## Changes
This is necessary to ensure that our Terraform provider can use the same
auxiliary programs (e.g. `az`, or `gcloud`) as the CLI.
## Tests
Unit test and manual verification.
~(this should be changed to target `main`)~
This reveals the template from
https://github.com/databricks/cli/pull/686 in CLI prompts for once #686
and #708 are merged.
---------
Co-authored-by: Andrew Nester <andrew.nester@databricks.com>
Co-authored-by: PaulCornellDB <paul.cornell@databricks.com>
Co-authored-by: Pieter Noordhuis <pieter.noordhuis@databricks.com>
## Changes
This adds a built-in "default-python" template to the CLI. This is based
on the new default-template support of
https://github.com/databricks/cli/pull/685.
The goal here is to offer an experience where customers can simply type
`databricks bundle init` to get a default template:
```
$ databricks bundle init
Template to use [default-python]: default-python
Unique name for this project [my_project]: my_project
✨ Successfully initialized template
```
The present template:
- [x] Works well with VS Code
- [x] Works well with the workspace
- [x] Works well with DB Connect
- [x] Uses minimal stubs rather than boiler-plate-heavy examples
I'll have a followup with tests + DLT support.
---------
Co-authored-by: Andrew Nester <andrew.nester@databricks.com>
Co-authored-by: PaulCornellDB <paul.cornell@databricks.com>
Co-authored-by: Pieter Noordhuis <pieter.noordhuis@databricks.com>
## Changes
The latest rendition of isServicePrincipal no longer worked for
non-admin users as it used the "principals get" API.
This new version relies on the property that service principals always
have a UUID as their userName. This was tested with the eng-jaws
principal (8b948b2e-d2b5-4b9e-8274-11b596f3b652).
## Changes
This reduces the latency of every workspace command by the duration of a
single API call to retrieve the current user (which can take up to a
full second).
Note: the better place to verify that a request can be authenticated is
the SDK itself.
## Tests
* Unit test to confirm an the empty `*http.Request` can be constructed
* Manually confirmed that the additional API call no longer happens
## Changes
JSON schema properties are a map and thus unordered.
This PR introduces a JSON schema extension field called `order` to allow
template authors to define the order in which template variables should
be resolved/prompted.
## Tests
Unit tests.
---------
Co-authored-by: Pieter Noordhuis <pieter.noordhuis@databricks.com>
## Changes
Before:
```
Usage:
databricks instance-pools [command]
Available Commands:
create Create a new instance pool.
delete Delete an instance pool.
edit Edit an existing instance pool.
get Get instance pool information.
get-permission-levels Get instance pool permission levels.
get-permissions Get instance pool permissions.
list List instance pool info.
set-permissions Set instance pool permissions.
update-permissions Update instance pool permissions.
```
After:
```
Usage:
databricks instance-pools [command]
Available Commands
create Create a new instance pool.
delete Delete an instance pool.
edit Edit an existing instance pool.
get Get instance pool information.
list List instance pool info.
Permission Commands
get-permission-levels Get instance pool permission levels.
get-permissions Get instance pool permissions.
set-permissions Set instance pool permissions.
update-permissions Update instance pool permissions.
```
## Tests
Manual.
## Changes
* Update Go SDK to v0.19.0
* Update commands per OpenAPI spec from Go SDK
* Incorporate `client.Do()` signature change to include a (nil) header
map
* Update `workspace.WorkspaceService` mock with permissions methods
* Skip `files` service in codegen; already implemented under the `fs`
command
## Tests
Unit and integration tests pass.
# Warning: breaking change
## Changes
Instead of having paths in bundle config files be relative to bundle
root even if the config file is nested, this PR makes such paths
relative to the folder where the config is located.
When bundle is initialised, these paths will be transformed to relative
paths based on bundle root. For example,
we have file structure like this
```
- mybundle
| - bundle.yml
| - subfolder
| -- resource.yml
| -- my.whl
```
Previously, we had to reference `my.whl` in resource.yml like this,
which was confusing because resource.yml is in the same subfolder
```
sync:
include:
- ./subfolder/*.whl
...
tasks:
- task_key: name
libraries:
- whl: ./subfolder/my.whl
...
```
After the change we can reference it like this (which is in line with
the current behaviour for notebooks)
```
sync:
include:
- ./*.whl
...
tasks:
- task_key: name
libraries:
- whl: ./my.whl
...
```
## Tests
Existing `translate_path_tests` successfully passed after refactoring.
Added a couple of uses cases for `Libraries` paths.
Added a bundle config tests with include config and sync section
---------
Co-authored-by: Pieter Noordhuis <pieter.noordhuis@databricks.com>
## Changes
@pietern this addresses a comment from you on a recently merged PR. It
also updates settings.json based on the settings VS Code adds as soon as
you edit a notebook.
## Changes
Version 4 enables caching by default so we no longer need to explicitly
enable it: https://github.com/actions/setup-go#v4.
