migrate to evaldata

src/poprox_recommender/evaluation/generate.py

@@ -37,7 +37,7 @@
 from poprox_concepts.api.recommendations import RecommendationRequest
 from poprox_concepts.domain import ArticleSet
 from poprox_recommender.config import default_device
-from poprox_recommender.data.data import Data
+from poprox_recommender.data.eval import EvalData
 from poprox_recommender.data.mind import TEST_REC_COUNT, MindData
 from poprox_recommender.data.poprox import PoproxData
 from poprox_recommender.lkpipeline import PipelineState
@@ -108,7 +108,7 @@ def extract_recs(
     return output_df
 
 
-def generate_user_recs(data: Data, pipe_names: list[str] | None = None, n_users: int | None = None):
+def generate_user_recs(data: EvalData, pipe_names: list[str] | None = None, n_users: int | None = None):
     pipelines = recommendation_pipelines(device=default_device())
     if pipe_names is not None:
         pipelines = {name: pipelines[name] for name in pipe_names}  # type: ignore
@@ -118,9 +118,9 @@ def generate_user_recs(data: Data, pipe_names: list[str] | None = None, n_users:
     logger.info("generating recommendations")
     user_recs = []
 
-    user_iter = data.iter_users()
+    user_iter = data.iter_profiles()
     if n_users is None:
-        n_users = data.n_users
+        n_users = data.n_profiles
         logger.info("recommending for all %d users", n_users)
     else:
         logger.info("running on subset of %d users", n_users)