From 8b5b2f659c2bf848918648f8564dcb832d79f57e Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Th=C3=A9o?= <theo.nardin@cri-paris.org>
Date: Thu, 5 Mar 2026 12:04:33 +0100
Subject: [PATCH 01/14] Remove unused cache dict

---
 src/app/services/sql_db/queries.py | 7 +------
 1 file changed, 1 insertion(+), 6 deletions(-)

diff --git a/src/app/services/sql_db/queries.py b/src/app/services/sql_db/queries.py
index 170f7b4..031709b 100644
--- a/src/app/services/sql_db/queries.py
+++ b/src/app/services/sql_db/queries.py
@@ -34,7 +34,6 @@
 from src.app.services.constants import APP_NAME
 from src.app.services.sql_db.sql_service import session_maker
 
-model_id_cache: dict[str, UUID] = {}
 model_id_lock = Lock()
 
 
@@ -215,7 +214,7 @@ def get_context_documents(
     return sdg_meta_documents
 
 
-def get_embeddings_model_id_according_name(model_name: str) -> UUID | None:
+def get_embeddings_model_id_according_name(model_name: str) -> list[UUID] | None:
     """
     Get the embeddings model ID according to its name.
 
@@ -225,10 +224,6 @@ def get_embeddings_model_id_according_name(model_name: str) -> UUID | None:
     Returns:
         The ID of the embeddings model if found, otherwise None.
     """
-    with model_id_lock:
-        if model_name in model_id_cache:
-            return model_id_cache[model_name]
-
     with session_maker() as session:
         model = (
             session.query(EmbeddingModel)

From 27279659e16b802ba9263307ad70d879e3de59d6 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Th=C3=A9o?= <theo.nardin@cri-paris.org>
Date: Thu, 5 Mar 2026 12:20:55 +0100
Subject: [PATCH 02/14] feat(micro-learning): update embedding model handling
 to support multiple models

---
 src/app/api/api_v1/endpoints/micro_learning.py |  7 +++++--
 src/app/services/helpers.py                    |  9 +++++----
 src/app/services/sql_db/queries.py             | 13 +++++++------
 3 files changed, 17 insertions(+), 12 deletions(-)

diff --git a/src/app/api/api_v1/endpoints/micro_learning.py b/src/app/api/api_v1/endpoints/micro_learning.py
index 460fcc8..b54a7c3 100644
--- a/src/app/api/api_v1/endpoints/micro_learning.py
+++ b/src/app/api/api_v1/endpoints/micro_learning.py
@@ -30,10 +30,13 @@
 async def get_subject_list(
     lang: str | None = None, sp: SearchService = Depends(get_search_service)
 ) -> list[str]:
-    collection_info, model_id = await collection_and_model_id_according_lang(
+    collection_info, models_ids = await collection_and_model_id_according_lang(
         sp=sp, lang=lang
     )
-    ret = [md.title for md in get_subjects(embedding_model_id=model_id)]
+    ret = [
+        md.title
+        for md in get_subjects(embedding_models_ids=[model.id for model in models_ids])
+    ]
     if len(ret) == 0:
         raise HTTPException(status_code=404, detail="No subjects found.")
     return ret
diff --git a/src/app/services/helpers.py b/src/app/services/helpers.py
index a4cee7c..d5374a8 100644
--- a/src/app/services/helpers.py
+++ b/src/app/services/helpers.py
@@ -8,6 +8,7 @@
 from json_repair import JSONReturnType
 from langdetect import detect_langs
 from qdrant_client.http.models import models
+from welearn_database.data.models import EmbeddingModel
 
 from src.app.models.collections import Collection
 from src.app.models.documents import JourneySectionType
@@ -184,7 +185,7 @@ def choose_readability_according_journey_section_type(
 @log_time_and_error_sync
 async def collection_and_model_id_according_lang(
     lang: str | None, sp
-) -> tuple[Collection, uuid]:
+) -> tuple[Collection, list[EmbeddingModel]]:
     """
     Get the collection info and model id according to the language.
     Args:
@@ -203,9 +204,9 @@ async def collection_and_model_id_according_lang(
         raise HTTPException(
             status_code=404, detail=f"No collection found for language '{lang}'."
         )
-    model_id = get_embeddings_model_id_according_name(collection_info.model)
-    if not model_id:
+    emb_models = get_embeddings_model_id_according_name(collection_info.model)
+    if not emb_models:
         raise ValueError(
             f"Embedding model '{collection_info.model}' not found in the database."
         )
-    return collection_info, model_id
+    return collection_info, emb_models
diff --git a/src/app/services/sql_db/queries.py b/src/app/services/sql_db/queries.py
index 031709b..1af13aa 100644
--- a/src/app/services/sql_db/queries.py
+++ b/src/app/services/sql_db/queries.py
@@ -173,7 +173,7 @@ def get_subject(subject: str, embedding_model_id: UUID) -> ContextDocument | Non
     return subject_meta_document
 
