feat(models): add interventions and eligibility_keywords to SearchAnchors

Support richer ClinicalTrials.gov queries by adding target drug names
(based on biomarker-to-drug mapping) and patient clinical features for
eligibility pre-filtering. Both default to empty lists for backward
compatibility.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

trialpath/models/search_anchors.py

@@ -26,6 +26,14 @@ class SearchAnchors(BaseModel):
     age: Optional[int] = None
     performance_status_max: Optional[int] = None
     trial_filters: TrialFilters = Field(default_factory=TrialFilters)
+    interventions: list[str] = Field(
+        default_factory=list,
+        description="Target drugs/treatments based on patient biomarkers and treatment history",
+    )
+    eligibility_keywords: list[str] = Field(
+        default_factory=list,
+        description="Key patient clinical features for eligibility pre-filtering",
+    )
     relaxation_order: list[str] = Field(
         default=["phase", "distance", "biomarker_strictness"],
         description="Order in which to relax criteria if too few results",