[CT-679] [Bug] when shelling out to dbt, disabling colored output still prints out the reset color code

[darin-reify]

Is there an existing issue for this?

• I have searched the existing issues

Current Behavior

when shelling out to dbt (a la Prefect), disabling colored output still prints the reset color code.

Expected Behavior

no color codes should be printed when colored output is disabled

Steps To Reproduce

1. install dbt >= 1.0.0
2. set up a dbt project
3. start python and shell out to dbt via your preferred method making sure to disable colored output
4. notice the reset color code from the stdout stream

Relevant log output

# running 'dbt --no-use-colors run'
�[0m23:06:09  Running with dbt=1.1.0
�[0m23:06:09  Found 11 models, 10 tests, 0 snapshots, 0 analyses, 172 macros, 1 operation, 0 seed files, 2 sources, 0 exposures, 0 metrics

### Environment

```markdown
- OS: Debian 11
- Python: 3.9.12
- dbt: 1.1.0

What database are you using dbt with?

redshift

Additional Context

this line here looks to be the culprit

[gshank]

I tried this in a standard terminal and it works fine. Is this behavior dependant on using that DbtShellTask that you linked to? Because if it works fine in a normal terminal and doesn't underneath that application, I'm wondering if there is an issue with DbtShellTask.

[darin-reify]

it's probably more of a problem with anything that wants programmatic access to the logs for whatever reason.

here's a super small snippet (modified from DbtShellTask's parent functionality) showing what's going on:

import os
from subprocess import Popen, PIPE


with Popen(["dbt", "--no-use-colors", "compile"], stdout=PIPE, shell=False) as sub_process:
    lines = []
    for raw_line in iter(sub_process.stdout.readline, b""):
        line = raw_line.decode("utf-8").rstrip()
        lines.append(line)

    sub_process.wait()

print(lines)

results in:

$ python test.py
['\x1b[0m18:17:51  Running with dbt=1.1.0', '\x1b[0m18:17:52  Found 358 models, 497 tests, 4 snapshots, 0 analyses, 638 macros, 12 operations, 2 seed files, 128 sources, 55 exposures, 0 metrics', '\x1b[0m18:17:52', "\x1b[0m18:17:56  Concurrency: 4 threads (target='testing')", '\x1b[0m18:17:56', '\x1b[0m18:18:18  Done.']

since it's not a "real" shell, it won't resolve that reset color code automatically and thus it's still hanging out there when prefect/<insert tool grabbing lines from stdout> reads the lines.

[gshank]

It looks like this might be due to a programming error in core/dbt/events/functions.py. The line color_tag: str = "" if this.format_color else Style.RESET_ALL should probably be color_tag: str = "" if not this.format_color else Style.RESET_ALL (in two places).

Would you be interested in trying this and doing a pull request?

[darin-reify]

sure, i'll give it a whirl! i found that line too but hadn't dug too deep to know if that was the right spot to change.