 
-def get_subjects(embedding_model_id: UUID) -> list[ContextDocument]:
+def get_subjects(embedding_models_ids: list[UUID]) -> list[ContextDocument]:
     """
     Get all the subject meta documents from the database.
     Returns: List of subject meta documents.
@@ -183,7 +183,7 @@ def get_subjects(embedding_model_id: UUID) -> list[ContextDocument]:
             session.query(ContextDocument)
             .filter(
                 ContextDocument.context_type == ContextType.SUBJECT.value.lower(),
-                ContextDocument.embedding_model_id == embedding_model_id,
+                ContextDocument.embedding_model_id.in_(embedding_models_ids),
             )
             .all()
         )
@@ -214,7 +214,9 @@ def get_context_documents(
     return sdg_meta_documents
 
 
-def get_embeddings_model_id_according_name(model_name: str) -> list[UUID] | None:
+def get_embeddings_model_id_according_name(
+    model_name: str,
+) -> list[EmbeddingModel | None]:
     """
     Get the embeddings model ID according to its name.
 
@@ -225,12 +227,11 @@ def get_embeddings_model_id_according_name(model_name: str) -> list[UUID] | None
         The ID of the embeddings model if found, otherwise None.
     """
     with session_maker() as session:
-        model = (
+        return (
             session.query(EmbeddingModel)
             .filter(EmbeddingModel.title == model_name)
-            .first()
+            .all()
         )
-        return model.id if model else None
 
 
 def write_new_data_quality_error(

From dbcfde0ed8b8891545ac4fd89820e1dfadf7fabf Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Th=C3=A9o?= <theo.nardin@cri-paris.org>
Date: Thu, 5 Mar 2026 15:01:50 +0100
Subject: [PATCH 03/14] feat(micro-learning): update functions to support
 multiple embedding models

---
 src/app/api/api_v1/endpoints/micro_learning.py |  8 ++++----
 src/app/services/search.py                     |  8 +++++---
 src/app/services/sql_db/queries.py             | 12 ++++++++----
 3 files changed, 17 insertions(+), 11 deletions(-)

diff --git a/src/app/api/api_v1/endpoints/micro_learning.py b/src/app/api/api_v1/endpoints/micro_learning.py
index b54a7c3..ea75559 100644
--- a/src/app/api/api_v1/endpoints/micro_learning.py
+++ b/src/app/api/api_v1/endpoints/micro_learning.py
@@ -53,13 +53,13 @@ async def get_full_journey(
     lang: str | None = None,
     sp: SearchService = Depends(get_search_service),
 ):
-    collection_info, model_id = await collection_and_model_id_according_lang(
+    collection_info, models_id = await collection_and_model_id_according_lang(
         sp=sp, lang=lang
     )
-
+    emb_models_ids = [model.id for model in models_id]
     journey_part = [i.lower() for i in JourneySectionType]
     sdg_meta_documents = get_context_documents(
-        journey_part=journey_part, sdg=sdg, embedding_model_id=model_id
+        journey_part=journey_part, sdg=sdg, embedding_models_ids=emb_models_ids
     )
     if not sdg_meta_documents:
         raise HTTPException(
@@ -67,7 +67,7 @@ async def get_full_journey(
         )
 
     subject_meta_document: ContextDocument | None = get_subject(
-        subject=subject, embedding_model_id=model_id
+        subject=subject, embedding_models_ids=emb_models_ids
     )
 
     if not subject_meta_document:
diff --git a/src/app/services/search.py b/src/app/services/search.py
index d52fb56..d98d461 100644
--- a/src/app/services/search.py
+++ b/src/app/services/search.py
@@ -25,7 +25,7 @@
 from src.app.services.data_quality import DataQualityChecker
 from src.app.services.exceptions import CollectionNotFoundError, ModelNotFoundError
 from src.app.services.helpers import convert_embedding_bytes
-from src.app.services.sql_db.queries import get_subject
+from src.app.services.sql_db.queries import get_subject, get_embeddings_model_id_according_name
 from src.app.utils.decorators import log_time_and_error, log_time_and_error_sync
 from src.app.utils.logger import logger as logger_utils
 
