added endpoint to embed query with text

gamechangerml/api/fastapi/routers/search.py

@@ -77,8 +77,28 @@ async def text_extract_infer(body: dict, extractType: str, response: Response) -
  raise
  return results
 
+@router.post("/embedSemanticQuery", status_code=200)
+async def semantic_search(
+ body: dict,
+ response: Response,
+) -> dict:
+ """
+ embedSemanticQuery - takes in a query and returns the txtai embeddings as a list of floats to be used by ElasticSearch
+
+ Args:
+ (dict) json format of the search query.\n 
+ query: (str, required) a string of any length to embed and use for semantic search.\n
+ Example: {"query": "Where is the science and technology reinvention laboratory?"}
+ Response: Response class; for status codes(apart of fastapi do not need to pass param)
+ Returns:
+ results: list of floats
+ """
+ query_text = body["query"]
+ embeddings = MODELS.semantic_searcher.embed_query(query_text)
+ return list(embeddings.astype(float))
+
 
-@ router.post("/semanticSearch", status_code=200)
+@router.post("/semanticSearch", status_code=200)
 async def semantic_search(
  body: dict,
  response: Response,

gamechangerml/src/search/sent_transformer/model.py

@@ -470,4 +470,8 @@ def search(
 
  return top_results
  else:
- return []
+ return []
+
+ def embed_query(self, query_text):
+ embeddings = self.embedder.transform(query_text)
+ return embeddings