rohit/probo
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Stop auto-creating org third parties on mapping

Browse Source 
The tracker-mapping worker materialized a per-org ThirdParty for every
categorized tracker, linking or creating one through heuristic and
disambiguation-agent matching. Concurrent mapping of two patterns for
the same common third party raced the load-then-create check and left
duplicate org third parties with the same name.

Reduce the worker to catalog resolution only: it resolves the shared
common_tracker_pattern_id / common_third_party_id link and leaves
third_party_id untouched, preserving any link set elsewhere. Org third
parties will instead be created through an explicit per-vendor import
action added in a later commit.

Remove resolveOrgThirdParty, prepareOrgThirdParty, the sibling
direct-link signal, and the disambiguation-agent wiring (including its
constructor parameter and buildTrackerAgents return), and update the
worker tests to assert the catalog link is resolved while third_party_id
is preserved.

Signed-off-by: Émile Ré <emile@probo.com>
This commit is contained in:
Émile Ré
2026-06-10 14:32:52 +02:00
parent 04a34c9757
commit 0f5ac03e70
4 changed files with 69 additions and 664 deletions
Download Patch File Download Diff File Expand all files Collapse all files

52
pkg/probod/tracker_agents.go
View File

@@ -22,27 +22,25 @@ import (
"go.gearno.de/kit/log"
"go.opentelemetry.io/otel/trace"
"go.probo.inc/probo/pkg/cookiebanner"
"go.probo.inc/probo/pkg/thirdparty"
)
// buildTrackerAgents wires the three tracker agents from the probod
// buildTrackerAgents wires the two tracker agents from the probod
// config, each with its own LLM client and tuning: the tracker-mapping
// agent (catalog identification), the common-pattern enrichment agent
// (description research), and the third-party disambiguation agent. All
// are opt-in: when `llm.tracker-mapping.provider` is empty it returns
// zero configs (nil LLM clients) so callers run without agent fallback.
// agent (catalog identification) and the common-pattern enrichment agent
// (description research). Both are opt-in: when
// `llm.tracker-mapping.provider` is empty it returns zero configs (nil
// LLM clients) so callers run without agent fallback.
//
// The enrichment and disambiguation agents fall back to the
// tracker-mapping config when their own provider slot is empty, so a
// deployment that configures only `tracker-mapping` keeps wiring all
// three agents.
// The enrichment agent falls back to the tracker-mapping config when its
// own provider slot is empty, so a deployment that configures only
// `tracker-mapping` keeps wiring both agents.
func (impl *Implm) buildTrackerAgents(
l *log.Logger,
tp trace.TracerProvider,
r prometheus.Registerer,
) (cookiebanner.TrackerMappingAgentConfig, cookiebanner.TrackerEnrichmentAgentConfig, thirdparty.DisambiguationAgentConfig, error) {
) (cookiebanner.TrackerMappingAgentConfig, cookiebanner.TrackerEnrichmentAgentConfig, error) {
if impl.cfg.Agents.TrackerMapping.Provider == "" {
return cookiebanner.TrackerMappingAgentConfig{}, cookiebanner.TrackerEnrichmentAgentConfig{}, thirdparty.DisambiguationAgentConfig{}, nil
return cookiebanner.TrackerMappingAgentConfig{}, cookiebanner.TrackerEnrichmentAgentConfig{}, nil
}
firecrawlAPIKey := impl.cfg.Agents.Tools.FirecrawlAPIKey
@@ -55,7 +53,7 @@ func (impl *Implm) buildTrackerAgents(
r,
)
if err != nil {
return cookiebanner.TrackerMappingAgentConfig{}, cookiebanner.TrackerEnrichmentAgentConfig{}, thirdparty.DisambiguationAgentConfig{}, fmt.Errorf("cannot resolve tracker mapping agent client: %w", err)
return cookiebanner.TrackerMappingAgentConfig{}, cookiebanner.TrackerEnrichmentAgentConfig{}, fmt.Errorf("cannot resolve tracker mapping agent client: %w", err)
}
mappingCfg := cookiebanner.TrackerMappingAgentConfig{
@@ -81,7 +79,7 @@ func (impl *Implm) buildTrackerAgents(
r,
)
if err != nil {
return cookiebanner.TrackerMappingAgentConfig{}, cookiebanner.TrackerEnrichmentAgentConfig{}, thirdparty.DisambiguationAgentConfig{}, fmt.Errorf("cannot resolve tracker enrichment agent client: %w", err)
return cookiebanner.TrackerMappingAgentConfig{}, cookiebanner.TrackerEnrichmentAgentConfig{}, fmt.Errorf("cannot resolve tracker enrichment agent client: %w", err)
}
enrichmentCfg := cookiebanner.TrackerEnrichmentAgentConfig{
@@ -94,29 +92,5 @@ func (impl *Implm) buildTrackerAgents(
MaxTurns: impl.cfg.CommonPatternEnrichmentWorker.AgentMaxTurns,
}
disambiguationSlot := impl.cfg.Agents.ThirdPartyDisambiguation
if disambiguationSlot.Provider == "" {
disambiguationSlot = impl.cfg.Agents.TrackerMapping
}
disambiguationAgentCfg, disambiguationClient, err := impl.resolveAgentClient(
"third-party-disambiguation",
disambiguationSlot,
l,
tp,
r,
)
if err != nil {
return cookiebanner.TrackerMappingAgentConfig{}, cookiebanner.TrackerEnrichmentAgentConfig{}, thirdparty.DisambiguationAgentConfig{}, fmt.Errorf("cannot resolve third party disambiguation agent client: %w", err)
}
disambiguationCfg := thirdparty.DisambiguationAgentConfig{
LLMClient: disambiguationClient,
Model: disambiguationAgentCfg.ModelName,
MaxTokens: disambiguationAgentCfg.MaxTokens,
Temperature: disambiguationAgentCfg.Temperature,
Timeout: time.Duration(impl.cfg.TrackerMappingWorker.DisambiguationAgentTimeout) * time.Second,
}
return mappingCfg, enrichmentCfg, disambiguationCfg, nil
return mappingCfg, enrichmentCfg, nil
}