@@ -443,7 +443,7 @@ def concatenate_same_doc_id_slices(
 
 
 @log_time_and_error_sync
-def get_subject_vector(subject: str | None) -> list[float] | None:
+def get_subject_vector(subject: str | None, model_name: str) -> list[float] | None:
     """
     Get the subject vector from the database.
     Args:
@@ -455,7 +455,9 @@ def get_subject_vector(subject: str | None) -> list[float] | None:
     if not subject:
         return None
 
-    subject_from_db = get_subject(subject=subject)
+    emb_models = get_embeddings_model_id_according_name(model_name)
+
+    subject_from_db = get_subject(subject=subject, embedding_models_ids=[m.id for m in emb_models])
     if not subject_from_db:
         return None
 
diff --git a/src/app/services/sql_db/queries.py b/src/app/services/sql_db/queries.py
index 1af13aa..370c44a 100644
--- a/src/app/services/sql_db/queries.py
+++ b/src/app/services/sql_db/queries.py
@@ -151,10 +151,13 @@ def register_endpoint(endpoint, session_id, http_code):
         session.commit()
 
 
-def get_subject(subject: str, embedding_model_id: UUID) -> ContextDocument | None:
+def get_subject(
+    subject: str, embedding_models_ids: list[UUID]
+) -> ContextDocument | None:
     """
     Get the subject meta document from the database.
     Args:
+        embedding_models_ids: Database IDs of embeddings models used for vectorize documents
         subject: The subject to get.
 
     Returns: The subject meta document.
@@ -166,7 +169,7 @@ def get_subject(subject: str, embedding_model_id: UUID) -> ContextDocument | Non
             .filter(
                 ContextDocument.context_type == ContextType.SUBJECT.value.lower(),
                 ContextDocument.title == subject,
-                ContextDocument.embedding_model_id == embedding_model_id,
+                ContextDocument.embedding_model_id.in_(embedding_models_ids),
             )
             .first()
         )
@@ -191,12 +194,13 @@ def get_subjects(embedding_models_ids: list[UUID]) -> list[ContextDocument]:
 
 
 def get_context_documents(
-    journey_part: JourneySection, sdg: int, embedding_model_id: UUID
+    journey_part: list[JourneySection], sdg: int, embedding_models_ids: list[UUID]
 ):
     """
     Get the context documents from the database.
 
     Args:
+        embedding_models_ids: Database IDs of embeddings models used for vectorize documents
         journey_part: The journey part to get the context documents for.
         sdg: The SDG to get the context documents for.
     Returns: List of context documents.
@@ -207,7 +211,7 @@ def get_context_documents(
             .filter(
                 ContextDocument.context_type.in_(journey_part),
                 ContextDocument.sdg_related.contains([sdg]),
-                ContextDocument.embedding_model_id == embedding_model_id,
+                ContextDocument.embedding_model_id.in_(embedding_models_ids),
             )
             .all()
         )

From cf5b0a2be959f2f5819b13a092069fbd0992b776 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Th=C3=A9o?= <theo.nardin@cri-paris.org>
Date: Thu, 5 Mar 2026 15:02:19 +0100
Subject: [PATCH 04/14] type ignore

---
 src/app/services/search.py | 9 +++++++--
 1 file changed, 7 insertions(+), 2 deletions(-)

diff --git a/src/app/services/search.py b/src/app/services/search.py
index d98d461..14cde2d 100644
--- a/src/app/services/search.py
+++ b/src/app/services/search.py
@@ -25,7 +25,10 @@
 from src.app.services.data_quality import DataQualityChecker
 from src.app.services.exceptions import CollectionNotFoundError, ModelNotFoundError
 from src.app.services.helpers import convert_embedding_bytes
-from src.app.services.sql_db.queries import get_subject, get_embeddings_model_id_according_name
+from src.app.services.sql_db.queries import (
+    get_embeddings_model_id_according_name,
+    get_subject,
+)
 from src.app.utils.decorators import log_time_and_error, log_time_and_error_sync
 from src.app.utils.logger import logger as logger_utils
 
@@ -457,7 +460,9 @@ def get_subject_vector(subject: str | None, model_name: str) -> list[float] | No
 
     emb_models = get_embeddings_model_id_according_name(model_name)
 
-    subject_from_db = get_subject(subject=subject, embedding_models_ids=[m.id for m in emb_models])
+    subject_from_db = get_subject(
+        subject=subject, embedding_models_ids=[m.id for m in emb_models]  # type: ignore
+    )
     if not subject_from_db:
         return None
 

