[CI] tests should be run in latest and dynamically created conda env

## 🔨Work Item

**IMPORTANT:**
* This template is only for dev team to track project progress. For feature request or bug report, please use the corresponding issue templates.
* DO NOT create a new work item if the purpose is to fix an existing issue or feature request. We will directly use the issue in the project tracker.

Project tracker: https://github.com/orgs/dmlc/projects/2

## Description
For now, CI tests are running on a fixed conda env such as `pytorch-ci` which might be out-dated. And if we need to add new package dependencies when supporting new feature, we need to manually update the conda env in docker image(Linux) or dedicated machine(Windows) in advance. This is really inconvenient.

The better way is to create conda env in runtime via a template such as https://github.com/dmlc/dgl/blob/master/script/dgl_dev.yml.template. Then any new dependency should be added into the template.

## Depending work items or issues
An example is https://github.com/dmlc/dgl/pull/5942.