The build cache only reuses a cache from a repo's default branch, which
for this repository is `main`. After enabling the merge queue, we no
longer run builds on the `main` branch after push, but on merge queue
branches. With no more builds on the `main` branch there is no longer a
cache to reuse.
This change fixes that by making the `release(-snapshot)?` workflows use
the same caching mechanism. These run off of the `main` branch, so the
cache they save can be reused by builds triggered on PRs or from the
merge queue.
## Tests
We have to merge this to see if it works.
## Changes
The installer doesn't respect the version constraints if they are
specified.
Source: [the vc argument is not
used](850464c601/releases/latest_version.go (L158-L177)).
## Tests
Confirmed manually.
## Changes
The provider at version 1.24.0 includes a regression for the MLflow
model resource.
To fix this, we explicitly pin the provider version at the version we
generate bindings for.
## Tests
Confirmed that a deploy of said MLflow model resource works with 1.23.0.
## Changes
***Note: this PR relies on sync.include functionality from here:
https://github.com/databricks/cli/pull/671***
Added transformation mutator for Python wheel task for them to work on
DBR <13.1
Using wheels upload to Workspace file system as cluster libraries is not
supported in DBR < 13.1
In order to make Python wheel work correctly on DBR < 13.1 we do the
following:
1. Build and upload python wheel as usual
2. Transform python wheel task into special notebook task which does the
following
a. Installs all necessary wheels with %pip magic
b. Executes defined entry point with all provided parameters
3. Upload this notebook file to workspace file system
4. Deploy transformed job task
This is also beneficial for executing on existing clusters because this
notebook always reinstall wheels so if there are any changes to the
wheel package, they are correctly picked up
## Tests
bundle.yml
```yaml
bundle:
name: wheel-task
workspace:
host: ****
resources:
jobs:
test_job:
name: "[${bundle.environment}] My Wheel Job"
tasks:
- task_key: TestTask
existing_cluster_id: "***"
python_wheel_task:
package_name: "my_test_code"
entry_point: "run"
parameters: ["first argument","first value","second argument","second value"]
libraries:
- whl: ./dist/*.whl
```
Output
```
andrew.nester@HFW9Y94129 wheel % databricks bundle run test_job
Run URL: ***
2023-08-03 15:58:04 "[default] My Wheel Job" TERMINATED SUCCESS
Output:
=======
Task TestTask:
Hello from my func
Got arguments v1:
['python', 'first argument', 'first value', 'second argument', 'second value']
```
## Changes
Now if the user reference local Python wheel files and do not specify
"artifacts" section, this file will be automatically uploaded by CLI.
Fixes#693
## Tests
Added unit tests
Ran bundle deploy for this configuration
```
resources:
jobs:
some_other_job:
name: "[${bundle.environment}] My Wheel Job"
tasks:
- task_key: TestTask
existing_cluster_id: ${var.job_existing_cluster}
python_wheel_task:
package_name: "my_test_code"
entry_point: "run"
libraries:
- whl: ./dist/*.whl
```
Result
```
andrew.nester@HFW9Y94129 wheel % databricks bundle deploy
artifacts.whl.AutoDetect: Detecting Python wheel project...
artifacts.whl.AutoDetect: No Python wheel project found at bundle root folder
Starting upload of bundle files
Uploaded bundle files at /Users/andrew.nester@databricks.com/.bundle/wheel-task/default/files!
artifacts.Upload(my_test_code-0.0.1-py3-none-any.whl): Uploading...
artifacts.Upload(my_test_code-0.0.1-py3-none-any.whl): Upload succeeded
```
## Changes
This pull request extends the templating support in preparation of a
new, default template (WIP, https://github.com/databricks/cli/pull/686):
* builtin templates that can be initialized using e.g. `databricks
bundle init default-python`
* builtin templates are embedded into the executable using go's `embed`
functionality, making sure they're co-versioned with the CLI
* new helpers to get the workspace name, current user name, etc. help
craft a complete template
* (not enabled yet) when the user types `databricks bundle init` they
can interactively select the `default-python` template
And makes two tangentially related changes:
* IsServicePrincipal now uses the "users" API rather than the
"principals" API, since the latter is too slow for our purposes.
* mode: prod no longer requires the 'target.prod.git' setting. It's hard
to set that from a template. (Pieter is planning an overhaul of warnings
support; this would be one of the first warnings we show.)
The actual `default-python` template is maintained in a separate PR:
https://github.com/databricks/cli/pull/686
## Tests
Unit tests, manual testing
## Changes
Added run_as section for bundle configuration.
This section allows to define an user name or service principal which
will be applied as an execution identity for jobs and DLT pipelines. In
the case of DLT, identity defined in `run_as` will be assigned
`IS_OWNER` permission on this pipeline.
## Tests
Added unit tests for configuration.