From df64c55f0e1c7c405c04886b75197d8f087813b3 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Th=C3=A9o?= <theo.nardin@cri-paris.org>
Date: Thu, 5 Mar 2026 15:50:24 +0100
Subject: [PATCH 05/14] refactor(queries): clean up unused imports in
 queries.py

---
 src/app/services/sql_db/queries.py | 4 +---
 1 file changed, 1 insertion(+), 3 deletions(-)

diff --git a/src/app/services/sql_db/queries.py b/src/app/services/sql_db/queries.py
index 370c44a..f70abf9 100644
--- a/src/app/services/sql_db/queries.py
+++ b/src/app/services/sql_db/queries.py
@@ -7,7 +7,7 @@
 from qdrant_client.http.models import ScoredPoint
 from sqlalchemy import func, select
 from welearn_database.data.enumeration import Step
-from welearn_database.data.models import (
+from welearn_database.data.models import (  # FilterType,; FilterUsedInQuery,
     ChatMessage,
     ContextDocument,
     Corpus,
@@ -17,8 +17,6 @@
     EmbeddingModel,
     EndpointRequest,
     ErrorDataQuality,
-    FilterType,
-    FilterUsedInQuery,
     ProcessState,
     QtyDocumentInQdrant,
     QtyDocumentInQdrantPerCorpus,

From d5b2a66a438fdfe65d238eb1ab747e9512c7298c Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Th=C3=A9o?= <theo.nardin@cri-paris.org>
Date: Thu, 5 Mar 2026 12:04:33 +0100
Subject: [PATCH 06/14] Remove unused cache dict

---
 src/app/services/sql_db/queries.py | 7 +------
 1 file changed, 1 insertion(+), 6 deletions(-)

diff --git a/src/app/services/sql_db/queries.py b/src/app/services/sql_db/queries.py
index 170f7b4..031709b 100644
--- a/src/app/services/sql_db/queries.py
+++ b/src/app/services/sql_db/queries.py
@@ -34,7 +34,6 @@
 from src.app.services.constants import APP_NAME
 from src.app.services.sql_db.sql_service import session_maker
 
-model_id_cache: dict[str, UUID] = {}
 model_id_lock = Lock()
 
 
@@ -215,7 +214,7 @@ def get_context_documents(
     return sdg_meta_documents
 
 
-def get_embeddings_model_id_according_name(model_name: str) -> UUID | None:
+def get_embeddings_model_id_according_name(model_name: str) -> list[UUID] | None:
     """
     Get the embeddings model ID according to its name.
 
@@ -225,10 +224,6 @@ def get_embeddings_model_id_according_name(model_name: str) -> UUID | None:
     Returns:
         The ID of the embeddings model if found, otherwise None.
     """
-    with model_id_lock:
-        if model_name in model_id_cache:
-            return model_id_cache[model_name]
-
     with session_maker() as session:
         model = (
             session.query(EmbeddingModel)

From f8b14cde6d69303a283229e3462753976dae9503 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Th=C3=A9o?= <theo.nardin@cri-paris.org>
Date: Thu, 5 Mar 2026 12:20:55 +0100
Subject: [PATCH 07/14] feat(micro-learning): update embedding model handling
 to support multiple models

---
 src/app/api/api_v1/endpoints/micro_learning.py |  7 +++++--
 src/app/services/helpers.py                    |  9 +++++----
 src/app/services/sql_db/queries.py             | 13 +++++++------
 3 files changed, 17 insertions(+), 12 deletions(-)

diff --git a/src/app/api/api_v1/endpoints/micro_learning.py b/src/app/api/api_v1/endpoints/micro_learning.py
index 460fcc8..b54a7c3 100644
--- a/src/app/api/api_v1/endpoints/micro_learning.py
+++ b/src/app/api/api_v1/endpoints/micro_learning.py
@@ -30,10 +30,13 @@
 async def get_subject_list(
     lang: str | None = None, sp: SearchService = Depends(get_search_service)
 ) -> list[str]:
-    collection_info, model_id = await collection_and_model_id_according_lang(
+    collection_info, models_ids = await collection_and_model_id_according_lang(
         sp=sp, lang=lang
     )
-    ret = [md.title for md in get_subjects(embedding_model_id=model_id)]
+    ret = [
+        md.title
+        for md in get_subjects(embedding_models_ids=[model.id for model in models_ids])
+    ]
     if len(ret) == 0:
         raise HTTPException(status_code=404, detail="No subjects found.")
     return ret
diff --git a/src/app/services/helpers.py b/src/app/services/helpers.py
index a4cee7c..d5374a8 100644
--- a/src/app/services/helpers.py
+++ b/src/app/services/helpers.py
@@ -8,6 +8,7 @@
 from json_repair import JSONReturnType
 from langdetect import detect_langs
 from qdrant_client.http.models import models
+from welearn_database.data.models import EmbeddingModel
 
