firestore

Multimodal image storage and retrieval with Firestore Vector Store

In this sample, you'll learn how to store image embeddings to Firestore Vector Store and later retrieve images based on similarity search on a keyword or image.

You can see the full sample in main.py.

Setup Firestore

Make sure you're logged in:

gcloud auth application-default login

Enable Firestore API:

gcloud services enable firestore.googleapis.com

Create a Firestore database:

gcloud firestore databases create --database image-database --location=europe-west1

Create a Firestore index for retrieval later:

gcloud alpha firestore indexes composite create --project=your-project-id \
 --database="image-database" --collection-group=ImageCollection --query-scope=COLLECTION \
 --field-config=vector-config='{"dimension":"1408","flat": "{}"}',field-path=embedding

Add images

Let's add some images.

Add an image from a Cloud Storage url:

python main.py --project_id=genai-atamel --image_paths gs://genai-atamel-firestore-images/landmark1.png

Add other images from a local folder:

python main.py --project_id=genai-atamel --image_paths ../images/landmark2.png ../images/landmark3.png

Add another image from an HTTP url:

python main.py --project_id=genai-atamel --image_paths https://atamel.dev/img/mete-512.jpg

At this point, you should see images and their embeddings saved to Firestore:

Retrieve images

Now, retrieve and display images with a keyword:

python main.py --project_id=genai-atamel --search_by_keyword="stadium"
python main.py --project_id=genai-atamel --search_by_keyword="temple"
python main.py --project_id=genai-atamel --search_by_keyword="statue"
python main.py --project_id=genai-atamel --search_by_keyword="man"

You can also retrieve by searching similar images:

python main.py --project_id=genai-atamel --search_by_image="../images/landmark4.png"

References

• Build LLM-powered applications using LangChain
• Firestore VectorStore + LangChain notebook