improved
Scoring API: raw_filters + score_threshold
16 days ago by Nedhir Ebnou
We’ve just upgraded the Scoring API with two new powerful parameters: raw_filters and score_threshold.
These additions give users much finer control over how profiles or jobs are scored, filtered, and ranked—bringing Elasticsearch Query DSL capabilities and threshold-based precision directly into the scoring endpoints.
😍 Why it’s a big deal for HrFlow.ai users?
This update unlocks advanced customization for scoring pipelines:
- ✅ raw_filters
Allows users to write complex, expressive filtering logic using Elasticsearch’s native Query DSL syntax—just like in the Searching API.
You can now filter profiles or jobs before scoring, based on structured fields or full-text conditions. - ✅ score_threshold
Lets you exclude results that fall below a minimum relevance score. This helps:- Remove noise from results.
- Customize scoring selectivity.
- Fine-tune matching quality.
🧪 Example Queries
// Job filtering before scoring
{
"query": {
"bool": {
"should": [
{ "match_phrase": { "name": "AI" } },
{ "match_phrase": { "name": "Researcher" } }
]
}
}
}
Profile filtering before scoring
{
"query": {
"bool": {
"should": [
{ "match_phrase": { "text": "Python" } },
{ "match_phrase": { "text": "Java" } }
]
}
}
}