Tests fail on Windows: UnicodeDecodeError when reading SQL query templates

[saurabh12nxf]

What happens

Running uv run pytest on Windows gives 11 test failures. 10 of them are caused by a UnicodeDecodeError when reading downloads_by_country.sql:

UnicodeDecodeError: 'charmap' codec can't decode byte 0x8f in position 2391: character maps to

The remaining 1 failure is a path separator assertion (\ vs /).

Why it happens

downloads_by_country.sql has emoji characters (⚠️) in its comments (lines 42–43, 55–56). On Windows, Python defaults to cp1252 encoding, which can't handle these UTF-8 multi-byte characters.

The issue is in two places:

• pipeline.py:195 — query_file.read_text() (production code)
• queries_test.py — multiple query_file.read_text() calls (tests)

Neither specifies encoding="utf-8", so Python falls back to the system default.

The uploads_by_country.sql tests pass fine because that file has no emoji.

How to reproduce

# On Windows
uv sync --dev
uv run pytest library/tests/iqb/queries_test.py -v

Suggested fix
Add encoding="utf-8" to all read_text() calls that load SQL templates:

Before

template_text = query_file.read_text()

After

template_text = query_file.read_text(encoding="utf-8")

Environment
Windows 10
Python 3.13.12
uv 0.10.4
Note
The CI only runs on ubuntu-latest, so this never showed up before. The 1 remaining failure is
test_init_with_custom_data_dir
asserting /custom/path but getting \custom\path on Windows — a separate path separator issue I can fix in the same PR if that's okay.