What are the main functions of the web ui console? I tried using this to create collections and save the code, but it doesn't seem like I actually created it. Thank you

Introduce GitHub actions.

Run on each commit:

• tests
• eslint checks

Use MUI open-source templates to define basic page components:

• Header
• Side bar
• Footer

Like a lot of other database client UIs, we can have a + Collection button in the homepage, where a user can quickly create a new collection using a form based interface and start working on it, instead of going to the console.

Something like this:

This will open up a form where user can fill in basic details of the collection, and also have an advanced section for further configurations

I'm keep getting this error do you guys know why?

Visualization Unsuccessful, error: {"message":"Request failed with status code 405","name":"AxiosError","stack":"AxiosError: Request failed with status code 405\n at qte (https://[CENSORED URL]/dashboard/assets/index-22a6d29c.js:295:959)\n at XMLHttpRequest.d (https://[CENSORED URL]/dashboard/assets/index-22a6d29c.js:295:4075)","config":{"transitional":{"silentJSONParsing":true,"forcedJSONParsing":true,"clarifyTimeoutError":false},"adapter":["xhr","http"],"transformRequest":[null],"transformResponse":[null],"timeout":0,"xsrfCookieName":"XSRF-TOKEN","xsrfHeaderName":"X-XSRF-TOKEN","maxContentLength":-1,"maxBodyLength":-1,"env":{},"headers":{"Accept":"application/json, text/plain, /","api-key":"[CENSORED API KEY]","Content-Type":"application/json"},"baseURL":"https://[CENSORED URL]/dashboard/","method":"post","url":"collections/me_BOT/points/scroll","data":"{"limit":500,"color_by":"source","with_vector":true}"},"code":"ERR_BAD_REQUEST","status":405}

Examples:

scroll request works properly and returns the necessary data.

Long类型精度丢失

It's not clear what the icon does, it can be misinterpreted as a "RUN all commands" or "Open new console" etc.

Adding a tooltip on hover that says "Commands" or something similar will make it clearer to the user.

If this issue is entertained, I would like to work on this.

We have the one file issuing requests other than the API itself

The Web-UI (http://localhost:6333/dashboard#/collections/<collection>/visualize) is very useful for quick exploration of a collection.
However, the UX might need some minor improvements. I guess that the most common use case is free exploration i.e., you take a look at your points and see which topics seem to form some kind of cluster. By doing so, currently when you hover over the points, first the raw JSON is displayed. Displaying the embeddings here is not needed.

When you click on the point, an overlay appears in a nice tabular format. The embeddings are hidden by default.

I think what would be better suited is an integration of the overlay table in the right side bar as there is lots of available space. Maybe two tabs would be good, one for the request and one for the point info?

If the tabular format was displayed for each point on hover in the right side bar, one could much easier explore a collection.

Anyway thanks a lot for providing the tool - much appreciated! :)

Hi! I'm getting this error, while visualizing my dataset on dashboard. Everything raedy but the visualization doesn't work out

Prepare and integrate a collection of useful examples which user can select to start with something.

Should have a menu to to select an example by name: click on example should insert in into the editor

Would love to contribute on this if @generall you agree.

We got a bug report: snapshot uploading is missing API key in request headers what prevent users from uploading to the cloud.

See, how the link previews are rendered right now.

Slack:

Discord:

This looks quite ugly, so let's change this

I suggest the following changes:
index.html

• change the <title> to: UI | Qdrant like you do already for when the SPA loads
• change the <meta name="description"> value to: Qdrant Web UI

public/manifest.json

• short_name: "UI | Qdrant"
• name: "Qdrant Web UI"

logos

• the logo being rendered in slack is either the logo192.png or logo512.png. Those need to be replaced by a qdrant logo of that size.

I build a dockerfile by myself looks below:

FROM node:18-alpine

WORKDIR /app

COPY . .

CMD ["npm", "start"]

And when I start the container, I got a error "sh: vite: not found". I'm not very familiar with front-end projects. It will make my life easier if you can provide a dockerfile which can handle all the dependencies.
Thank you.

I mostly use the auto completion to write the endpoint I want to query in the console but then I do not remember all the parameters which are required / optional. I just want to click on a link which deeplinks to the endpoint I want to query. The link could be next to the "RUN | BEAUTIFY" buttons. What do you think?

The UI Dashboard loads the monaco-editor package from an online CDN.
So the Dashboard console can't work while offline or on Kubernetes Pods without internet access

Current Behavior

Console won't load

Steps to Reproduce

1. Try to use the dashboard console on an offline machine

Expected Behavior

Loading the dependencies from the UI server.

Possible Solution

Add monaco-editor as a local dependency

As I navigate to different route eg /collections and come back to /console the Code lens is registered again and return 2 Run button

Design a page, which should allow to view all existing collections in Qdrant

https://drive.google.com/file/d/12ZcOreTsS8m4Oay4GSSG0oGDf_Qp8FsP/view?usp=sharing

There is a PR with multiple useful fixes: #38

• introduction of Vite
• hotkeys in console
• typescript intgration

But it is too huge to be merged in one run.
Let's split it function-wise and merge one-by-one

Make commentaries (lines, starting with //) a valid syntax construction in console

Advanced: make any text after // commentary, not only the whole line

A Qdrant admin UI is an amazing idea! It would be perfect if we could just run it from a pre-built docker image though...
Even better would be to have a version of qdrant vector store docker image with the UI served there as well, in a separate port.

would be very helpful if we migrate the entire codebase of the qdrant-web-ui as it will give more type safety to the code

Desired features:

• Be able to see first N records of the collection
• Should be possible to click on records to see similar ones
• If payload contains links, try to render links as images
• Nested payload values render as json

I noticed some issues with the dashboard when using sparse vectors. So creating an issue here for us to track those.

