Describe the bug

If the settings folders are not set at some point the downloads will silently fail

Expected behavior

downloads should work

Screenshots

Versions (please complete the following information):

• OS: Windows 11
• Captain 1.0.0-beta.7

Additional context

User Story

Title: Updating Local App Catalog Data
As a Captain user
I want a process to manually or automatically update my local app catalog data
So that I can ensure that my local marketplace reflects the most current apps, updates, and information available, even when I'm offline

Acceptance Criteria

• Users have the option to manually initiate an update of the local app catalog data whenever they are online.
• The system provides an option for users to set up automatic updates, which periodically refresh the local app catalog data when an internet connection is available.

User Story

Title: Seamless Offline App Download and Installation
As a Captain user
I want to queue AI applications for download when online and have them automatically downloaded and ready for installation, leveraging the locally stored marketplace data
So that I can install and use these applications seamlessly, even without an internet connection, while ensuring that the marketplace experience remains consistent and up-to-date with the manually or periodically updated app catalog data

Acceptance Criteria

• Users can select applications for download from the locally stored and periodically updated marketplace data when online.
• The offline installation process is integrated with offline-accessible documentation and support resources, ensuring users have all the information they need for a successful setup.

Is your feature request related to a problem? Please describe.
When starting the app, the installer process is always shown when there was an update to the app version.

Describe the solution you'd like
The installer should only ever be shown when the app gets installed, not when there is a version update of the app. I don't want to change my language / theming settings. I might only need to download the python-emebedded if that changed.

User Story

Title: Design a System for Storing and Retrieving Marketplace Data Locally
As a Captain developer
I want to design a system that fetches marketplace data from a JSON file in our GitHub repo's releases, stores this data locally on the user's device, and uses it to populate the marketplace
So that Captain users can browse, search, and manage AI applications seamlessly, with or without an internet connection, ensuring they have access to the most up-to-date information

Acceptance Criteria

• The system can automatically check for and download the latest JSON file from the GitHub repository's releases when the user is online.
• The downloaded JSON file is parsed, and its data is stored locally, updating the existing local data store with the latest marketplace information.
• Users have the option to manually trigger an update when connected to the internet.
• Robust error handling is in place for download and parsing processes, ensuring reliability and a smooth user experience.

Test feedback from app

Enable users to access GPU-based cloud computing within Captain for intensive AI tasks by integrating with RunPod's infrastructure.

Initial Spike: Assess Effort for RunPod GPU-Cloud Integration

• #106
• #107
• #108

Core Development Tasks for RunPod Integration:

• #218
• #219
• #220
• #233
• #234
• #221
• #222
• #223

Would it be possible to add the captions with moondream? It works quite well.

Thank you very much for your work :)

https://huggingface.co/spaces/vikhyatk/moondream1

As a usewr,
I want to del;ete projects,
So that I can get rid of faulty or old ones.

I expect a delete button on each product in the product view.
It should be a 2 step confirmation, so that I don't accidently delete a project.

As a user ,
I want a feedback button in the app,
So that I have an easy acces to the feedback form

I expect a button that opens a github issue where I can state my issue or request a feature

Objective: Create a comprehensive Inventory system within Captain that tracks and organizes all user-downloaded and user-created content, including apps, models, LoRAs, themes, and more.

Core Development Tasks for Inventory Enhancement:

• Design and implement the Inventory UI, ensuring it provides a clear, organized view of all user content.
• #125
• #126
• #127
• Enable detailed view options for items in the Inventory, providing users with information such as download date, source (marketplace or user-created), version, and usage instructions.
• #128
• Develop the ability to update items directly from the Inventory, including user notifications for available updates for both marketplace and user-created content.
• Implement a backup and export feature for the Inventory, allowing users to securely backup their Inventory items or export them for use in other environments.
• Ensure privacy and security are maintained for user-created and downloaded content, implementing appropriate access controls and encryption where necessary.
• Provide support for bulk actions within the Inventory, such as updating, deleting, or categorizing multiple items at once.

Describe the bug
When a model is downloaded as archive and then unpacked, the subpath contains the folder that was used when the model was actually packed. This leads to subpaths like this: \Captain_Data\downloads\stable-diffusion\checkpoints\stabilityai\sd-turbo\fp16\sd-turbo-fp16

Expected behavior
The model itself should be in the subpath as defined by Captain, not containing the folder that was used when it was packed: \Captain_Data\downloads\stable-diffusion\checkpoints\stabilityai\sd-turbo\fp16

Versions:

• OS: Windows 11
• Captain: 1.0.0-alpha.23

User Story

Title: Cost-Benefit Analysis for RunPod Subscription Models
As a product owner
I want to perform a cost-benefit analysis for different RunPod subscription models
So that we can choose the most efficient and cost-effective subscription model

Acceptance Criteria

• Identify and list all available RunPod subscription models.
• Analyze the cost implications of each subscription model.
• Evaluate the benefits, including but not limited to, computing power, availability, and support provided by each model.
• Create a comprehensive comparison chart or matrix that includes cost, benefits, and potential limitations for each model.
• Recommend the most suitable subscription model based on the analysis.

