk.integrations.sanic import SanicIntegration
+from sentry_sdk.integrations.rust_tracing import RustTracingIntegration
+
 import requests
 from sanic import HTTPResponse, Sanic, response
 from sanic.log import logger
 from sanic_cors import CORS
 
-import ark_resolver.base64url_check_digit as base64url_check_digit_py
-from ark_resolver.ark_url import (ArkUrlException, ArkUrlFormatter, ArkUrlInfo, ArkUrlSettings)
+import _rust
+from ark_resolver import ark_url
+import ark_resolver.check_digit as check_digit_py
+import ark_resolver.health
+
+
+#################################################################################################
+# OpenTelemetry
+
+provider = TracerProvider()
+processor = BatchSpanProcessor(ConsoleSpanExporter())
+provider.add_span_processor(processor)
+
+# Sets the global default tracer provider
+trace.set_tracer_provider(provider)
+
+# Creates a tracer from the global tracer provider
+tracer = trace.get_tracer("my.tracer.name")
 
 #################################################################################################
 # Server implementation.
@@ -35,8 +63,53 @@
 app = Sanic('ark_resolver')
 CORS(app)
 
+# Register health check route
+app.blueprint(ark_resolver.health.health_bp)
+
+
+@app.before_server_start
+async def init_sentry(_):
+    sentry_dsn = os.environ.get("ARK_SENTRY_DSN", None)
+    sentry_debug = os.environ.get("ARK_SENTRY_DEBUG", "False")
+    sentry_environment = os.environ.get("ARK_SENTRY_ENVIRONMENT", None)
+    sentry_release = os.environ.get("ARK_SENTRY_RELEASE", None)
+    if sentry_dsn:
+        sentry_sdk.init(
+            dsn=sentry_dsn,
+            debug=sentry_debug,
+            environment=sentry_environment,
+            release=sentry_release,
+            # Add data like request headers and IP for users;
+            # see https://docs.sentry.io/platforms/python/data-management/data-collected/ for more info
+            send_default_pii=True,
+            # Set traces_sample_rate to 1.0 to capture 100%
+            # of transactions for tracing.
+            traces_sample_rate=1.0,
+            # Set profiles_sample_rate to 1.0 to profile 100%
+            # of sampled transactions.
+            # We recommend adjusting this value in production.
+            profiles_sample_rate=1.0,
+            instrumenter="otel",
+            integrations=[
+                AsyncioIntegration(),
+                RustTracingIntegration(
+                    "_rust",
+                    _rust.initialize_tracing,
+                    include_tracing_fields=True,
+                ),
+                SanicIntegration(
+                    # Configure the Sanic integration so that we generate
+                    # transactions for all HTTP status codes, including 404
+                    unsampled_statuses=None,
+                ),
+            ],
+        )
+        logger.info("Sentry initialized.")
+    else:
+        logger.info("No SENTRY_DSN found in environment variables. Sentry will not be initialized.")
+
 
-def get_config() -> str:
+def get_safe_config() -> str:
     """
     Returns the app's configuration
     """
@@ -55,20 +128,22 @@ def get_config() -> str:
     return safe_config_output.getvalue()
 
 
+@tracer.start_as_current_span("config_get")
 @app.get("/config")
-async def config_get(_) -> HTTPResponse:
+async def safe_config_get(_) -> HTTPResponse:
     """
     Returns the app's configuration
     """
-    return response.text(get_config())
+    return response.text(get_safe_config())
 
 
+@tracer.start_as_current_span("config_head")
 @app.head("/config")
-async def config_head(_) -> HTTPResponse:
+async def safe_config_head(_) -> HTTPResponse:
     """
     Returns only the head of the config response
     """
-    config_str = get_config()
+    config_str = get_safe_config()
 
     headers = {
         "Content-Length": str(len(config_str)),
@@ -78,6 +153,7 @@ async def config_head(_) -> HTTPResponse:
     return response.text("", headers=headers)
 
 
+@tracer.start_as_current_span("reload")
 @app.post("/reload")
 async def reload(req) -> HTTPResponse:
     """
@@ -108,23 +184,28 @@ async def reload(req) -> HTTPResponse:
     else:
         return response.text("Unauthorized", status=401)
 
-
+@tracer.start_as_current_span("redirect")
 @app.get('/<path:path>')
 async def catch_all(_, path="") -> HTTPResponse:
     """
     Catch all URL. Tries to redirect the given ARK ID.
     """
     try:
-        redirect_url = ArkUrlInfo(settings=app.config.settings, ark_url=path, path_only=True).to_redirect_url()
-    except ArkUrlException as ex:
+        redirect_url = ark_url.ArkUrlInfo(settings=app.config.settings, ark_id=path).to_redirect_url()
+
+    except ark_url.ArkUrlException as ex:
+        logger.error(f"Invalid ARK ID: {path}", exc_info=ex)
         return response.text(body=ex.message, status=400)
 
-    except base64url_check_digit_py.CheckDigitException as ex:
+    except check_digit_py.CheckDigitException as ex:
+        logger.error(f"Invalid ARK ID: {path}", exc_info=ex)
         return response.text(body=ex.message, status=400)
 
-    except KeyError:
+    except KeyError as ex:
+        logger.error(f"Invalid ARK ID: {path}", exc_info=ex)
         return response.text(body="Invalid ARK ID", status=400)
 
+    logger.info(f"Redirecting {path} to {redirect_url}")
     return response.redirect(redirect_url)
 
 
@@ -158,7 +239,7 @@ def reload_config() -> None:
 #################################################################################################
 # Loading of config and registry files.
 