1. Sparse vector dims are not displayed in the dashboard
   

2. Sparse vector visualization doesn't work because the response format is slightly different

Both of these are low-priority tasks at the moment.

https://github.com/Lv-291/wasm-bhtsne

Very similar to #104, but for the download endpoint.

Discord thread link: https://discord.com/channels/907569970500743200/1143482730009612318

When testing queries a lot of time, I end up with the error:

{
  "stack": "Error: Failed to execute 'setItem' on 'Storage': Setting the value of 'history' exceeded the quota.\n    at https://127.0.0.1/dashboard/assets/index-92508d58.js:250:7963\n    at async https://127.0.0.1/dashboard/assets/index-92508d58.js:262:12108"
}

By clearing some queries in the History pop-up, I can finally send requests again.

Reset button doesn't work, it doesn't clear the history

I think old requests should be deleted automatically or limit must be increased (seems to be around 470)

This will reduce ambiguity and provide ease-of-mind for companies and developers that want to use/improve this tool.

The Qdrant DB uses Apache 2.0, maybe keep going with that?

It would be nice to provide an official container on your Docker Hub account.

Color mode couldn't presist it's states after reloading
its better to store the state in a local storage variable

Can you solve it? I can't take it anymore

Failed to fetch
version:
qdrant-web-ui 0f1243e
VITE v4.3.3
qdrant "1.2.2"

Add ability to save previously used editor states and, if necessary, go back to them.
Saved states should be placed into the local storage of the browser.

Hi,

I would like to request the addition of a color mode based on media queries. It would be great to have a color theme for both light and dark modes that switches automatically based on the user's current choice.

There are many different implementations of this and I'm not a front-end dev, but I like the way Material for MkDocs does it with a 3-state button.

Thanks in advance!

Implement autocomplete for requests based on OpenAPI schema of qdrant: https://qdrant.github.io/qdrant/redoc/index.html

Autocomplete should suggest:

• next allowed work in API url. Examples

  • /collect -> ["collections"]
  • /collections/mycollection/ -> ["points", "snapshots"]

• Allowed fields in JSON:

{
  "filter": {
    "|
  }
}

should suggest

must
should
must_not

The dashboard expects Qdrant to be hosted at the root of a domain otherwise it breaks. Our clusters work by hosting each service, or which Qdrant is one, on the same domain each with their own sub path. So for instance www.foo.com/qdrant.

If someone tries to run a time-consuming command to the server user didn't get any feedback that query is runing

I'm currently using Qdrant and I tried it for the first time the UI and I know it is currently working in progress,

But it is better we have a docker image for this so that we can add it to kubernetees.

Thanks

Hi - I have created the ID of the document using the following hash function.

record_id, _ = mmh3.hash64(file_data, signed=False)

If I check the ID using Python client, I get the actual ID. On Qdrant dashboard, it seems that we approximate the value to the nearest thousand
Suppose the actual id is 298631332630714982. On UI, we get ID 298631332630715000
It makes it slightly challenging to use the qdrant dashboard using record_id.

I would like to suggest moving to Next.js or Vite with TypeScript as CRA (create-react-app) is not recommended anymore.

Below are few reasons for the same -

1. Faster build times
2. Hot Code Reloading with Vite while Next.js provides a simplified API for handling server-side rendering and static site generation
3. Built-in TypeScript support - As the project grows, it would be helpful to have type safety within the app and the ability to access features like autocomplete through TypeScript

Hi !
It seems the fold button on the monaco editor is not fully accurate. I need to click upper than the button I see if I want to fold my JSON.

I guess this is implemented yet. But looks wrong to put CTA with an unimplemented functionality.
I think we should hide VISUALIZE button.

I tried to check github history of src/components/Collections/CollectionCard.jsx. This button was from start.
I am curious to know, is visualize coming soon ?

The visualisation tool requires a vector name to build a map

A user does not necessarily remembers vector names, they might be bulky or they can be autogenerated (e.g. with fastembed)

To find a vector name, one need to do some extra actions, like visiting other pages or making requests to the DB.

It would be nice to have autocomplete for vector_name field, or have available vector names somewhere in the vicinity.

If there is only one vector in a collection, then it might also be helpful to fill vector_name automatically.

Implementing Vector Visualization Page

This Issue proposes the implementation of a vector visualization page, which involves converting n-dimensional vectors into 2 dimensions and plotting the resulting points on a 2D graph. The page allows users to filter points, assign colours based on specified criteria (color_by), and select vectors using vector_name if a point has multiple vectors.

Route

/#/collections/:collection_name/visualize

Changes to be Made

• Implemented a filter editor window for users to customize point filtering.
• Performed dimension reduction on the n-dimensional vectors to plot them in 2D.
• Created a graph plotting component for visualizing the converted vectors.
• Included optimizers to improve the efficiency of the visualization process.

Task List

• Filter Editor window
• Dimension reduction
• Plotting Graph
• Optimizers

Please review and provide feedback on the proposed changes. Thank you!