eric/seriatim
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases 4 Wiki Activity

Added initial segment overlap resolution logic

Browse Source
This commit is contained in:
Eric Rakestraw
2026-04-27 15:52:53 -05:00
parent e42a2326e8
commit 1b9f4bd922
16 changed files with 1357 additions and 59 deletions
Download Patch File Download Diff File Expand all files Collapse all files

51
internal/model/model.go
View File

@@ -11,6 +11,7 @@ type RawSegment struct {
Start float64 `json:"start"`
End float64 `json:"end"`
Text string `json:"text"`
Words []Word `json:"words,omitempty"`
}
// CanonicalTranscript is a per-speaker transcript in seriatim's internal model.
@@ -45,16 +46,18 @@ type OutputMetadata struct {
// Segment is the canonical transcript segment shape used by the framework.
type Segment struct {
ID int `json:"id,omitempty"`
InternalRef string `json:"internal_ref,omitempty"`
Source string `json:"source"`
SourceSegmentIndex int `json:"source_segment_index"`
Speaker string `json:"speaker"`
Start float64 `json:"start"`
End float64 `json:"end"`
Text string `json:"text"`
Words []Word `json:"words,omitempty"`
OverlapGroupID int `json:"overlap_group_id,omitempty"`
ID int `json:"id,omitempty"`
InternalRef string `json:"internal_ref,omitempty"`
Source string `json:"source"`
SourceSegmentIndex *int `json:"source_segment_index,omitempty"`
SourceRef string `json:"source_ref,omitempty"`
DerivedFrom []string `json:"derived_from,omitempty"`
Speaker string `json:"speaker"`
Start float64 `json:"start"`
End float64 `json:"end"`
Text string `json:"text"`
Words []Word `json:"words,omitempty"`
OverlapGroupID int `json:"overlap_group_id,omitempty"`
}
// Word preserves optional word-level timing data.
@@ -64,6 +67,7 @@ type Word struct {
End float64 `json:"end"`
Score float64 `json:"score,omitempty"`
Speaker string `json:"speaker,omitempty"`
Timed bool `json:"-"`
}
// OverlapGroup describes a detected overlapping speech region.
@@ -76,3 +80,30 @@ type OverlapGroup struct {
Class string `json:"class"`
Resolution string `json:"resolution"`
}
// SegmentLess defines the deterministic chronological ordering used by merge
// and postprocessing modules.
func SegmentLess(left Segment, right Segment) bool {
if left.Start != right.Start {
return left.Start < right.Start
}
if left.End != right.End {
return left.End < right.End
}
if left.Source != right.Source {
return left.Source < right.Source
}
if left.SourceSegmentIndex != nil && right.SourceSegmentIndex != nil && *left.SourceSegmentIndex != *right.SourceSegmentIndex {
return *left.SourceSegmentIndex < *right.SourceSegmentIndex
}
if left.SourceSegmentIndex == nil && right.SourceSegmentIndex != nil {
return false
}
if left.SourceSegmentIndex != nil && right.SourceSegmentIndex == nil {
return true
}
if left.SourceRef != right.SourceRef {
return left.SourceRef < right.SourceRef
}
return left.Speaker < right.Speaker
}