additional_headers param not getting set in invoke calls

[ssvaidyanathan]

Checked other resources

• I added a very descriptive title to this question.
• I searched the LangChain documentation with the integrated search.
• I used the GitHub search to find a similar question and didn't find it.

Commit to Help

• I commit to help with one of those options 👆

Example Code

from langchain_google_vertexai import VertexAI
# Define project information
PROJECT_ID = "project"  # @param {type:"string"}
LOCATION = "region"  # @param {type:"string"}
API_ENDPOINT = "https://foo.com"  # @param {type:"string"}
API_KEY = "abcd"  # @param {type:"string"}
# Initialize Langchain
model = VertexAI(
      project=PROJECT_ID,
      location=LOCATION,
      api_endpoint=API_ENDPOINT,
      api_transport="rest",
      streaming=True,
      model_name="gemini-1.5-pro",
      additional_headers={"x-apikey": API_KEY})

prompts = ["Why is the sky blue?",
           "What makes the sky blue?",
           "Why does the sky look blue colored?",
           "Can you explain why the sky is blue?",
           "The sky is blue, why is that?"]

def invoke_model(prompt, model_to_invoke=model):
  model_to_invoke.invoke(prompt)

for prompt in prompts:
  model.invoke(prompt)

Description

• The above code snippet used to work. But recently we notice that the call being made to the custom API endpoint is not setting the additional headers that is required for us to verify.
• When I try to print the model variable using print(vars(model)), it prints all the params correctly including the additional_headers
• When I force the SDK to use the previous version

!pip install langchain==0.3.18
!pip install langchain-google-vertexai==2.0.12
!pip install google-cloud-aiplatform

It seems to work. Looks like its a regression issue on the new version released.

System Info

langchain==0.3.19
langchain-community==0.3.18
langchain-core==0.3.40
langchain-google-vertexai==2.0.12
langchain-text-splitters==0.3.6

platform: linux
python version: Python 3.10.12