perf: cache parsed manifest in APIDbtRunner#2196
Closed
dtaniwaki wants to merge 2 commits intoelementary-data:masterfrom
Closed
perf: cache parsed manifest in APIDbtRunner#2196dtaniwaki wants to merge 2 commits intoelementary-data:masterfrom
dtaniwaki wants to merge 2 commits intoelementary-data:masterfrom
Conversation
Reuse the parsed manifest across sequential dbt run-operation calls instead of re-parsing it on every invocation. This reduces edr report overhead from ~18 manifest parses down to 1.
Cover three scenarios: - manifest is cached after the first successful invocation - manifest is not cached when invocation fails - cached manifest is reused on subsequent calls
Contributor
|
👋 @dtaniwaki |
dtaniwaki
requested a deployment
to
elementary_test_env
GitHub Actions
Waiting
📝 WalkthroughWalkthroughThis change implements manifest caching in the Changes
Sequence DiagramsequenceDiagram
participant Caller
participant APIDbtRunner as APIDbtRunner<br/>(_manifest=None)
participant dbtRunner as dbtRunner
participant Manifest as Manifest
Caller->>APIDbtRunner: _inner_run_command() [1st call]
APIDbtRunner->>dbtRunner: dbtRunner(..., manifest=None)
dbtRunner->>Manifest: run & generate manifest
dbtRunner-->>APIDbtRunner: result + manifest
APIDbtRunner->>APIDbtRunner: self._manifest = result.manifest
APIDbtRunner-->>Caller: result
Caller->>APIDbtRunner: _inner_run_command() [2nd call]
APIDbtRunner->>dbtRunner: dbtRunner(..., manifest=self._manifest)
dbtRunner->>Manifest: run with cached manifest
dbtRunner-->>APIDbtRunner: result
APIDbtRunner-->>Caller: result
Estimated Code Review Effort🎯 2 (Simple) | ⏱️ ~12 minutes Poem
🚥 Pre-merge checks | ✅ 4 | ❌ 1❌ Failed checks (1 warning)
✅ Passed checks (4 passed)
✏️ Tip: You can configure your own custom pre-merge checks in the settings. ✨ Finishing Touches🧪 Generate unit tests (beta)
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
There was a problem hiding this comment.
Actionable comments posted: 1
🤖 Prompt for all review comments with AI agents
Verify each finding against the current code and only fix it if needed.
Inline comments:
In `@tests/unit/clients/dbt_runner/test_api_dbt_runner.py`:
- Around line 17-23: Replace the helper that bypasses APIDbtRunner.__init__ with
one that calls the real initializer (so _manifest is set by the class) and avoid
the hardcoded "/tmp" by using a temporary directory fixture; e.g., change
_make_runner to call APIDbtRunner(...) with project_dir=str(tmp_path) (or a
tempfile.TemporaryDirectory context) and monkeypatch/stub any heavy external
behavior if needed instead of assigning runner._manifest manually, ensuring
tests exercise the actual __init__ behavior and remove the Ruff S108 `/tmp`
literal.
🪄 Autofix (Beta)
Fix all unresolved CodeRabbit comments on this PR:
- Push a commit to this branch (recommended)
- Create a new PR with the fixes
ℹ️ Review info
⚙️ Run configuration
Configuration used: defaults
Review profile: CHILL
Plan: Pro
Run ID: aa6ee93f-7cc2-43d8-865d-e4273823bcf5
📒 Files selected for processing (2)
elementary/clients/dbt/api_dbt_runner.pytests/unit/clients/dbt_runner/test_api_dbt_runner.py
Comment on lines
+17
to
+23
| def _make_runner(): | ||
| runner = APIDbtRunner.__new__(APIDbtRunner) | ||
| runner._manifest = None | ||
| runner.project_dir = "/tmp/fake" | ||
| runner.env_vars = None | ||
| runner.raise_on_failure = False | ||
| return runner |
There was a problem hiding this comment.
Cover the real initializer and avoid the linted /tmp path.
_make_runner() is fine for isolating _inner_run_command, but it manually sets _manifest, so these tests would still pass if APIDbtRunner.__init__ stopped initializing _manifest before super().__init__(). Also, Line 20 is flagged by Ruff S108.
Proposed test/helper adjustment
def _make_runner():
runner = APIDbtRunner.__new__(APIDbtRunner)
runner._manifest = None
- runner.project_dir = "/tmp/fake"
+ runner.project_dir = "fake_project_dir"
runner.env_vars = None
runner.raise_on_failure = False
return runner
+
+
+def test_manifest_initialized_before_parent_init():
+ def assert_manifest_initialized(self, *args, **kwargs):
+ assert hasattr(self, "_manifest")
+ assert self._manifest is None
+
+ with mock.patch(
+ "elementary.clients.dbt.api_dbt_runner.CommandLineDbtRunner.__init__",
+ assert_manifest_initialized,
+ ):
+ APIDbtRunner(project_dir="fake_project_dir")🧰 Tools
🪛 Ruff (0.15.10)
[error] 20-20: Probable insecure usage of temporary file or directory: "/tmp/fake"
(S108)
🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed.
In `@tests/unit/clients/dbt_runner/test_api_dbt_runner.py` around lines 17 - 23,
Replace the helper that bypasses APIDbtRunner.__init__ with one that calls the
real initializer (so _manifest is set by the class) and avoid the hardcoded
"/tmp" by using a temporary directory fixture; e.g., change _make_runner to call
APIDbtRunner(...) with project_dir=str(tmp_path) (or a
tempfile.TemporaryDirectory context) and monkeypatch/stub any heavy external
behavior if needed instead of assigning runner._manifest manually, ensuring
tests exercise the actual __init__ behavior and remove the Ruff S108 `/tmp`
literal.
Author
|
Closing in favor of #2197 which should address this more effectively. |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Hi, thanks for building and maintaining Elementary! I'd love to get your review on this small performance improvement.
Problem
edr reportexecutes ~18 sequentialdbt run-operationcalls via the Python API. Each call creates a freshdbtRunner()without passing a manifest, which forces a full manifest re-parse every time. On a project with hundreds of models, sources, and macros, each parse can take ~10 seconds — adding up to ~3 minutes of overhead while the actual queries finish in seconds.Solution
Cache the parsed manifest on
APIDbtRunnerafter the first successful invocation and reuse it for all subsequentdbtRunner(manifest=...)calls. This brings the number of manifest parses from ~18 down to 1.Key details:
self._manifestis initialized beforesuper().__init__()because the parent constructor may call_inner_run_command()via_run_deps_if_needed().dbtRunner.manifestis a public API available since dbt-core 1.5+.APIDbtRunneris affected —SubprocessDbtRunnerandDbtFusionRunnerare unchanged.Summary by CodeRabbit