> ## Documentation Index
> Fetch the complete documentation index at: https://docs.honeyhive.ai/llms.txt
> Use this file to discover all available pages before exploring further.
# Client-Side Evaluations
> Learn how to log external evaluation results (metrics) with your trace.
export const enrichmentType_4 = "Metrics"
export const paramName_3 = "metrics"
export const enrichmentType_3 = "Metrics"
export const paramName_2 = "metrics"
export const enrichmentType_2 = "Metrics"
export const paramName_1 = "metrics"
export const enrichmentType_1 = "Metrics"
export const paramName_0 = "metrics"
export const enrichmentType_0 = "Metrics"
## Introduction
Client-side evaluations allow you to log external evaluation results (metrics) with your trace.
### Prerequisites
You have already set tracing for your code as [described in our quickstart guide](/introduction/quickstart).
## Setting {enrichmentType_0}
You can set {enrichmentType_0} on both the trace level or the span level. If the {enrichmentType_0} applies to the entire trace, then set it on the trace level. If the {enrichmentType_0} applies to a specific span, then set it on the span level. For more details, refer to the [enrich traces](/tracing/enrich-traces) documentation.
In Python, you can use the `enrich_session` function to set {enrichmentType_1} on the trace level.
To pass {enrichmentType_1} to HoneyHive, pass it to the {paramName_0} param in the `enrich_session` function. This function is used to enrich the session with additional information. Remember that `enrich_session` will update, not overwrite, the existing {paramName_0} object on the trace.
Read more about the `enrich_session` function in the [Python SDK reference](/sdk-reference/python-tracer-ref#enrich-session).
Here's an example of how to set {enrichmentType_1} on the trace level in Python:
```python Python theme={null}
from honeyhive import HoneyHiveTracer, enrich_session
HoneyHiveTracer.init(
api_key="my-api-key",
project="my-project",
)
# ...
enrich_session(metrics={
"json_validated": True,
"num_actions": 10,
# any other custom fields and values as you need
"step_evals": [
{
"invalid_grammar": False,
"unable_to_locate_UI": True
}
],
})
```
In Python, you can use the `enrich_span` function to set {enrichmentType_3} on the span level.
To pass {enrichmentType_3} to HoneyHive, pass it to the {paramName_2} param in the `enrich_span` function. This function is used to enrich the span with additional information. Remember that `enrich_span` will update, not overwrite, the existing {paramName_2} object linked to the span.
Read more about the `enrich_span` function in the [Python SDK reference](/sdk-reference/python-tracer-ref#enrich-span).
Here's an example of how to set {enrichmentType_3} on the span level in Python:
```python Python theme={null}
from honeyhive import HoneyHiveTracer, trace, enrich_span
HoneyHiveTracer.init(
api_key="my-api-key",
project="my-project",
)
# ...
@trace
def my_function(input, offline_eval_score):
# ...
enrich_span(metrics={
"json_validated": True,
"num_actions": 10,
"offline_eval_score": offline_eval_score,
# any other custom fields and values as you need
"step_evals": [
{
"invalid_grammar": False,
"unable_to_locate_UI": True
}
],
})
# ...
return response
# ...
```
In TypeScript, you can use the `tracer.enrichSession` function to set {enrichmentType_2} on the trace level.
To pass {enrichmentType_2} to HoneyHive, pass it to the {paramName_1} param in the `tracer.enrichSession` function. This function is used to enrich the session with additional information. Remember that `tracer.enrichSession` will update, not overwrite, the existing {paramName_1} object linked to the trace.
Read more about the `tracer.enrichSession` function in the [TypeScript SDK reference](/sdk-reference/typescript-tracer-ref#enrichsession).
Here's an example of how to set {enrichmentType_2} on the trace level in TypeScript:
```typescript TypeScript theme={null}
import { HoneyHiveTracer } from "honeyhive";
const tracer = await HoneyHiveTracer.init({
apiKey: "my-api-key",
project: "my-project",
});
// ...
tracer.enrichSession({
metrics: {
"json_validated": true,
"num_actions": 10,
// any other custom fields and values as you need
"step_evals": [
{
"invalid_grammar": false,
"unable_to_locate_UI": true
}
],
}
})
```
In TypeScript, you can use the `tracer.enrichSpan` function to set {enrichmentType_4} on the span level.
To pass {enrichmentType_4} to HoneyHive, pass it to the {paramName_3} param in the `tracer.enrichSpan` function. This function is used to enrich the span with additional information. Remember that `tracer.enrichSpan` will update, not overwrite, the existing {paramName_3} object linked to the span.
Read more about the `tracer.enrichSpan` function in the [TypeScript SDK reference](/sdk-reference/typescript-tracer-ref#enrichspan).
Here's an example of how to set {enrichmentType_4} on the span level in TypeScript:
```typescript TypeScript theme={null}
import { HoneyHiveTracer } from "honeyhive";
const tracer = await HoneyHiveTracer.init({
apiKey: "my-api-key",
project: "my-project",
});
// ...
const myTracedFunction = tracer.traceFunction()(function (param1, param2) {
// span enrichment
tracer.enrichSpan({
metrics: {
"json_validated": true,
"num_actions": 10,
// any other custom fields and values as you need
"step_evals": [
{
"invalid_grammar": false,
"unable_to_locate_UI": true
}
],
}
});
// Your function code here
return `Result with ${param1} and ${param2}`;
});
// ...
```
## Concepts
If you are defining any evaluators in your application execution i.e. client-side, you can add the resulting metrics to any trace or span in HoneyHive for monitoring and analysis.
Most guardrails (eg: format, safety, PII, etc.) are ideal to compute client-side at execution time instead of server-side in HoneyHive post-ingestion.
### What are evaluators?
Evaluators are tests (code or LLM based) that compute a score/metric to measure the quality of inputs and/or outputs for your application or specific steps within it.
They can be computed and instrumented within your application runtime (client-side) or computed in our platform (server-side) post-ingestion.
Evaluators computed client-side are not overwritten by server-side evaluators of the same name.
## Return Types
We accept all primary data types as evaluator metrics. This includes:
| Return Type | Available Measurements | Notes | Uses |
| ----------- | ----------------------------------------- | ----------------------------- | ---------------------------------------- |
| Boolean | Percentage True/False | - | Evaluations |
| Number | Sum, Avg, Median, Min, Max, P95, P98, P99 | - | Evaluations |
| String | - | Used for filters and group by | Classification, feature extraction, etc. |
For complex data types, we allow you to drill down to the nested fields or specific positions in the array.
So, for example, if you pass metrics like:
```json theme={null}
{
"step_evals": [
{
"invalid_grammar": true,
"user_interevened": true
},
{
"invalid_grammar": false,
"unable_to_locate_UI": true
}
],
"trajectory_eval": {
"overall": 5,
"clarified_user_intent": "yes"
}
}
```
You can chart `metrics.step_evals.0.user_intervened` as a boolean field or `trajectory_eval.overall` as a numeric field.
**Nesting Limitations:** For complex data types like objects and arrays, we allow max 5 levels of nested objects and max 2 levels of nested arrays.
## Learn more
Learn how to chart metrics from your traces
Learn how to compute metrics server-side over your logs via code
Learn how to compute metrics server-side over your logs via a LLM
Learn how to setup annotation for your logs via domain experts
## SDK Reference
Read more about the `enrich_session` function in the [Python SDK reference](/sdk-reference/python-tracer-ref#enrich-session).