Also ran deploy for the following bundle configuration
```
bundle:
name: "run_as"
run_as:
# service_principal_name: "f7263fcc-56d0-4981-8baf-c2a45296690b"
user_name: "lennart.kats@databricks.com"
resources:
pipelines:
andrew_pipeline:
name: "Andrew Nester pipeline"
libraries:
- notebook:
path: ./test.py
jobs:
job_one:
name: Job One
tasks:
- task_key: "task"
new_cluster:
num_workers: 1
spark_version: 13.2.x-snapshot-scala2.12
node_type_id: i3.xlarge
runtime_engine: PHOTON
notebook_task:
notebook_path: "./test.py"
```
This command takes the user through the interactive flow to set up OAuth
for a fresh account, where only Basic authentication works.
---------
Co-authored-by: Andrew Nester <andrew.nester@databricks.com>
## Changes
The pattern `.*` in a `.gitignore` file can match `.` when walking all
files in a repository. If it does, then the walker immediately aborts
and no files are returned. The root directory (an unnamed directory)
must never be ignored.
Reported in https://github.com/databricks/databricks-vscode/issues/837.
## Tests
New tests pass.
## Changes
This flag allows users to initialize a template from a subdirectory in
the repo root. Also enables multi template repositories.
## Tests
Manually
## Changes
This PR:
1. Renames the project-dir flag to output-dir
2. Makes the project dir flag optional. When unspecified we default to
the current working directory.
## Tests
Manually
---------
Co-authored-by: Pieter Noordhuis <pieter.noordhuis@databricks.com>
## Changes
Renamed Environments to Targets in bundle.yml.
The change is backward-compatible and customers can continue to use
`environments` in the time being.
## Tests
Added tests which checks that both `environments` and `targets` sections
in bundle.yml works correctly
## Changes
This is not desirable and will be addressed by representing our
configuration in a different structure (e.g. with cty, or with
plain `any`), instead of Go structs.
## Tests
Pass.
## Changes
Go text templates allows only specifying one input argument for
invocations of associated templates (ie `{{template ...}}`). This PR
introduces the map and pair functions which allow template authors to
work around this limitation by passing multiple arguments as key value
pairs in a map.
This PR is based on feedback from the mlops stacks migration where
otherwise a bunch of duplicate code is required for computed values and
fixtures.
## Tests
Unit test
## Changes
Prompt UI glitches often. We are switching to a custom implementation of
a simple prompter which is much more stable.
This also allows new lines in prompts which has been an ask by the
mlflow team.
## Tests
Tested manually
## Changes
Adds a function to validate json schema types added by the author. The
default json unmarshaller does not validate that the parsed type matches
the enum defined in `jsonschema.Type`
Includes some other improvements to provide better error messages.
This PR was prompted by usability difficulties reported by @mingyu89
during mlops stack migration.
## Tests
Unit tests
## Changes
Originally, these blocks were merged with overrides. This was
(inadvertently) disabled in #94. This change re-enables merging these
blocks with overrides, such that any field set in an environment
override always takes precedence over the field set in the base
definition.
## Tests
New unit test passes.
## Changes
#629 introduced a change to autopopulate the host from .databrickscfg if
the user is logging back into a host they were previously using. This
did not respect the DATABRICKS_CONFIG_FILE env variable, causing the
flow to stop working for users with no .databrickscfg file in their home
directory.
This PR refactors all config file loading to go through one interface,
`databrickscfg.GetDatabricksCfg()`, and an auxiliary
`databrickscfg.GetDatabricksCfgPath()` to get the configured file path.
Closes#655.
## Tests
```
$ databricks auth login --profile abc
Error: open /Users/miles/.databrickscfg: no such file or directory
$ ./cli auth login --profile abc
Error: cannot load Databricks config file: open /Users/miles/.databrickscfg: no such file or directory
$ DATABRICKS_CONFIG_FILE=~/.databrickscfg.bak ./cli auth login --profile abc
Databricks Host: https://asdf
```
## Changes
While they are a slice, we can identify a job cluster by its job cluster
key. A job definition with multiple job clusters with the same key is
always invalid. We can therefore merge definitions with the same key
into one. This is compatible with how environment overrides are applied;
merging a slice means appending to it. The override will end up in the
job cluster slice of the original, which gives us a deterministic way to
merge them.
Since the alternative is an invalid configuration, this doesn't change
behavior.
## Tests
New test coverage.
## Changes
The `.tmpl` extension is only meant as a qualifier for whether the file
content is executed as a template. All file paths in the `template`
directory should be treated as valid go text templates.
Before only paths with the `.tmpl` extensions would be resolved as
templates, after this change, all file paths are interpreted as
templates.
## Tests
Unit test. The newly added unit tests also asserts that the file path is
correct, even when the `.tmpl` extension is missing.
Pieter Noordhuis
shreyas-goenka
Lennart Kats (databricks)
Andrew Nester
Serge Smertin
Miles Yucht