414749681f
Addresses three review concerns from the prior PR:
1. No tests existed for any of the stuck-parsing fixes (PR① / PR②.5).
This commit adds coverage for the four most regression-prone
surfaces: span repo upsert/cascade/attempt isolation, span
tracker cascade-cancel and cross-process LookupStage,
housekeeping false-kill protection, and handler tree assembly.
2. Housekeeping was using only knowledge.updated_at as its staleness
signal, but knowledge.updated_at advances only at parse_status
transitions — a long DocReader call (or large embedding batch)
can run for an hour with no updated_at change, so a tight
DocumentProcessTimeout setting would falsely flip an actively
running parse to "failed".
The sweep now does a two-stage check: candidates by knowledge
updated_at, then filtered by MAX(spans.updated_at). Every
SpanTracker Begin/End/Fail/Skip now also pokes
knowledge.updated_at as a side-channel heartbeat, so the
filter sees recent activity even when no parse_status
transition has fired.
3. parseHeartbeatTime accepts the timestamp formats both Postgres
and SQLite emit for an aggregated MAX() column (the SQLite
driver doesn't auto-cast aggregates to time.Time the way
Postgres does), so the same code path works in Lite mode.
The new TestHousekeeping_NoFalseKill_ActiveSpan is the regression
test for the user-flagged scenario: a 3-hour-stale knowledge.updated_at
combined with a 2-minute-fresh span row must NOT be killed.
101 lines
4.6 KiB
Go
101 lines
4.6 KiB
Go
package handler
|
|
|
|
import (
|
|
"testing"
|
|
"time"
|
|
|
|
"github.com/Tencent/WeKnora/internal/types"
|
|
"github.com/stretchr/testify/assert"
|
|
)
|
|
|
|
// TestBuildSpanTree_AssemblesParentChild covers the basic shape: a root
|
|
// with stage children and an image generation grandchild. The handler's
|
|
// JSON contract depends on this tree topology; if the linking logic
|
|
// regresses, the UI sees a flat list and renders nothing.
|
|
func TestBuildSpanTree_AssemblesParentChild(t *testing.T) {
|
|
now := time.Now()
|
|
rows := []types.KnowledgeProcessingSpan{
|
|
{KnowledgeID: "kid", Attempt: 1, SpanID: "root", Name: "knowledge_processing", Kind: types.SpanKindRoot, Status: types.SpanStatusRunning, StartedAt: &now},
|
|
{KnowledgeID: "kid", Attempt: 1, SpanID: "doc", ParentSpanID: "root", Name: types.StageDocReader, Kind: types.SpanKindStage, Status: types.SpanStatusDone, StartedAt: &now},
|
|
{KnowledgeID: "kid", Attempt: 1, SpanID: "mm", ParentSpanID: "root", Name: types.StageMultimodal, Kind: types.SpanKindStage, Status: types.SpanStatusRunning, StartedAt: &now},
|
|
{KnowledgeID: "kid", Attempt: 1, SpanID: "img0", ParentSpanID: "mm", Name: "multimodal.image[0]", Kind: types.SpanKindGeneration, Status: types.SpanStatusRunning, StartedAt: &now},
|
|
}
|
|
|
|
tree, currentStage, lastFail := buildSpanTree("kid", 1, rows)
|
|
require := assert.New(t)
|
|
require.NotNil(tree)
|
|
require.Equal("root", tree.SpanID)
|
|
require.Equal(types.StageMultimodal, currentStage,
|
|
"running stage span surfaces as current_stage")
|
|
require.Nil(lastFail)
|
|
|
|
// The 5 canonical stages must all appear under root (real or
|
|
// synthesized placeholder).
