odctl 0.2.0__tar.gz → 0.2.2__tar.gz

{odctl-0.2.0 → odctl-0.2.2}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: odctl
-Version: 0.2.0
+Version: 0.2.2
 Summary: Orchestrator for the Open Data Stack
 Project-URL: Homepage, https://github.com/jaehyeon-kim/odctl
 Project-URL: Repository, https://github.com/jaehyeon-kim/odctl

{odctl-0.2.0 → odctl-0.2.2}/pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "odctl"
-version = "0.2.0"
+version = "0.2.2"
 description = "Orchestrator for the Open Data Stack"
 readme = "README.md"
 authors = [

{odctl-0.2.0 → odctl-0.2.2}/src/odctl/resources/compose-infra.yml

@@ -124,10 +124,13 @@ services:
       - CATALOG_WAREHOUSE=s3://warehouse/
       - CATALOG_IO__IMPL=org.apache.iceberg.aws.s3.S3FileIO
       - CATALOG_S3_ENDPOINT=http://seaweed:8333
-      - CATALOG_S3_ACCESS_KEY_ID=user
-      - CATALOG_S3_SECRET_ACCESS_KEY=password
-      - CATALOG_S3_PATH_STYLE_ACCESS=true
-      - CATALOG_IO__S3_PATH_STYLE_ACCESS=true
+      # The image maps a double underscore to a hyphen and a single one to a dot,
+      # so hyphenated Iceberg properties need __ or they resolve to names Iceberg
+      # ignores. With path-style access silently dropped, S3FileIO addressed
+      # buckets virtual-host style and SeaweedFS read the namespace as the bucket.
+      - CATALOG_S3_ACCESS__KEY__ID=user
+      - CATALOG_S3_SECRET__ACCESS__KEY=password
+      - CATALOG_S3_PATH__STYLE__ACCESS=true
       # Explicit AWS SDK overrides (just in case)
       - AWS_REGION=us-east-1
       - AWS_ACCESS_KEY_ID=${AWS_ACCESS_KEY_ID:-user}

{odctl-0.2.0 → odctl-0.2.2}/src/odctl/resources/registry.yml

@@ -202,8 +202,8 @@ stacks:
         • Trino UI & JDBC connection (user: user, password: none):
           * External (Localhost): http://localhost:8080
           * Internal (Docker): http://trino:8080
-        • Catalogs: iceberg, kafka (JSON topics), kafka_avro (Avro via
-          the schema registry), clickhouse, postgres, valkey
+        • Catalogs: iceberg, kafka (topics discovered from the schema
+          registry), clickhouse, postgres, valkey
       metabase: |
         • Metabase Dashboard Engine:
           * External (Localhost): http://localhost:3000

odctl-0.2.0/src/odctl/resources/trino/catalog/kafka_avro.properties → odctl-0.2.2/src/odctl/resources/trino/catalog/kafka.properties

@@ -2,9 +2,9 @@ connector.name=kafka
 kafka.nodes=broker-1:19092
 kafka.default-schema=default
 # Tables are discovered from the schema registry, so topics carrying Confluent
-# Avro appear without being listed and their fields arrive typed. This supplier
-# replaces file-based discovery rather than extending it, which is why plain JSON
-# topics live in the separate 'kafka' catalog.
+# Avro appear without being listed and their fields arrive typed. The file-based
+# supplier cannot work this way: it requires every topic to be named up front,
+# which no general-purpose stack can know.
 kafka.table-description-supplier=CONFLUENT
 kafka.confluent-schema-registry-url=http://karapace:8081
 kafka.hide-internal-columns=false

odctl-0.2.0/src/odctl/resources/trino/catalog/kafka.properties

@@ -1,7 +0,0 @@
-connector.name=kafka
-kafka.nodes=broker-1:19092
-kafka.default-schema=default
-kafka.table-names=web_events,logs
-# Expose _message, _key and the other underscore columns so raw JSON topics can
-# be read without a table description file
-kafka.hide-internal-columns=false