-def load_settings(config_path: str) -> ArkUrlSettings:
+def load_settings(config_path: str) -> ark_url.ArkUrlSettings:
     """
     Loads configuration from given path and returns an ArkUrlSettings.
     """
@@ -187,7 +268,7 @@ def load_settings(config_path: str) -> ArkUrlSettings:
     else:
         config.read_file(open(registry_path))
 
-    settings = ArkUrlSettings(config)
+    settings = ark_url.ArkUrlSettings(config)
 
     return settings
 
@@ -205,8 +286,10 @@ def main() -> None:
     parser.add_argument("-c", "--config", help="config file (default {})".format(default_config_path))
     group = parser.add_mutually_exclusive_group()
     group.add_argument("-s", "--server", help="start server", action="store_true")
-    group.add_argument("-i", "--iri", help="print the converted ARK URL from a given DSP resource IRI (add -v and -d optionally)")
-    group.add_argument("-a", "--ark", help="print the converted DSP resource IRI (requires -r) or DSP URL from a given ARK URL")
+    group.add_argument("-i", "--iri",
+                       help="print the converted ARK URL from a given DSP resource IRI (add -v and -d optionally)")
+    group.add_argument("-a", "--ark",
+                       help="print the converted DSP resource IRI (requires -r) or DSP URL from a given ARK ID")
     parser.add_argument("-r", "--resource", help="generate resource IRI", action="store_true")
     parser.add_argument("-v", "--value", help="value UUID (has to be provided with -i)")
     parser.add_argument("-d", "--date", help="DSP ARK timestamp (has to be provided with -i)")
@@ -226,20 +309,20 @@ def main() -> None:
             server(settings)
         elif args.iri:
             # prints the converted ARK URL from a given DSP resource IRI
-            print(ArkUrlFormatter(settings).resource_iri_to_ark_url(args.iri, args.value, args.date))
+            print(ark_url.ArkUrlFormatter(settings).resource_iri_to_ark_url(args.iri, args.value, args.date))
         elif args.ark:
             if args.resource:
                 # prints the converted DSP resource IRI from a given ARK URL
-                print(ArkUrlInfo(settings, args.ark).to_resource_iri())
+                print(ark_url.ArkUrlInfo(settings, args.ark).to_resource_iri())
             else:
                 # prints the converted DSP URL from a given ARK URL
-                print(ArkUrlInfo(settings, args.ark).to_redirect_url())
+                print(ark_url.ArkUrlInfo(settings, args.ark).to_redirect_url())
         else:
             parser.print_help()
-    except ArkUrlException as ex:
+    except ark_url.ArkUrlException as ex:
         print(ex.message)
         exit(1)
-    except base64url_check_digit_py.CheckDigitException as ex:
+    except check_digit_py.CheckDigitException as ex:
         print(ex.message)
         exit(1)
 

diff --git a/python/ark_resolver/ark_url.py → python/src/ark_resolver/ark_url.py b/python/ark_resolver/ark_url.py → python/src/ark_resolver/ark_url.py
@@ -9,7 +9,7 @@
 from string import Template
 from urllib import parse
 
-import ark_resolver.base64url_check_digit as base64url_check_digit_py
+import ark_resolver.check_digit as check_digit_py
 
 
 #################################################################################################
@@ -56,42 +56,27 @@ def __init__(self, message):
 
 class ArkUrlInfo:
     """
-    Represents the information retrieved from a DSP ARK URL.
+    Represents the information retrieved from a DSP ARK ID.
     """
 
-    def __init__(self, settings, ark_url, path_only=False):
+    def __init__(self, settings, ark_id):
         self.settings = settings
 
-        # Are we matching just the path part of the URL?
-        # TODO: path_only=True should be tested in unit tests
-        if path_only:
-            # Yes. Is it a version 1 ARK ID?
-            match = settings.ark_path_regex.match(ark_url)
-
-            if match:
-                # Yes.
-                self.url_version = int(match.group(1))
-            else:
-                # No. Is it a version 0 ARK ID?
-                match = settings.v0_ark_path_regex.match(ark_url)
-
-                if match is not None:
-                    self.url_version = 0
+        match = settings.ark_path_regex.match(ark_id)
 
