fix(preprocess): bind DWH artifacts to workspace

This commit is contained in:
2026-07-12 06:46:06 +02:00
parent 0a2421c513
commit b43075289e
4 changed files with 146 additions and 24 deletions
+7 -4
View File
@@ -20,9 +20,12 @@ def run_dwh_from_config(
):
from tht.cli.lsh_cmd import build_lsh_artifacts
from tht.cli.schema_cmd import _load_config_or_exit, physical_path, refresh_catalog
from tht.jobs.dwh_pipeline import DwhPreprocessPipeline, active_generation_dir, fingerprint
from tht.jobs.dwh_pipeline import (
DwhPreprocessPipeline, active_generation_dir, config_dwh_binding,
)
cfg = _load_config_or_exit(config)
binding = config_dwh_binding(cfg)
workspace_root = cfg.paths.artifacts.parent
active = active_generation_dir(workspace_root)
lsh_names = (
@@ -31,10 +34,10 @@ def run_dwh_from_config(
f"{cfg.database.db_schema}_meta.json",
)
pipeline = DwhPreprocessPipeline(
workspace_id=config.stem.lower().replace(".", "-").replace("_", "-"),
workspace_id=binding["workspace_id"],
workspace_root=workspace_root,
config_fingerprint=fingerprint(cfg.model_dump_json()),
input_fingerprint=fingerprint(config.resolve().as_posix()),
config_fingerprint=binding["config_fingerprint"],
input_fingerprint=binding["input_fingerprint"],
introspect=lambda output: refresh_catalog(cfg, output_path=output),
build_lsh=lambda physical, output: build_lsh_artifacts(
cfg, physical_file=physical, output_dir=output