 from src.app.models.collections import Collection
 from src.app.models.documents import JourneySectionType
@@ -184,7 +185,7 @@ def choose_readability_according_journey_section_type(
 @log_time_and_error_sync
 async def collection_and_model_id_according_lang(
     lang: str | None, sp
-) -> tuple[Collection, uuid]:
+) -> tuple[Collection, list[EmbeddingModel]]:
     """
     Get the collection info and model id according to the language.
     Args:
@@ -203,9 +204,9 @@ async def collection_and_model_id_according_lang(
         raise HTTPException(
             status_code=404, detail=f"No collection found for language '{lang}'."
         )
-    model_id = get_embeddings_model_id_according_name(collection_info.model)
-    if not model_id:
+    emb_models = get_embeddings_model_id_according_name(collection_info.model)
+    if not emb_models:
         raise ValueError(
             f"Embedding model '{collection_info.model}' not found in the database."
         )
-    return collection_info, model_id
+    return collection_info, emb_models
diff --git a/src/app/services/sql_db/queries.py b/src/app/services/sql_db/queries.py
index 031709b..1af13aa 100644
--- a/src/app/services/sql_db/queries.py
+++ b/src/app/services/sql_db/queries.py
@@ -173,7 +173,7 @@ def get_subject(subject: str, embedding_model_id: UUID) -> ContextDocument | Non
     return subject_meta_document
 
 
-def get_subjects(embedding_model_id: UUID) -> list[ContextDocument]:
+def get_subjects(embedding_models_ids: list[UUID]) -> list[ContextDocument]:
     """
     Get all the subject meta documents from the database.
     Returns: List of subject meta documents.
@@ -183,7 +183,7 @@ def get_subjects(embedding_model_id: UUID) -> list[ContextDocument]:
             session.query(ContextDocument)
             .filter(
                 ContextDocument.context_type == ContextType.SUBJECT.value.lower(),
-                ContextDocument.embedding_model_id == embedding_model_id,
+                ContextDocument.embedding_model_id.in_(embedding_models_ids),
             )
             .all()
         )
@@ -214,7 +214,9 @@ def get_context_documents(
     return sdg_meta_documents
 
 
-def get_embeddings_model_id_according_name(model_name: str) -> list[UUID] | None:
+def get_embeddings_model_id_according_name(
+    model_name: str,
+) -> list[EmbeddingModel | None]:
     """
     Get the embeddings model ID according to its name.
 
@@ -225,12 +227,11 @@ def get_embeddings_model_id_according_name(model_name: str) -> list[UUID] | None
         The ID of the embeddings model if found, otherwise None.
     """
     with session_maker() as session:
-        model = (
+        return (
             session.query(EmbeddingModel)
             .filter(EmbeddingModel.title == model_name)
-            .first()
+            .all()
         )
-        return model.id if model else None
 
 
 def write_new_data_quality_error(

From cd3a57a6180c98537115d928099b3bc776e49bac Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Th=C3=A9o?= <theo.nardin@cri-paris.org>
Date: Thu, 5 Mar 2026 15:01:50 +0100
Subject: [PATCH 08/14] feat(micro-learning): update functions to support
 multiple embedding models

---
 src/app/api/api_v1/endpoints/micro_learning.py |  8 ++++----
 src/app/services/search.py                     |  8 +++++---
 src/app/services/sql_db/queries.py             | 12 ++++++++----
 3 files changed, 17 insertions(+), 11 deletions(-)

diff --git a/src/app/api/api_v1/endpoints/micro_learning.py b/src/app/api/api_v1/endpoints/micro_learning.py
index b54a7c3..ea75559 100644
--- a/src/app/api/api_v1/endpoints/micro_learning.py
+++ b/src/app/api/api_v1/endpoints/micro_learning.py
@@ -53,13 +53,13 @@ async def get_full_journey(
     lang: str | None = None,
     sp: SearchService = Depends(get_search_service),
 ):
-    collection_info, model_id = await collection_and_model_id_according_lang(
+    collection_info, models_id = await collection_and_model_id_according_lang(
         sp=sp, lang=lang
     )
-
+    emb_models_ids = [model.id for model in models_id]
     journey_part = [i.lower() for i in JourneySectionType]
     sdg_meta_documents = get_context_documents(
-        journey_part=journey_part, sdg=sdg, embedding_model_id=model_id
+        journey_part=journey_part, sdg=sdg, embedding_models_ids=emb_models_ids
     )
     if not sdg_meta_documents:
         raise HTTPException(
@@ -67,7 +67,7 @@ async def get_full_journey(
         )
 
