acryl-datahub 0.15.0.6rc3__py3-none-any.whl → 1.0.0__py3-none-any.whl

datahub/ingestion/source/metabase.py

@@ -69,9 +69,19 @@ class MetabaseConfig(DatasetLineageProviderConfigBase, StatefulIngestionConfigBa
         default=None,
         description="optional URL to use in links (if `connect_uri` is only for ingestion)",
     )
-    username: Optional[str] = Field(default=None, description="Metabase username.")
+    username: Optional[str] = Field(
+        default=None,
+        description="Metabase username, used when an API key is not provided.",
+    )
     password: Optional[pydantic.SecretStr] = Field(
-        default=None, description="Metabase password."
+        default=None,
+        description="Metabase password, used when an API key is not provided.",
+    )
+
+    # https://www.metabase.com/learn/metabase-basics/administration/administration-and-operation/metabase-api#example-get-request
+    api_key: Optional[pydantic.SecretStr] = Field(
+        default=None,
+        description="Metabase API key. If provided, the username and password will be ignored. Recommended method.",
     )
     # TODO: Check and remove this if no longer needed.
     # Config database_alias is removed from sql sources.
@@ -178,30 +188,40 @@ class MetabaseSource(StatefulIngestionSourceBase):
         self.source_config: MetabaseConfig = config
 
     def setup_session(self) -> None:
-        login_response = requests.post(
-            f"{self.config.connect_uri}/api/session",
-            None,
-            {
-                "username": self.config.username,
-                "password": (
-                    self.config.password.get_secret_value()
-                    if self.config.password
-                    else None
-                ),
-            },
-        )
+        self.session = requests.session()
+        if self.config.api_key:
+            self.session.headers.update(
+                {
+                    "x-api-key": self.config.api_key.get_secret_value(),
+                    "Content-Type": "application/json",
+                    "Accept": "*/*",
+                }
+            )
+        else:
+            # If no API key is provided, generate a session token using username and password.
+            login_response = requests.post(
+                f"{self.config.connect_uri}/api/session",
+                None,
+                {
+                    "username": self.config.username,
+                    "password": (
+                        self.config.password.get_secret_value()
+                        if self.config.password
+                        else None
+                    ),
+                },
+            )
 
-        login_response.raise_for_status()
-        self.access_token = login_response.json().get("id", "")
+            login_response.raise_for_status()
+            self.access_token = login_response.json().get("id", "")
 
-        self.session = requests.session()
-        self.session.headers.update(
-            {
-                "X-Metabase-Session": f"{self.access_token}",
-                "Content-Type": "application/json",
-                "Accept": "*/*",
-            }
-        )
+            self.session.headers.update(
+                {
+                    "X-Metabase-Session": f"{self.access_token}",
+                    "Content-Type": "application/json",
+                    "Accept": "*/*",
+                }
+            )
 
         # Test the connection
         try:
@@ -217,15 +237,17 @@ class MetabaseSource(StatefulIngestionSourceBase):
             )
 
     def close(self) -> None:
-        response = requests.delete(
-            f"{self.config.connect_uri}/api/session",
-            headers={"X-Metabase-Session": self.access_token},
-        )
-        if response.status_code not in (200, 204):
-            self.report.report_failure(
-                title="Unable to Log User Out",
-                message=f"Unable to logout for user {self.config.username}",
+        # API key authentication does not require session closure.
+        if not self.config.api_key:
+            response = requests.delete(
+                f"{self.config.connect_uri}/api/session",
+                headers={"X-Metabase-Session": self.access_token},
             )
+            if response.status_code not in (200, 204):
+                self.report.report_failure(
+                    title="Unable to Log User Out",
+                    message=f"Unable to logout for user {self.config.username}",
+                )
         super().close()
 
     def emit_dashboard_mces(self) -> Iterable[MetadataWorkUnit]:
@@ -291,7 +313,7 @@ class MetabaseSource(StatefulIngestionSourceBase):
             return None
 
