[CT-28] Spike moving dbt to a binary installation

[nathaniel-may]

We should explore the tradeoffs for moving from pip installing source to distributing a binary. We would bundle one version of the Python interpreter to run the Python source code.

Pros

• No more maintaining python compatibility for old versions
• dbt developers and maintainers get access to the latest Python features
• Users won't be subject to experiencing strange behaviors while using unsupported, newer versions of Python, and also won't log that class of ticket anymore.
• Significantly less CI steps to run
• Debugging has one less dimension to check for developers and the support team
• Solidifies interface boundaries because consumers of dbt would no longer be able to directly call arbitrary parts of dbt
• pip install could still work if we distribute the dbt binary as a wheel.
• All installs of the same dbt version use the same versions of dependencies that went through our CI regardless of the version ranges we specify in dbt source.

Cons

• Consumers of dbt would no longer be able to directly call arbitrary parts of dbt, which would require consumers currently relying on this strategy to use the proper interfaces before upgrading.
• Our current strategy for plugins is to rely on pip placing code inside dbt-core with directory matching which would no longer work. A new strategy would need to be explored and decided on. One possible alternative is to have adapters be the single point of install where each adapter depends on a specific version of dbt core.

Issues

A small sample of issues this packaging strategy would prevent:

• [CT-255] [Bug] ImportError: cannot import name soft_unicode from markupsafe #4745
• [CT-32] [Bug] cannot run dbt - module typing has no attribute OrderedDict #4564
• TODO add others

[nathaniel-may]

An example of a user issue that would be eliminated by this packaging strategy: #4564

[github-actions]

This issue has been marked as Stale because it has been open for 180 days with no activity. If you would like the issue to remain open, please remove the stale label or comment on the issue, or it will be closed in 7 days.

[github-actions]

Although we are closing this issue as stale, it's not gone forever. Issues can be reopened if there is renewed community interest; add a comment to notify the maintainers.

[github-actions]

This issue has been marked as Stale because it has been open for 180 days with no activity. If you would like the issue to remain open, please comment on the issue or else it will be closed in 7 days.

[github-actions]

This issue has been marked as Stale because it has been open for 180 days with no activity. If you would like the issue to remain open, please comment on the issue or else it will be closed in 7 days.

[github-actions]

Although we are closing this issue as stale, it's not gone forever. Issues can be reopened if there is renewed community interest. Just add a comment to notify the maintainers.