Merge branch 'main' into er/build-tooling

README.md

@@ -1,8 +1,6 @@
-** replace `dbt-common` with your repository name in all docs
-
 ## Understanding dbt-common
 
-A short description of the purpose of this repository.  Include relevant images.
+The shared common utilities for dbt-core and adapter implementations use
 
 ## Getting started
 

dbt_common/__about__.py

@@ -1 +1 @@
-version = "0.0.1"
+version = "0.1.0"

docs/arch/adr-0001-build-tooling.md

@@ -0,0 +1,80 @@
+# Build Tool
+
+
+## Context
+
+We need to select a build tool for managing dependencies for, building, and distributing `dbt-common`.  While tooling can vary, this repo can serve as an OSS template for other OSS at dbt Labs and beyond.
+
+
+## Options
+
+- `setuptools` (`twine`, `build`)
+- `hatch`
+- `poetry`
+
+
+### setuptools
+
+#### Pro's
+
+- most popular option
+- supported by Python Packaging Authority
+- build tool of record for dbt-core & existing internal adapters
+
+#### Con's
+
+- less flexible; forced to support backwards compatibility more so than other options
+- no dependency management (manually add to `pyproject.toml`)
+
+
+### hatch
+
+#### Pro's
+
+- supported by Python Packaging Authority
+- already used by `dbt-semantic-layer` and `internal-actions`
+- supports running tests against multiple versions of python locally (same functionality as `tox`)
+- supports configuring workflows in `pyproject.toml` (same functionality as `make`)
+- incorporates new PEP's quickly
+- Manages python distributions itself without need of pyenv.  This allows Windows and non-Windows users to both work locally in the same way.
+- used by black, tox, pipx, Jupyter Notebook, Datadog
+
+#### Con's
+
+- far less popular than other options
+- no dependency management (manually add to `pyproject.toml`)
+- only one maintainer (but is officially part of the larger PyPA working group)
+- Hatch does not allow for the installation of specific patch release versions of itself but rather only uses minor release granularity that tracks the latest patch release
+
+
+### poetry
+
+#### Pro's
+
+- second most popular option, similar in popularity to `setuptools`
+- dependency management (`poetry add "my-dependency"`)
+- provides a lock file
+- more than one maintainer
+
+#### Con's
+
+- incorporates new PEP's slowly
+
+
+## Decision
+
+#### Selected: `hatch`
+
+This option aligns with `dbt-adapter` and `dbt-semantic-layer`, which minimizes confusion
+for anyone working in multiple repositories.
+`hatch` also replaces `tox` and `make`, which consolidates our toolset to make working locally and with CI more consistant.
+
+
+## Consequences
+
+- [+] retire `tox`
+- [+] retire `make`
+- [+] rewriting the release workflows will create a more intuitive release for hatch projects
+- [-] we cannot reuse the existing release workflows
+- [-] write more detailed docs given lower familiarity
+- [-] learning curve

pyproject.toml

@@ -86,7 +86,9 @@ extra-dependencies = [
     "types-protobuf~=4.24.0",
     "types-python-dateutil~=2.8",
     "types-PyYAML~=6.0",
-    "types-requests<2.31.0" # types-requests 2.31.0.8 requires urllib3>=2, but we pin urllib3 ~= 1.0 because of openssl requirement for requests
+    "types-requests<2.31.0" # types-requests 2.31.0.8 requires 
+
+    lib3>=2, but we pin urllib3 ~= 1.0 because of openssl requirement for requests
 
 ]