         dashboard_urn = builder.make_dashboard_urn(
-            self.platform, dashboard_details.get("id", "")
+            self.platform, str(dashboard_details.get("id", ""))
         )
         dashboard_snapshot = DashboardSnapshot(
             urn=dashboard_urn,
@@ -315,7 +337,7 @@ class MetabaseSource(StatefulIngestionSourceBase):
             card_id = card_info.get("card").get("id", "")
             if not card_id:
                 continue  # most likely a virtual card without an id (text or heading), not relevant.
-            chart_urn = builder.make_chart_urn(self.platform, card_id)
+            chart_urn = builder.make_chart_urn(self.platform, str(card_id))
             chart_urns.append(chart_urn)
 
         dashboard_info_class = DashboardInfoClass(
@@ -437,7 +459,7 @@ class MetabaseSource(StatefulIngestionSourceBase):
             )
             return None
 
-        chart_urn = builder.make_chart_urn(self.platform, card_id)
+        chart_urn = builder.make_chart_urn(self.platform, str(card_id))
         chart_snapshot = ChartSnapshot(
             urn=chart_urn,
             aspects=[],

datahub/ingestion/source/metadata/business_glossary.py

@@ -1,5 +1,6 @@
 import logging
 import pathlib
+import re
 import time
 from dataclasses import dataclass, field
 from typing import Any, Dict, Iterable, List, Optional, TypeVar, Union
@@ -118,17 +119,58 @@ class BusinessGlossaryConfig(DefaultConfig):
         return v
 
 
+def clean_url(text: str) -> str:
+    """
+    Clean text for use in URLs by:
+    1. Replacing spaces with hyphens
+    2. Removing special characters (preserving hyphens and periods)
+    3. Collapsing multiple hyphens and periods into single ones
+    """
+    # Replace spaces with hyphens
+    text = text.replace(" ", "-")
+    # Remove special characters except hyphens and periods
+    text = re.sub(r"[^a-zA-Z0-9\-.]", "", text)
+    # Collapse multiple hyphens into one
+    text = re.sub(r"-+", "-", text)
+    # Collapse multiple periods into one
+    text = re.sub(r"\.+", ".", text)
+    # Remove leading/trailing hyphens and periods
+    text = text.strip("-.")
+    return text
+
+
 def create_id(path: List[str], default_id: Optional[str], enable_auto_id: bool) -> str:
+    """
+    Create an ID for a glossary node or term.
+
+    Args:
+        path: List of path components leading to this node/term
+        default_id: Optional manually specified ID
+        enable_auto_id: Whether to generate GUIDs
+    """
     if default_id is not None:
-        return default_id  # No need to create id from path as default_id is provided
+        return default_id  # Use explicitly provided ID
 
     id_: str = ".".join(path)
 
-    if UrnEncoder.contains_extended_reserved_char(id_):
-        enable_auto_id = True
+    # Check for non-ASCII characters before cleaning
+    if any(ord(c) > 127 for c in id_):
+        return datahub_guid({"path": id_})
 
     if enable_auto_id:
+        # Generate GUID for auto_id mode
         id_ = datahub_guid({"path": id_})
+    else:
+        # Clean the URL for better readability when not using auto_id
+        id_ = clean_url(id_)
+
+        # Force auto_id if the cleaned URL still contains problematic characters
+        if UrnEncoder.contains_extended_reserved_char(id_):
+            logger.warning(
+                f"ID '{id_}' contains problematic characters after URL cleaning. Falling back to GUID generation for stability."
+            )
+            id_ = datahub_guid({"path": id_})
+
     return id_
 
 

datahub/ingestion/source/metadata/lineage.py

@@ -104,8 +104,8 @@ class FineGrainedLineageConfig(ConfigModel):
 
 class EntityNodeConfig(ConfigModel):
     entity: EntityConfig
-    upstream: Optional[List["EntityNodeConfig"]]
-    fineGrainedLineages: Optional[List[FineGrainedLineageConfig]]
+    upstream: Optional[List["EntityNodeConfig"]] = None
+    fineGrainedLineages: Optional[List[FineGrainedLineageConfig]] = None
 
 
 # https://pydantic-docs.helpmanual.io/usage/postponed_annotations/ required for when you reference a model within itself