litellm Added support to additional MIME types for Gemini model

Title

Add support for additional MIME types in Gemini models

Fixes #9416

Please complete all items before asking a LiteLLM maintainer to review your PR

[x] I have Added testing in the [tests/litellm/](https://github.com/BerriAI/litellm/tree/main/tests/litellm) directory, Adding at least 1 test is a hard requirement - [see details](https://docs.litellm.ai/docs/extras/contributing_code)
[x] I have added a screenshot of my new test passing locally
[ ] My PR passes all unit tests on (make test-unit)[https://docs.litellm.ai/docs/extras/contributing_code]
[x] My PR's scope is as isolated as possible, it only solves 1 specific problem

🆕 New Feature

This PR adds support for non-image/PDF file types with Gemini models, specifically addressing:

Fixed handling of text-based formats (Markdown, JSON, XML, CSV) that don't require base64 encoding
Resolved token counting errors for text files by preventing unnecessary encoding
Expanded MIME type detection to support additional file formats
Maintains backward compatibility with existing image/PDF handling

litellm/types/files.py:
- Added requires_base64_encoding() function
- Added new file types (Markdown, XML) to supported types
- Updated GEMINI_1_5_ACCEPTED_FILE_TYPES to include text-based formats
litellm/llms/vertex_ai/gemini/transformation.py:
- Renamed _process_gemini_image() to _process_gemini_file()
- Added conditional logic to handle text files without base64 encoding
- Expanded _get_mime_type_from_url() to detect additional formats
- Fixed typo in error message ("filr" → "file")
- Updated function references

tests/litellm/llms/vertex_ai/gemini/test_transformation.py:
- Added tests for MIME type detection
- Added tests for text vs binary file handling
- Added tests for file type validation