Escape non-ASCII characters in tojson like Python's json.dumps - #16908

Resolves #16830

Problem

tojson writes non-ASCII characters as raw UTF-8, but dbt 1.x uses json.dumps with its default ensure_ascii=True, which writes \uXXXX. So local_md5(tojson(...)) gives a different hash on v2 for anything with non-ASCII text, which is what makes Elementary re-insert its artifacts every time you switch engines.

Solution

Added write_string_fragment to PythonStyleFormatter so it escapes everything outside printable ASCII the way Python does: lowercase \uXXXX, surrogate pairs above U+FFFF, and DEL (0x7f) too, since Python escapes that one as well. Same idea as PythonJsonFormatter in dbt-state. Added a test with あ, é, an emoji, U+2028 and DEL, plus the local_md5 hash from the issue (f3dd52a0cdda6cf470609dcfc4759191), all checked against json.dumps in Python.

Checklist

• I have read the contributing guide and understand what's expected of me.
• I have run this code in development, and it appears to resolve the stated issue.
• This PR includes tests, or tests are not required or relevant for this PR.
• This PR has no interface changes ... (tojson output changes for non-ASCII input, but only back to what 1.x does)

This fix was AI-assisted and reviewed by me. Thanks