migration to google-genai

requirements.txt

@@ -1,5 +1,7 @@
 streamlit
 chromadb
 langchain
-langchain-google-vertexai
+langchain-core
+langchain_community
+langchain-google-genai
 pypdf

tasks/task_10/task_10.py

@@ -12,9 +12,7 @@
 if __name__ == "__main__":
 
     embed_config = {
-        "model_name": "textembedding-gecko@003",
-        "project": "YOUR-PROJECT-ID-HERE",
-        "location": "us-central1"
+        "model_name": "models/embedding-001"
     }
 
     # Add Session State

tasks/task_4/task_4.py

@@ -1,20 +1,20 @@
 # embedding_client.py
 
-from langchain_google_vertexai import VertexAIEmbeddings
+from langchain_google_genai import GoogleGenerativeAIEmbeddings
 
 class EmbeddingClient:
     """
     Task: Initialize the EmbeddingClient class to connect to Google Cloud's VertexAI for text embeddings.
 
     The EmbeddingClient class should be capable of initializing an embedding client with specific configurations
     for model name, project, and location. Your task is to implement the __init__ method based on the provided
-    parameters. This setup will allow the class to utilize Google Cloud's VertexAIEmbeddings for processing text queries.
+    parameters. This setup will allow the class to utilize Google Cloud's GoogleGenerativeAIEmbeddings for processing text queries.
 
     Steps:
     1. Implement the __init__ method to accept 'model_name', 'project', and 'location' parameters.
-       These parameters are crucial for setting up the connection to the VertexAIEmbeddings service.
+       These parameters are crucial for setting up the connection to the GoogleGenerativeAIEmbeddings service.
 
-    2. Within the __init__ method, initialize the 'self.client' attribute as an instance of VertexAIEmbeddings
+    2. Within the __init__ method, initialize the 'self.client' attribute as an instance of GoogleGenerativeAIEmbeddings
        using the provided parameters. This attribute will be used to embed queries.
 
     Parameters:
@@ -23,17 +23,17 @@ class EmbeddingClient:
     - location: The location of the Google Cloud project, such as 'us-central1'.
 
     Instructions:
-    - Carefully initialize the 'self.client' with VertexAIEmbeddings in the __init__ method using the parameters.
+    - Carefully initialize the 'self.client' with GoogleGenerativeAIEmbeddings in the __init__ method using the parameters.
     - Pay attention to how each parameter is used to configure the embedding client.
 
     Note: The 'embed_query' method has been provided for you. Focus on correctly initializing the class.
     """
 
-    def __init__(self, model_name, project, location):
-        # Initialize the VertexAIEmbeddings client with the given parameters
-        # Read about the VertexAIEmbeddings wrapper from Langchain here
+    def __init__(self, model_name):
+        # Initialize the GoogleGenerativeAIEmbeddings client with the given parameters
+        # Read about the GoogleGenerativeAIEmbeddings wrapper from Langchain here
         # https://python.langchain.com/docs/integrations/text_embedding/google_generative_ai
-        self.client = VertexAIEmbeddings(
+        self.client = GoogleGenerativeAIEmbeddings(
             #### YOUR CODE HERE ####
         )
 
@@ -61,11 +61,9 @@ def embed_documents(self, documents):
             return None
 
 if __name__ == "__main__":
-    model_name = "textembedding-gecko@003"
-    project = "YOUR PROJECT ID HERE"
-    location = "us-central1"
+    model_name = "models/embedding-001"
 
-    embedding_client = EmbeddingClient(model_name, project, location)
+    embedding_client = EmbeddingClient(model_name)
     vectors = embedding_client.embed_query("Hello World!")
     if vectors:
         print(vectors)

tasks/task_5/task_5.py

@@ -92,9 +92,7 @@ def query_chroma_collection(self, query) -> Document:
     processor.ingest_documents()
 
     embed_config = {
-        "model_name": "textembedding-gecko@003",
-        "project": "YOUR PROJECT ID HERE",
-        "location": "us-central1"
+        "model_name": "models/embedding-001"
     }
 
     embed_client = EmbeddingClient(**embed_config) # Initialize from Task 4

tasks/task_6/task_6.py

@@ -44,9 +44,7 @@
 
     # Configuration for EmbeddingClient
     embed_config = {
-        "model_name": "textembedding-gecko@003",
-        "project": "YOUR PROJECT ID HERE",
-        "location": "us-central1"
+        "model_name": "models/embedding-001"
     }
 
     screen = st.empty() # Screen 1, ingest documents

tasks/task_7/task_7.py

@@ -1,5 +1,5 @@
 import streamlit as st
-from langchain_google_vertexai import VertexAI
+from langchain_google_genai import GoogleGenerativeAI
 from langchain_core.prompts import PromptTemplate
 import os
 import sys
@@ -65,12 +65,12 @@ def init_llm(self):
         4. Initialize the LLM with the specified parameters to be ready for generating quiz questions.
 
         Implementation:
-        - Use the VertexAI class to create an instance of the LLM with the specified configurations.
+        - Use the GoogleGenerativeAI class to create an instance of the LLM with the specified configurations.
         - Assign the created LLM instance to the 'self.llm' attribute for later use in question generation.
 
         Note: Ensure you have appropriate access or API keys if required by the model or platform.
         """
-        self.llm = VertexAI(
+        self.llm = GoogleGenerativeAI(
             ############# YOUR CODE HERE ############
         )
 
@@ -140,9 +140,7 @@ def generate_question_with_vectorstore(self):
 
 
     embed_config = {
-        "model_name": "textembedding-gecko@003",
-        "project": "YOUR-PROJECT-ID-HERE",
-        "location": "us-central1"
+        "model_name": "models/embedding-001"
     }
 
     screen = st.empty()

tasks/task_8/task_8.py

@@ -8,7 +8,7 @@
 from tasks.task_5.task_5 import ChromaCollectionCreator
 
 from langchain_core.prompts import PromptTemplate
-from langchain_google_vertexai import VertexAI
+from langchain_google_genai import GoogleGenerativeAI
 
 class QuizGenerator:
     def __init__(self, topic=None, num_questions=1, vectorstore=None):
@@ -66,8 +66,8 @@ def init_llm(self):
 
         :return: An instance or configuration for the LLM.
         """
-        self.llm = VertexAI(
-            model_name = "gemini-pro",
+        self.llm = GoogleGenerativeAI(
+            model = "gemini-1.5-pro",
             temperature = 0.8, # Increased for less deterministic questions 
             max_output_tokens = 500
         )
@@ -177,9 +177,7 @@ def validate_question(self, question: dict) -> bool:
 if __name__ == "__main__":
 
     embed_config = {
-        "model_name": "textembedding-gecko@003",
-        "project": "YOUR-PROJECT-ID-HERE",
-        "location": "us-central1"
+        "model_name": "models/embedding-001"
     }
 
     screen = st.empty()

tasks/task_9/task_9.py

@@ -70,9 +70,7 @@ def next_question_index(self, direction=1):
 if __name__ == "__main__":
 
     embed_config = {
-        "model_name": "textembedding-gecko@003",
-        "project": "YOUR-PROJECT-ID-HERE",
-        "location": "us-central1"
+        "model_name": "models/embedding-001"
     }
 
     screen = st.empty()