Look At Me Sideways (LAMS)
LAMS is a style guide and linter for Looker's LookML data modeling language. It is designed to help a team of developers to produce more maintainable LookML projects.
- The style guide alone can help your project, even without enforcement by the linter.
- The linter comes with a number of built-in rules from the style guide.
- The linter also allows you to conveniently specify custom rules to enforce.
- The linter can be deployed to enforce rules for all commits to your master branch.
Interested? See a video of LAMS in action!
Contents
Functionality & Features
Built-in Linter Rules
The linter comes with built-in rules that can enforce rules K1-4, F1-4, E1-2, T1-2, and W1 from the style guide.
As of LAMS v3, you must opt-in via your manifest.lkml
file to use the built-in rules. Here is an example declaration opting in to all the currently available built-in rules:
#LAMS
#rule: K1{} # Primary key naming
#rule: K3{} # Primary keys first
#rule: K4{} # Primary keys hidden
#rule: K7{} # Provide one `primary_key`
#rule: K8{} # `primary_key` uses PK dims
#rule: F1{} # No cross-view fields
#rule: F2{} # No view-labeled fields
#rule: F3{} # Count fields filtered
#rule: F4{} # Description or hidden
#rule: E1{} # Join with subst'n operator
#rule: E2{} # Join on PK for "one" joins
#rule: E6{} # FK joins are m:1
#rule: E7{} # Explore label 25-char max
#rule: T1{} # Triggers use datagroups
#rule: T2{} # Primary keys in DT
#rule: W1{} # Block indentation
#rule: W1{} # Block indentation
Custom Rules
In addition to linting against its style guide, LAMS also lets you specify your own rules. See Customizing LAMS.
Rule Exemptions
You can opt-out of rules granularly by locally specifying rule_exemptions
.
The rule exemption syntax encourages developers to document the reason for each such exemption:
view: rollup {
sql_table_name: my_table ;;
#LAMS
#rule_exemptions: {
# K3: "2018-11-12 - Bob said it's ok"
#}
dimension: info {...}
...
(BETA) You can also opt-out of rules granularly from a centrally maintained lams-exemptions.ndjson
file. Simply specify the rule name and location to exempt in a series of newline-terminated JSON objects:
{"rule":"K3","location":"model:my_model/view:rollup"}
{"rule":"K3","location":"model:my_other_model/view:foo"}
You may also apply rule_exemptions globally in your project.manifest, but this is generally unnecessary as of LAMS v3.
Output
Once LAMS has evaluated your project against the necessary rules, the resulting list of messages are communicated back to you through one of several output modes.
The default mode is a human-readable tabble logged to the command line / stdout, where each line represents a distinct message:
Another available output mode is formatting the messages into a markdown file in your project, so that they can be viewed in Looker's IDE. Here is an example of a resulting markdown file as displayed in Looker:
Deployment Examples
Although LAMS can be deployed in many ways to fit your specific CI flow, we have put together a few examples and resources to get you up and running quicker. (If you'd like to contribute your configuration, get in touch!)
Regardless of which example you follow, we recommend pinning your LAMS version to a particular major version.
- Local Interactive Usage - To use LAMS with the least overhead for simple interactive local use and testing:
npm install -g @looker/look-at-me-sideways@3
cd <your-lookml-project>
lams
- Github Action - This option is very quick to get started if you're using Github, and offers a compromise between convenience of setup and per-commit run performance.
The following examples were prepared for v1 of LAMS, though updating them for v2+ should be straightforward. Please review v2 release notes for details. In particular, look for error messages on the console's standard output rather than a file output to be committed back to the repo.
- GitLab CI - A community-contributed configuration for GitLab, which offers similarly low overhead as our dockerized Jenkins configuration
- Dockerized Jenkins Server - We have provided a Docker image with an end-to-end configuration including a Jenkins server, LAMS, and Github protected branches & status checks configuration.
- CircleCI - A community-contributed configuration for CircleCI (external link)
Configuration
Command-line arguments
- reporting - Required. One of
yes
,no
,save-yes
, orsave-no
. See PRIVACY.md for details. - report-user - An email address to use in reporting. See PRIVACY.md for details.