|
|
stageNames := map[string]string{}
|
|
for _, child := range tree.Children {
|
|
if child.Kind == types.SpanKindStage {
|
|
stageNames[child.Name] = child.Status
|
|
}
|
|
}
|
|
for _, name := range types.AllStages {
|
|
_, ok := stageNames[name]
|
|
require.True(ok, "stage %s must appear in tree", name)
|
|
}
|
|
require.Equal(types.SpanStatusDone, stageNames[types.StageDocReader])
|
|
require.Equal(types.SpanStatusRunning, stageNames[types.StageMultimodal])
|
|
require.Equal(types.SpanStatusPending, stageNames[types.StageEmbedding],
|
|
"missing stage rows must synthesize as pending placeholders")
|
|
|
|
// The image subspan must hang under multimodal, not at root level.
|
|
var mmNode *types.SpanTreeNode
|
|
for _, c := range tree.Children {
|
|
if c.Name == types.StageMultimodal {
|
|
mmNode = c
|
|
break
|
|
}
|
|
}
|
|
require.NotNil(mmNode)
|
|
require.Len(mmNode.Children, 1, "image generation subspan must be a child of multimodal stage")
|
|
require.Equal("multimodal.image[0]", mmNode.Children[0].Name)
|
|
}
|
|
|
|
// TestBuildSpanTree_NoRows_SynthesizesPlaceholderRoot ensures the API
|
|
// keeps a stable shape for fresh / never-parsed knowledge: the frontend
|
|
// always gets a `trace` with five pending stage children, never a
|
|
// nil/empty response.
|
|
func TestBuildSpanTree_NoRows_SynthesizesPlaceholderRoot(t *testing.T) {
|
|
tree, currentStage, lastFail := buildSpanTree("kid-empty", 0, nil)
|
|
a := assert.New(t)
|
|
a.NotNil(tree)
|
|
a.Equal(types.SpanKindRoot, tree.Kind)
|
|
a.Equal(types.SpanStatusPending, tree.Status)
|
|
a.Equal("", currentStage, "no rows means no running stage")
|
|
a.Nil(lastFail)
|
|
// All 5 stages must be present as pending placeholders so the UI
|
|
// renders a complete timeline even pre-parse.
|
|
a.Len(tree.Children, len(types.AllStages))
|
|
for _, child := range tree.Children {
|
|
a.Equal(types.SpanStatusPending, child.Status)
|
|
a.Equal(types.SpanKindStage, child.Kind)
|
|
}
|
|
}
|
|
|
|
// TestBuildSpanTree_LastFailureSurfaces records that a failed stage is
|
|
// reported as last_error so the UI can highlight the responsible step
|
|
// even if a later stage was cancelled by cascade.
|
|
func TestBuildSpanTree_LastFailureSurfaces(t *testing.T) {
|
|
now := time.Now()
|
|
finished := now.Add(5 * time.Second)
|
|
rows := []types.KnowledgeProcessingSpan{
|
|
{KnowledgeID: "kid", Attempt: 1, SpanID: "root", Name: "knowledge_processing", Kind: types.SpanKindRoot, Status: types.SpanStatusFailed, StartedAt: &now, FinishedAt: &finished},
|
|
{KnowledgeID: "kid", Attempt: 1, SpanID: "doc", ParentSpanID: "root", Name: types.StageDocReader, Kind: types.SpanKindStage, Status: types.SpanStatusFailed, ErrorCode: "DOCREADER_TIMEOUT", ErrorMessage: "slow", StartedAt: &now, FinishedAt: &finished},
|
|
{KnowledgeID: "kid", Attempt: 1, SpanID: "chunk", ParentSpanID: "root", Name: types.StageChunking, Kind: types.SpanKindStage, Status: types.SpanStatusCancelled, ErrorCode: "UPSTREAM_FAILED"},
|
|
}
|
|
_, _, lastFail := buildSpanTree("kid", 1, rows)
|
|
a := assert.New(t)
|
|
a.NotNil(lastFail)
|
|
a.Equal(types.StageDocReader, lastFail.Name,
|
|
"last_error must point at the actually-failed span, not the cascade-cancelled downstream")
|
|
a.Equal("DOCREADER_TIMEOUT", lastFail.ErrorCode)
|
|
}
|