     subject_meta_document: ContextDocument | None = get_subject(
-        subject=subject, embedding_model_id=model_id
+        subject=subject, embedding_models_ids=emb_models_ids
     )
 
     if not subject_meta_document:
diff --git a/src/app/services/search.py b/src/app/services/search.py
index d52fb56..d98d461 100644
--- a/src/app/services/search.py
+++ b/src/app/services/search.py
@@ -25,7 +25,7 @@
 from src.app.services.data_quality import DataQualityChecker
 from src.app.services.exceptions import CollectionNotFoundError, ModelNotFoundError
 from src.app.services.helpers import convert_embedding_bytes
-from src.app.services.sql_db.queries import get_subject
+from src.app.services.sql_db.queries import get_subject, get_embeddings_model_id_according_name
 from src.app.utils.decorators import log_time_and_error, log_time_and_error_sync
 from src.app.utils.logger import logger as logger_utils
 
@@ -443,7 +443,7 @@ def concatenate_same_doc_id_slices(
 
 
 @log_time_and_error_sync
-def get_subject_vector(subject: str | None) -> list[float] | None:
+def get_subject_vector(subject: str | None, model_name: str) -> list[float] | None:
     """
     Get the subject vector from the database.
     Args:
@@ -455,7 +455,9 @@ def get_subject_vector(subject: str | None) -> list[float] | None:
     if not subject:
         return None
 
-    subject_from_db = get_subject(subject=subject)
+    emb_models = get_embeddings_model_id_according_name(model_name)
+
+    subject_from_db = get_subject(subject=subject, embedding_models_ids=[m.id for m in emb_models])
     if not subject_from_db:
         return None
 
diff --git a/src/app/services/sql_db/queries.py b/src/app/services/sql_db/queries.py
index 1af13aa..370c44a 100644
--- a/src/app/services/sql_db/queries.py
+++ b/src/app/services/sql_db/queries.py
@@ -151,10 +151,13 @@ def register_endpoint(endpoint, session_id, http_code):
         session.commit()
 
 
-def get_subject(subject: str, embedding_model_id: UUID) -> ContextDocument | None:
+def get_subject(
+    subject: str, embedding_models_ids: list[UUID]
+) -> ContextDocument | None:
     """
     Get the subject meta document from the database.
     Args:
+        embedding_models_ids: Database IDs of embeddings models used for vectorize documents
         subject: The subject to get.
 
     Returns: The subject meta document.
@@ -166,7 +169,7 @@ def get_subject(subject: str, embedding_model_id: UUID) -> ContextDocument | Non
             .filter(
                 ContextDocument.context_type == ContextType.SUBJECT.value.lower(),
                 ContextDocument.title == subject,
-                ContextDocument.embedding_model_id == embedding_model_id,
+                ContextDocument.embedding_model_id.in_(embedding_models_ids),
             )
             .first()
         )
@@ -191,12 +194,13 @@ def get_subjects(embedding_models_ids: list[UUID]) -> list[ContextDocument]:
 
 
 def get_context_documents(
-    journey_part: JourneySection, sdg: int, embedding_model_id: UUID
+    journey_part: list[JourneySection], sdg: int, embedding_models_ids: list[UUID]
 ):
     """
     Get the context documents from the database.
 
     Args:
+        embedding_models_ids: Database IDs of embeddings models used for vectorize documents
         journey_part: The journey part to get the context documents for.
         sdg: The SDG to get the context documents for.
     Returns: List of context documents.
@@ -207,7 +211,7 @@ def get_context_documents(
             .filter(
                 ContextDocument.context_type.in_(journey_part),
                 ContextDocument.sdg_related.contains([sdg]),
-                ContextDocument.embedding_model_id == embedding_model_id,
+                ContextDocument.embedding_model_id.in_(embedding_models_ids),
             )
             .all()
         )

From 94291d9893c9663fd6a1db133ef754d4d0979302 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Th=C3=A9o?= <theo.nardin@cri-paris.org>
Date: Thu, 5 Mar 2026 15:02:19 +0100
Subject: [PATCH 09/14] type ignore

