Improved Environment Variable Handling for Enhanced Flexibility in TensorBoard Launch

[lcolok]

Dear Project Maintainer,

I would like to extend my sincerest gratitude and respect for your diligent work on this project. It has immensely contributed to the efficiency of my workflow, particularly in monitoring and managing the training of machine learning models.

During routine usage, I have identified an opportunity to enhance the startup process of TensorBoard, which is particularly critical in multi-user server environments or containerized deployments. I have proposed an improvement to optimize the logic for reading the TENSORBOARD_PORT environment variable.

The specific improvements are as follows:

• Before launching TensorBoard, the program will check whether the TENSORBOARD_PORT environment variable is set. If it is, the program will use this value as the port for TensorBoard; otherwise, it will default to port 6006. This modification allows users to flexibly configure the port in different environments to avoid port conflicts.
• By utilizing an environment variable, we improve the program's flexibility and configurability, enabling it to adapt to the actual conditions when running in servers or containers, among other environments.
• This approach avoids hard-coding the port number, aiding in better port configuration management during automated deployments and continuous integration processes.

I am confident that this improvement will make the tool more robust and easily configurable, thus serving a wider range of use scenarios. I hope you will consider adopting this modification. Should there be any concerns or further clarifications needed, please feel free to contact me.

Thank you for your contributions to this project and the open-source community at large. I look forward to your response.

Warm regards

[bmaltais]

Thank you for the contribution. I will merge it in dev 2 and release it this weekend