- Renamed fields in LLMResponseUsage struct to better reflect their purpose, changing InputTokens and OutputTokens to PromptTokens and CompletionTokens.
- Modified GetUsage method in ChatMessage to accept an optional contents parameter, improving flexibility in usage tracking.
- Ensured consistent field naming and improved clarity in token usage aggregation logic.
- Added BSON tags to fields in LLMResponseUsage struct to ensure proper serialization with MongoDB.
- Updated the Usage field in ChatMessage model to exclude it from JSON serialization while retaining BSON compatibility.
- Introduced GetUsage method in ChatMessage to aggregate token usage from contents, enhancing usage tracking capabilities.
- Added LLMResponseUsage struct to encapsulate token usage details for language model responses.
- Updated ChatMessageContent and ChatMessage models to include a Usage field of type *entity.LLMResponseUsage, enhancing message handling capabilities.
- Improved structure and clarity of models by ensuring consistent field definitions and descriptions.
