fix(files): fall back to text/plain for unsupported text mime types (…

[sk-reddy17]

Summary

Fixes #744.

When you upload a file without a mime_type, the SDK uses mimetypes.guess_type() and sends the result as-is. For lots of text/code files the guess is something the API rejects, so the upload 400s
even though it's plain text — e.g. .cu → application/cu-seeme, .ts → text/vnd.trolltech.linguist, .py → text/x-python.

client.files.upload(file="example.cu")

-> 400 INVALID_ARGUMENT: Unsupported MIME type: application/cu-seeme

Change

Added _resolve_upload_mime_type() in google/genai/_extra_utils.py, used only when the type is auto-detected (an explicit mime_type still wins). If the guess is one the API rejects but the file is
actually UTF-8 text, it falls back to text/plain. Standard text/*, image/audio/video, and application/json|xml|rtf are kept as-is; binary files are never relabeled (a file counts as text only if its
first 8 KB have no NUL bytes and decode as UTF-8).

Tests

Added google/genai/tests/files/test_upload_mime_type.py — 10 cases, no API key needed (fallbacks, type preservation, binary safety, upload header, explicit-mime_type precedence).

$ python -m pytest google/genai/tests/files/test_upload_mime_type.py -q
10 passed

Open question

Which application/* types should be kept vs. downgraded? Currently keeping json, xml, rtf — happy to adjust.

[google-cla]

Thanks for your pull request! It looks like this may be your first contribution to a Google open source project. Before we can look at your pull request, you'll need to sign a Contributor License Agreement (CLA).

View this failed invocation of the CLA check for more information.

For the most up to date status, view the checks section at the bottom of the pull request.