feat(zhipuai): ZhipuAI add thinking and response_format parameter support

[YunKuiLu]

• Add thinking and response_format fields to ZhiPuAiApi and ZhiPuAiChatOptions
• Add ZhiPuAiChatOptionsTests with 16 test methods covering all aspects of the class
• Test builder pattern with all fields including responseFormat and thinking
• Test copy functionality, setters, default values, and equals/hashCode
• Test tool callbacks, tool names validation, and collection handling
• Test stop sequences alias and fluent setters
• Add documentation for response-format.type and thinking.type properties

Official API:
https://docs.bigmodel.cn/api-reference/%E6%A8%A1%E5%9E%8B-api/%E5%AF%B9%E8%AF%9D%E8%A1%A5%E5%85%A8#body-response-format

[ilayaperumalg]

@mxsl-gr Requesting your review for this PR as well. Thank you

[mxsl-gr]

looks good.
not related to this PR's changes, with GLM-4.5 family models, the default behavior is to have thinking mode enabled.
while it does improve the output quality, I still feel it’s not ideal since it noticeably slows down the response speed

[YunKuiLu]

looks good.
not related to this PR's changes, with GLM-4.5 family models, the default behavior is to have thinking mode enabled. while it does improve the output quality, I still feel it’s not ideal since it noticeably slows down the response speed

@mxsl-gr Thank you for your review.

If not explicitly set, this PR will not set the thinking.type parameter. And GLM-4.5 enables CoT by default when thinking.type is missing.
That's why I marked enabled as default in the comment.

[ilayaperumalg]

Wouldn't the default value be null if it's not explicitly set?

[YunKuiLu]

@ilayaperumalg Yes, the thinking in the request body will be null, but for models GLM-4.5 and above, setting thinking=null is equivalent to enabling chain-of-thought reasoning. Here's the English documentation: https://docs.z.ai/api-reference/llm/chat-completion#body-thinking-type

[ilayaperumalg]

@YunKuiLu Thanks for the clarification!

[ilayaperumalg]

@mxsl-gr Thanks for the review. @YunKuiLu Thanks for the PR and following it up on the comments. Rebased and merged as cae6364