When I want to add a new point to the collection, should I create it from scratch?

[serhatBilal]

Hi everyone, I started using qdrant today, I am very new to this subject.

I am working on similarity search in images and I added my images to qdrant, but later when I want to add other images to my database, it deletes the existing one and creates a new database, so I lose my old images.

Can I add as many images as I want to my existing collection at any time, can you help me with this? I share my code with you.
Thank you for your interest

`
class ImageEmbedding:

def image_to_database(self):

    base_directory = "Images"
    all_image_urls = os.listdir(base_directory)

    sample_image_urls = all_image_urls
    sample_image_urls = list(map(lambda item: f"{base_directory}/{item}",sample_image_urls))

    payloads = DataFrame.from_records({"image_url": sample_image_urls})
    payloads["model_id"] = 2

    images = list(map(lambda el:Image.open(el),payloads["image_url"]))

    target_width = 256

    def resize_image(image_url):
        pil_image = Image.open(image_url)
        image_aspect_ratio = pil_image.width / pil_image.height
        resized_pil_image = pil_image.resize([target_width,math.floor(target_width * image_aspect_ratio)])
        return resized_pil_image

    def convert_image_to_base64(pil_image):
        image_data = BytesIO()
        pil_image.save(image_data,format="JPEG")
        base64_string = base64.b64encode(image_data.getvalue()).decode("utf-8")
        return base64_string

    resized_images = list(map(lambda el: resize_image(el), sample_image_urls))
    base64_strings = list(map(lambda el: convert_image_to_base64(el), resized_images))
    payloads["base64"] = base64_strings

    processor = AutoImageProcessor.from_pretrained("microsoft/resnet-50")
    model = ResNetForImageClassification.from_pretrained("microsoft/resnet-50")

    inputs = processor(images,return_tensors="pt",)
    outputs = model(**inputs)
    embeddings = outputs.logits

    embedding_length = len(embeddings[0])

    load_dotenv()

    qclient = QdrantClient(
        url = os.getenv('QDRANT_DB_URL'),
        api_key = os.getenv('QDRANT_API_KEY'),
    )

    collection_name = "die_models_images"

    collection = qclient.recreate_collection(
        collection_name=collection_name,
        vectors_config=VectorParams(
            size=embedding_length,
            distance=Distance.COSINE
        )
    )

    payload_dicts = payloads.to_dict(orient="records")

    records = [
        models.Record(
            id=idx,
            payload=payload_dicts[idx],
            vector=embeddings[idx]
        )
        for idx, _ in  enumerate(payload_dicts)
    ]

    qclient.upload_records(
        collection_name=collection_name,
        records=records
    )  

`

[timvisee]

There are two things to consider:

• do not recreate the collection on each call, that will dete your current one
• make sure to use unique IDs for every point you insert

Currently you're using enumerate which will result in the same IDs in each call. If you don't care about the IDs the easiest option may be to assign a random UUID as point ID for each point.

For reference, I answered the same question here: https://stackoverflow.com/a/77988925/1000145

[serhatBilal]

thanks sir