---
 src/app/services/search.py | 9 +++++++--
 1 file changed, 7 insertions(+), 2 deletions(-)

diff --git a/src/app/services/search.py b/src/app/services/search.py
index d98d461..14cde2d 100644
--- a/src/app/services/search.py
+++ b/src/app/services/search.py
@@ -25,7 +25,10 @@
 from src.app.services.data_quality import DataQualityChecker
 from src.app.services.exceptions import CollectionNotFoundError, ModelNotFoundError
 from src.app.services.helpers import convert_embedding_bytes
-from src.app.services.sql_db.queries import get_subject, get_embeddings_model_id_according_name
+from src.app.services.sql_db.queries import (
+    get_embeddings_model_id_according_name,
+    get_subject,
+)
 from src.app.utils.decorators import log_time_and_error, log_time_and_error_sync
 from src.app.utils.logger import logger as logger_utils
 
@@ -457,7 +460,9 @@ def get_subject_vector(subject: str | None, model_name: str) -> list[float] | No
 
     emb_models = get_embeddings_model_id_according_name(model_name)
 
-    subject_from_db = get_subject(subject=subject, embedding_models_ids=[m.id for m in emb_models])
+    subject_from_db = get_subject(
+        subject=subject, embedding_models_ids=[m.id for m in emb_models]  # type: ignore
+    )
     if not subject_from_db:
         return None
 

From c04374e8914453da1b720fb83bc7fb8b54817b48 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Th=C3=A9o?= <theo.nardin@cri-paris.org>
Date: Thu, 5 Mar 2026 15:50:24 +0100
Subject: [PATCH 10/14] refactor(queries): clean up unused imports in
 queries.py

---
 src/app/services/sql_db/queries.py | 4 +---
 1 file changed, 1 insertion(+), 3 deletions(-)

diff --git a/src/app/services/sql_db/queries.py b/src/app/services/sql_db/queries.py
index 370c44a..f70abf9 100644
--- a/src/app/services/sql_db/queries.py
+++ b/src/app/services/sql_db/queries.py
@@ -7,7 +7,7 @@
 from qdrant_client.http.models import ScoredPoint
 from sqlalchemy import func, select
 from welearn_database.data.enumeration import Step
-from welearn_database.data.models import (
+from welearn_database.data.models import (  # FilterType,; FilterUsedInQuery,
     ChatMessage,
     ContextDocument,
     Corpus,
@@ -17,8 +17,6 @@
     EmbeddingModel,
     EndpointRequest,
     ErrorDataQuality,
-    FilterType,
-    FilterUsedInQuery,
     ProcessState,
     QtyDocumentInQdrant,
     QtyDocumentInQdrantPerCorpus,

From ebe50045128a15369d405bb4f5e0f9365c694e60 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Th=C3=A9o?= <theo.nardin@cri-paris.org>
Date: Mon, 9 Mar 2026 16:02:40 +0100
Subject: [PATCH 11/14] Fix test

---
 src/app/tests/api/api_v1/test_micro_learning.py | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)

diff --git a/src/app/tests/api/api_v1/test_micro_learning.py b/src/app/tests/api/api_v1/test_micro_learning.py
index fcbdfd5..7162a24 100644
--- a/src/app/tests/api/api_v1/test_micro_learning.py
+++ b/src/app/tests/api/api_v1/test_micro_learning.py
@@ -1,8 +1,9 @@
 import unittest
+import uuid
 from unittest import mock
 
 from fastapi.testclient import TestClient
-from welearn_database.data.models import ContextDocument
+from welearn_database.data.models import ContextDocument, EmbeddingModel
 
 from src.app.core.config import settings
 from src.app.models.collections import Collection
@@ -38,7 +39,7 @@ async def test_get_full_journey(
         with TestClient(app) as client:
             mock_collection_and_model_id_according_lang.return_value = (
                 Collection(name="test_collection", lang="en", model="test_model"),
-                "model_id",
+                [EmbeddingModel(id=uuid.uuid4(), title="test_model", lang="en")],
             )
             # Mock data
             mock_get_context_docs.return_value = [

From be8668e47f41f61b3da092e230e79bb1824e3a46 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Th=C3=A9o?= <theo.nardin@cri-paris.org>
Date: Mon, 9 Mar 2026 16:09:25 +0100
Subject: [PATCH 12/14] fix(micro-learning): pass model to get_subject_vector
 in search.py

