feat: support data science scenario on macOS

[chaosddp]

Description

1. add homebrew to bin_path of LocalEnv by default
2. add virtual env to bin_path if "VIRTUAL_ENV" environment variable exist for LocalEnv in data science scenario
3. use gtimeout instead timeout on macOS. (need brew to install coreutils)
4. evaluate without strace info, as there is not a good alternative to strace on macOS, but it would be ok for local test
5. update kaggle crawler to support configurable chromedriver location, so that we do not need to move chromedrive to system folder if we already have chromedriver previously.
6. update readme about installing gtimeout on macOS for developers

States of strace alternatives on macOS:

• dtruss/dtrace is limited by SIP, only a few information is provided by default
• ktrace produce too much information that not suitable for this case

Motivation and Context

related issue: #1227

Currently for developers on macOS, we can only use devcontainers as develop environment, it not very convenient.

How Has This Been Tested?

• If you are adding a new feature, test on your own test scripts.

  I have tested with competition "titanic", and this is the script for the tests, it can generate the final submission.csv with following command on macOS Tahoe 26.0.1 (M4).

  bash run_rdagent_competition_local.sh titanic 2

  Though only tested on data science scenario, but this patch would be work for other scenarios that not running in docker, as it only try to resolve dependency issue.

Screenshots of Test Results (if appropriate):

1. Your own tests:

Types of changes

• Fix bugs
• Add new feature
• Update documentation

---

📚 Documentation preview 📚: https://RDAgent--1261.org.readthedocs.build/en/1261/