+        if match:
+            # Yes. Is it a version 1 ARK ID?
+            self.url_version = int(match.group(1))
         else:
-            # We are matching a whole URL. Does it contain a version 1 ARK ID?
-            match = settings.ark_url_regex.match(ark_url)
+            # No. Is it a version 0 ARK ID?
+            match = settings.v0_ark_path_regex.match(ark_id)
+
+            # If NOT None!, then it is a version 0 ARK ID.
             if match is not None:
-                # Yes.
-                self.url_version = int(match.group(1))
-            else:
-                # No. Does it contain a version 0 ARK ID?
-                match = settings.v0_ark_url_regex.match(ark_url)
+                self.url_version = 0
 
-                if match is not None:
-                    self.url_version = 0
         if match is None:
-            raise ArkUrlException(f"Invalid ARK ID: {ark_url}")
+            raise ArkUrlException(f"Invalid ARK ID: {ark_id}")
 
         # Which version of ARK ID did we match?
         if self.url_version == settings.dsp_ark_version:
@@ -101,15 +86,15 @@ def __init__(self, settings, ark_url, path_only=False):
 
             if escaped_resource_id_with_check_digit is not None:
                 self.resource_id = unescape_and_validate_uuid(
-                    ark_url=ark_url,
+                    ark_url=ark_id,
                     escaped_uuid=escaped_resource_id_with_check_digit
                 )
 
                 escaped_value_id_with_check_digit = match.group(4)
 
                 if escaped_value_id_with_check_digit is not None:
                     self.value_id = unescape_and_validate_uuid(
-                        ark_url=ark_url,
+                        ark_url=ark_id,
                         escaped_uuid=escaped_value_id_with_check_digit
                     )
                 else:
@@ -136,9 +121,10 @@ def __init__(self, settings, ark_url, path_only=False):
             project_config = self.settings.config[self.project_id]
 
             if not project_config.getboolean("AllowVersion0"):
-                raise ArkUrlException(f"Invalid ARK ID (version 0 not allowed): {ark_url}")
+                raise ArkUrlException(f"Invalid ARK ID (version 0 not allowed): {ark_id}")
         else:
-            raise ArkUrlException(f"Invalid ARK ID {ark_url}. The version of the ARK ID doesn't match the version defined in the settings.")
+            raise ArkUrlException(
+                f"Invalid ARK ID {ark_id}. The version of the ARK ID doesn't match the version defined in the settings.")
 
         self.template_dict = {
             "url_version": self.url_version,
@@ -181,7 +167,8 @@ def to_resource_iri(self) -> str:
             # in case of an ARK URL version 0, the resource_id generated from the salsah ID has to be converted to a
             # base64 UUID version 5
             generic_namespace_url = uuid.NAMESPACE_URL
-            dasch_uuid_ns = uuid.uuid5(generic_namespace_url, "https://dasch.swiss")  # cace8b00-717e-50d5-bcb9-486f39d733a2
+            dasch_uuid_ns = uuid.uuid5(generic_namespace_url,
+                                       "https://dasch.swiss")  # cace8b00-717e-50d5-bcb9-486f39d733a2
             resource_id = template_dict["resource_id"]
             dsp_iri = base64.urlsafe_b64encode(uuid.uuid5(dasch_uuid_ns, resource_id).bytes).decode("utf-8")
             # remove the padding ('==') from the end of the string
@@ -273,7 +260,7 @@ def add_check_digit_and_escape(uuid) -> str:
     """
     Adds a check digit to a Base64-encoded UUID, and escapes the result.
     """
-    check_digit = base64url_check_digit_py.calculate_check_digit(uuid)
+    check_digit = check_digit_py.calculate_check_digit(uuid)
     uuid_with_check_digit = uuid + check_digit
 
     # Escape '-' as '=' in the resource ID and check digit, because '-' can be ignored in ARK URLs.
@@ -287,7 +274,7 @@ def unescape_and_validate_uuid(ark_url, escaped_uuid) -> str:
     # '-' is escaped as '=' in the UUID and check digit, because '-' can be ignored in ARK URLs.
     unescaped_uuid = escaped_uuid.replace('=', '-')
 
-    if not base64url_check_digit_py.is_valid(unescaped_uuid):
+    if not check_digit_py.is_valid(unescaped_uuid):
         raise ArkUrlException(f"Invalid ARK ID: {ark_url}")
 
     return unescaped_uuid[0:-1]

diff --git a/python/ark_resolver/base64url_check_digit.py → python/src/ark_resolver/check_digit.py b/python/ark_resolver/base64url_check_digit.py → python/src/ark_resolver/check_digit.py
diff --git a/python/test/create_mock_settings.py b/python/test/create_mock_settings.py
diff --git a/python/test/__init__.py → python/tests/__init__.py b/python/test/__init__.py → python/tests/__init__.py