---
 src/app/services/search.py                      | 4 +++-
 src/app/tests/api/api_v1/test_micro_learning.py | 2 +-
 2 files changed, 4 insertions(+), 2 deletions(-)

diff --git a/src/app/services/search.py b/src/app/services/search.py
index 14cde2d..d30c62b 100644
--- a/src/app/services/search.py
+++ b/src/app/services/search.py
@@ -256,7 +256,9 @@ async def search_handler(
         assert isinstance(qp.query, str)
 
         collection = await self.get_collection_by_language(lang="mul")
-        subject_vector = await run_in_threadpool(get_subject_vector, qp.subject)
+        subject_vector = await run_in_threadpool(
+            get_subject_vector, qp.subject, collection.model
+        )
         embedding = await self.get_query_embed(
             model=collection.model,
             query=qp.query,
diff --git a/src/app/tests/api/api_v1/test_micro_learning.py b/src/app/tests/api/api_v1/test_micro_learning.py
index 7162a24..025c729 100644
--- a/src/app/tests/api/api_v1/test_micro_learning.py
+++ b/src/app/tests/api/api_v1/test_micro_learning.py
@@ -115,7 +115,7 @@ async def test_get_subject_list(
             ]
             mock_collection_and_model_id_according_lang.return_value = (
                 None,
-                "model_id",
+                [],
             )
 
             # API call

From 23908cf17126401bdb341ecf660c7d4fce6ca1cf Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Th=C3=A9o?= <theo.nardin@cri-paris.org>
Date: Tue, 10 Mar 2026 10:49:53 +0100
Subject: [PATCH 13/14] refactor: organize imports in agents.py, helpers.py,
 and queries.py

---
 src/app/services/helpers.py        | 1 -
 src/app/services/sql_db/queries.py | 4 +++-
 src/app/services/tutor/agents.py   | 2 +-
 3 files changed, 4 insertions(+), 3 deletions(-)

diff --git a/src/app/services/helpers.py b/src/app/services/helpers.py
index d5374a8..9fce3d5 100644
--- a/src/app/services/helpers.py
+++ b/src/app/services/helpers.py
@@ -1,4 +1,3 @@
-import uuid
 from functools import cache
 from typing import Any, List
 
diff --git a/src/app/services/sql_db/queries.py b/src/app/services/sql_db/queries.py
index f70abf9..370c44a 100644
--- a/src/app/services/sql_db/queries.py
+++ b/src/app/services/sql_db/queries.py
@@ -7,7 +7,7 @@
 from qdrant_client.http.models import ScoredPoint
 from sqlalchemy import func, select
 from welearn_database.data.enumeration import Step
-from welearn_database.data.models import (  # FilterType,; FilterUsedInQuery,
+from welearn_database.data.models import (
     ChatMessage,
     ContextDocument,
     Corpus,
@@ -17,6 +17,8 @@
     EmbeddingModel,
     EndpointRequest,
     ErrorDataQuality,
+    FilterType,
+    FilterUsedInQuery,
     ProcessState,
     QtyDocumentInQdrant,
     QtyDocumentInQdrantPerCorpus,
diff --git a/src/app/services/tutor/agents.py b/src/app/services/tutor/agents.py
index 6bb1947..0c73a7a 100644
--- a/src/app/services/tutor/agents.py
+++ b/src/app/services/tutor/agents.py
@@ -1,5 +1,5 @@
-import time
 import json
+import time
 from pathlib import Path
 
 from langchain_core.language_models import BaseChatModel

From e0435feca3a3f57450ea31191db158298bbec695 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Th=C3=A9o?= <133012334+lpi-tn@users.noreply.github.com>
Date: Tue, 10 Mar 2026 11:13:10 +0100
Subject: [PATCH 14/14] Apply suggestions from code review

Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>
---
 src/app/services/search.py | 6 +++++-
 1 file changed, 5 insertions(+), 1 deletion(-)

diff --git a/src/app/services/search.py b/src/app/services/search.py
index d30c62b..fdad4c1 100644
--- a/src/app/services/search.py
+++ b/src/app/services/search.py
@@ -461,9 +461,13 @@ def get_subject_vector(subject: str | None, model_name: str) -> list[float] | No
         return None
 
     emb_models = get_embeddings_model_id_according_name(model_name)
+    embedding_model_ids = [m.id for m in emb_models if m is not None]
+    if not embedding_model_ids:
+        return None
 
     subject_from_db = get_subject(
-        subject=subject, embedding_models_ids=[m.id for m in emb_models]  # type: ignore
+        subject=subject,
+        embedding_models_ids=embedding_model_ids,
     )
     if not subject_from_db:
         return None