🚀 Feature: Move custom histogram buckets to `opentelemetry-semantic-conventions-ai`

Problem

@nirga Currently, optimized histogram for specific LLM meters are only defined in traceloop-sdk, while standalone instrumentation packages will use OpenTelemetry's default buckets.

Taking the token usage histogram as an example, when integrated with traceloop-sdk, the instrumentation package utilizes the specified histogram, but upon standalone distribution as opentelemetry-instrumentation-openai, without defined histogram, the package falls back to using the OTel default histogram.

Proposed Implementation

We should move these meters definitions from traceloop-sdk to opentelemetry-semantic-conventions-ai just like opentelemetry-python, enabling the standalone distributed package to utilize them.

✌️ How do you aim to achieve this?

As mentioned above.

🔄️ Additional Information

No response

👀 Have you spent some time to check if this feature request has been raised before?

• I checked and didn't find similar issue

Are you willing to submit PR?

Yes I am willing to submit a PR!

Start by reading packages/traceloop-sdk/traceloop/sdk/metrics/metrics.py and locating the custom histogram definitions, then inspect the opentelemetry-semantic-conventions-ai package structure and its tests. Move the definitions to that package and verify that standalone instrumentation packages use the same buckets without relying on traceloop-sdk.