Changelog

v2.0 (compared to v1.0)

• Added features
  • Database ingestion via new connectors for Postgres, Snowflake, Redshift, and Athena with optional AWS Secrets Manager hydration for credentials. Data sources can also be external CSVs and annotations merged into the same daily frame.
  • Data loading now auto-merges multiple queries and CSV sources keyed on Date, supporting aliasing per query name for collision-free metrics.

• Annotations support via an annotations list in the WBR YAML, allowing one or more CSVs (local paths or URLs) to be loaded and rendered as special events alongside metrics in the report.

• Changed behavior

  • CSV upload is now optional; if omitted the backend loads data from setup.db_config_url using the data_sources map. Queries must alias the first column as "Date" and metrics reference query_name.column.
  • A CSV provided at request time overrides all YAML data_sources/csv_files; to merge CSV with DB queries, list CSVs under data_sources.csv_files in the YAML instead of uploading ad hoc.
  • Validation enforces setup.week_ending presence/format and tighter metric config checks for aggregation/function definitions.
  • YAML generation endpoint now reuses the new data loading path and falls back to the rules-based generator if the plugin fails.

• Removed or deprecated functions

  • No removals; existing endpoints remain.

• Test updates

  • New unit tests cover connector factory and individual connectors plus validator/controller utility behaviors.

• Documentation updates

  • README expanded with database configuration guidance; new docs/user_manual.md added covering CSV override precedence, DB/CSV merging rules, and updated sample YAML showing data_sources aliasing conventions.

• Dependency updates

  • Bumped core libs (Flask/Werkzeug, pandas, numpy, matplotlib), cloud SDKs (boto3, google-cloud-, azure-), DB drivers (psycopg2-binary, snowflake-connector-python), and pytest to current minor versions. Kept cryptography pinned for Snowflake compatibility.