fix: bugs #5 and #8 (embeddings model name + error handling)

Bug #5 — UiPathFireworksEmbeddings: self.model returned wrong model name
  Without an explicit field override, Pydantic kept two separate fields:
  FireworksEmbeddings.model (default 'nomic-ai/nomic-embed-text-v1.5') and
  UiPathBaseLLMClient.model_name (populated by the caller via alias 'model').
  self.model therefore always resolved to the stale FireworksEmbeddings default,
  silently ignoring whatever model name the caller passed.
  Fix: add model: str = Field(default='', alias='model_name') — the same override
  pattern used by every other embedding class — to collapse the two fields into
  one. Replace self.model with self.model_name in all four embed methods.

Bug #8 — UiPathEmbeddings: silent HTTP error swallowing + missing model field
  Two issues in embed_documents / aembed_documents:
  1. uipath_request / uipath_arequest were called without raise_status_error=True,
     so HTTP 4xx/5xx responses were silently accepted. The subsequent
     response.json()['data'] access would then raise a confusing KeyError or
     AttributeError rather than a clear HTTP error.
  2. The request body only contained {'input': texts}. The normalized API
     requires a 'model' field for routing, matching the pattern already used
     by the normalized chat model (_default_params includes 'model').
  Fix: pass raise_status_error=True and add 'model': self.model_name to the
  request body in both the sync and async paths.

Author: cosminacho

packages/uipath_langchain_client/src/uipath_langchain_client/clients/fireworks/embeddings.py

@@ -1,6 +1,6 @@
 from typing import Self
 
-from pydantic import model_validator
+from pydantic import Field, model_validator
 
 from uipath_langchain_client.base_client import UiPathBaseEmbeddings
 from uipath_langchain_client.settings import (
@@ -31,6 +31,14 @@ class UiPathFireworksEmbeddings(UiPathBaseEmbeddings, FireworksEmbeddings):
         freeze_base_url=True,
     )
 
+    # Override FireworksEmbeddings.model so that Pydantic merges it with
+    # UiPathBaseLLMClient.model_name (alias="model") into a single field.
+    # Without this, two separate fields coexist: FireworksEmbeddings.model
+    # (default "nomic-ai/nomic-embed-text-v1.5") and model_name, causing
+    # self.model to always return the stale FireworksEmbeddings default
+    # instead of the model name supplied by the caller.
+    model: str = Field(default="", alias="model_name")
+
     @model_validator(mode="after")
     def setup_uipath_client(self) -> Self:
         self.client = OpenAI(
@@ -48,7 +56,8 @@ def setup_uipath_client(self) -> Self:
     def embed_documents(self, texts: list[str]) -> list[list[float]]:
         """Embed search docs."""
         return [
-            i.embedding for i in self.client.embeddings.create(input=texts, model=self.model).data
+            i.embedding
+            for i in self.client.embeddings.create(input=texts, model=self.model_name).data
         ]
 
     def embed_query(self, text: str) -> list[float]:
@@ -59,7 +68,9 @@ async def aembed_documents(self, texts: list[str]) -> list[list[float]]:
         """Embed search docs asynchronously."""
         return [
             i.embedding
-            for i in (await self.async_client.embeddings.create(input=texts, model=self.model)).data
+            for i in (
+                await self.async_client.embeddings.create(input=texts, model=self.model_name)
+            ).data
         ]
 
     async def aembed_query(self, text: str) -> list[float]:

packages/uipath_langchain_client/src/uipath_langchain_client/clients/normalized/embeddings.py

@@ -18,14 +18,20 @@ class UiPathEmbeddings(UiPathBaseEmbeddings, Embeddings):
     )
 
     def embed_documents(self, texts: list[str]) -> list[list[float]]:
-        response = self.uipath_request(request_body={"input": texts})
+        response = self.uipath_request(
+            request_body={"input": texts, "model": self.model_name},
+            raise_status_error=True,
+        )
         return [r["embedding"] for r in response.json()["data"]]
 
     def embed_query(self, text: str) -> list[float]:
         return self.embed_documents([text])[0]
 
     async def aembed_documents(self, texts: list[str]) -> list[list[float]]:
-        response = await self.uipath_arequest(request_body={"input": texts})
+        response = await self.uipath_arequest(
+            request_body={"input": texts, "model": self.model_name},
+            raise_status_error=True,
+        )
         return [r["embedding"] for r in response.json()["data"]]
 
     async def aembed_query(self, text: str) -> list[float]: