Skip to content

v1.0.6 — dbt migration overhaul

Choose a tag to compare

@roypulseai roypulseai released this 30 Jun 16:38

1.0.6 — dbt migration overhaul + field-test fixes

Fixed (field-test feedback from jaffle-shop migration)

  • CTE parser false circular dependencies — CTE aliases no longer create outgoing dependency edges; prevents false cycle detection when a CTE name collides with a real model name
  • .sql seeds don't create tables — bare SELECT seed files now auto-wrapped in CREATE OR REPLACE TABLE
  • Jinja macros not converted during import — auto-translates cents_to_dollars, dbt.date_trunc, dbt_utils.generate_surrogate_key, safe_divide, is_incremental, dbt_date.get_base_dates
  • Materialization configs not preserved — parses dbt_project.yml per-folder +materialized: configs; emits -- materialized: headers
  • SchemaYaml empty after import — scans ALL .yml/.yaml files with models: or sources: key
  • compare --dbt silent failure — pre-flight check for dbt target/ directory
  • preview rendering on Windowslegacy_windows=False for proper grid rendering
  • security classify --table — added --table/-t option alias
  • Windows state DB file lock — retry with sleep in kelpmesh clean

Added

  • 12 new dbt_compat runtime macros (44 total): cents_to_dollars, dollars_to_cents, dbt_current_timestamp, dbt_type_string/numeric/bigint/int/timestamp/date/boolean/float
  • Interactive kelpmesh migrate wizard — scans, reports, confirms, imports, generates report
  • --compat dbt flag for kelpmesh init — dbt-style directory layout (staging/marts/intermediate)
  • Migration reportMIGRATION_REPORT.md auto-generated after import
  • CSV seed preservation — original .csv files kept alongside .sql wrappers
  • Schema YAML + sources preservation — copied into output project
  • Migration docs rewrite — macro translation table, wizard usage, comparison instructions

Full changelog: https://github.com/roypulseai/kelpmesh/blob/main/CHANGELOG.md
Docs: https://roypulseai.github.io/kelpmesh/
PyPI: pip install kelpmesh-core==1.0.6