- report-license-key - A Looker license key to use in reporting. See PRIVACY.md for details.
- output - A comma-separated string of output modes from among:
lines
(default),markdown
,markdown-developer
,jenkins
,legacy-cli
, or (BETA)add-exemptions
- source - A glob specifying which files to read. Defaults to
**/{*.model,*.explore,*.view,manifest}.lkml
. - cwd - A path for LAMS to use as its current working directory. Useful if you are not invoking lams from your LookML repo directory.
- project-name - An optional name for the project, used to generate links back to the project in mardown output. Specifying this in manifest.lkml is preferred.
- manifest - A JSON-encoded object to override any properties that are normally set via the manifest.lkml file.
- on-parser-error - Set to "info" to indicate that LookML parsing errors should not fail the linter, but yield an
info
level message instead (not all output modes displayinfo
level messages) - verbose - Set to also output
verbose
level messages, for output modes that support it (lines
) - date-output - Set to "none" to skip printing the date at the top of the
issues.md
file. - allow-custom-rules - Experimental and not recommended. Used to approve the running of Javascript-based custom rules. DO NOT USE TO RUN UNTRUSTED CODE. See custom rules for details.
Manifest.lkml arguments
If your LookML project doesn't have a manifest.lkml file, you may want to consider adding one! LAMS uses the following information from your project's mainfest.lkml file:
- name - Recommended. A name for the project, used to generate links back to the project in mardown output. If the native LookML validator complains about an unnecessary project name, you can use a conditional #LAMS comment to specify it.
- rule: rule_name - Recommended. Used to opt-in to built-in rules and to specify custom rules. See customizing LAMS
- rule_exemptions - Optional. Originally used in 1 & v2 to opt-out of rules globally. A global opt-out can still be useful for opting-out of certain "sub rules" that a rule may return without opting-out of the entire rule. See customizing LAMS
About
Release Notes
Privacy Policy
LAMS respects user privacy. See PRIVACY.md for details.
License
LAMS is Copyright (c) 2023 Looker Data Sciences, Inc. and is licensed under the MIT License. See LICENSE.txt for license details.
Support
LAMS is NOT officially supported by Looker. Please do not contact Looker support for issues with LAMS. Issues may be reported via the Issues tracker, but no SLA or warranty exists that they will be resolved.
Authors
LAMS has primarily been developed by Joseph Axisa and Fabio Beltramini. See all contributors
Contributing
Bug reports and pull requests are welcome on GitHub at https://github.com/looker-open-source/look-at-me-sideways.
Trying to install LAMS for development?
git clone [email protected]:looker-open-source/look-at-me-sideways.git
cd look-at-me-sideways
mv npm-shrinkwrap.dev.json npm-shrinkwrap.json
npm install
Publishing an update? The following hooks will run:
npm version {minor|major|patch}
> Pre-verion: npm run lint-fix
> Pre-verion: npm run test
npm publish
> Pre-publish: npm shrinkwrap
> Pre-publish: mv npm-shrinkwrap.json npm-shrinkwrap.dev.json
> Pre-publish: npm prune --prod
> Pre-publish: npm shrinkwrap
Code of Conduct
This project is intended to be a safe, welcoming space for collaboration, and contributors are expected to adhere to the Contributer Covenant Code of Conduct. Concerns or incidents may be reported confidentially to [email protected].
Other LookML Linting Projects
Interested in LookML linting but looking to shop around a bit? The community has come up with a few other options:
- https://github.com/rbob86/lookml-linter
- Spectacles.dev's style validator
- https://github.com/ww-tech/lookml-tools/blob/master/README_LINTER.md
- https://pypi.org/project/lookmlint/
Or if you want to write your own, you may want to start with one of these LookML parsers:
- https://www.npmjs.com/package/lookml-parser (JS, used by LAMS)
- https://github.com/joshtemple/lkml (Python)
- https://github.com/drewgillson/lkml (JS port of the above)
>_>