Potential Subtasks

• Conduct a market comparison to evaluate RunPod's subscription models against competitors, focusing on unique features and value propositions.
• Assess potential risks and mitigation strategies associated with each RunPod subscription model, including service downtimes and support issues.

Describe the bug
When opening a caption method with select element (open select) and none is selected.
And click away
Then the dropdown stays in place

Expected behavior

empty dropdown should not stay in place

Screenshots

Versions (please complete the following information):

• OS: Windows 11
• Captain v1.0.0-beat-9

Additional context

Describe the bug
The installer just starts installing the software without the users input

Expected behavior
Inform the user of where the program will be installed, and let the user install it wherever they want.
Additionally, there needs to be a functioning CANCEL button, right now you have even disabled the X-button.
This is beyond scummy and looks very suspicious.
There's no way I let something like this finish installing.

Screenshots

User Story

Title: Offline-Accessible Documentation and Support for Marketplace
As a Captain user
I want to access documentation and support resources for the marketplace and its apps without needing an internet connection
So that I can get help, learn about app features, and troubleshoot issues independently, enhancing my experience with the platform, even while offline

Acceptance Criteria

• The marketplace includes a comprehensive set of documentation and support resources, which are downloadable for offline access.
• Users can easily download and update these resources from within the app interface when online, preparing them for offline use.
• The offline documentation is searchable and well-organized, making it easy for users to find the information they need.

Is your feature request related to a problem? Please describe. A clear and concise description
The live-painting process should provide some kind of API, so that we can see if the process is still running. This comes in handy, when we go from the live-painting screen to somewhere else and back to live-painting.

User Story

Title: Develop File Preview and Display in Captain

As a user of Captain,

I want to preview files directly from the Explorer,

So that I can quickly view the contents of files, including markdown and images, with enhanced viewing features such as zooming for images.

Acceptance Criteria (Behavior Outline)

Feature: File Preview and Display

  Scenario: Previewing markdown files
    Given the user selects a markdown file in the Explorer
    When the user chooses to preview the file
    Then the markdown file is rendered in a readable format in the Preview app

  Scenario: Viewing images with zoom capabilities
    Given the user selects an image file in the Explorer
    When the user opens the file
    Then the image is displayed with options to zoom in and out in the Preview app

  Scenario: Managing open preview windows
    Given the user has multiple files open in preview mode
    When the user navigates between different preview windows
    Then the Explorer maintains the state of each previewed file, allowing seamless switching

Potential Subtasks

• Design the UI/UX for an integrated file preview feature within the Explorer, focusing on ease of use and quick access.
• Implement frontend functionality to render markdown files and display images, including interactive zoom controls.
• Develop backend support to handle file previews efficiently, ensuring that files are loaded and displayed quickly.
• Ensure responsive feedback and smooth interactions in the UI, particularly when users switch between previewed files or adjust viewing settings.

There is currently no way of generating captions for only uncaptioned or selected images.

1. allow captioning all uncaptioned
2. allow multi selection to caption

The settings should be available on all methods

Make GPT-Vision asynchronous. Currentlty batches wait for another, we can just run all of them at the same time which would exponentially rduce the waiting time.

Example:
100 images (batchsize of 10) is 10 rounds at 1 minute per round = 10 minutes

If we just run them all at once it should only take 1 minute to complete all.

(fictional duration above for simplicity)

As a safety measure we could also limit the calls to N, So I guess a batch of batches 😸

Objective: Implement an Explorer feature within Captain that allows users to efficiently browse files, with a focus on integrating collections and utilizing the vector store for data access.

Core Development Tasks for Explorer Enhancement:

• Design and implement the UI for the Explorer feature, ensuring it's intuitive and aligns with user expectations from traditional file explorers.
• #118
• Integrate "Collections" within the Explorer, allowing users to organize files into customizable groupings for easy access.
• #119
• #120
• #121
• Create a tagging system for files and collections to enhance organization and retrieval.
• #122
• #123
• Develop a system for tracking and displaying recently accessed or modified files to aid in user productivity.

Describe the bug
In very rare cases, the output image is broken for one frame.

Expected behavior
The output is never broken.

• OS: Windows 11
• Captain: 1.0.0-alpha.5

Objective: Create a dynamic and informative Dashboard within Captain that provides users with insights, updates, and quick access to their preferred AI tools and resources.

Core Development Tasks for Dashboard Enhancement:

• Design and implement a "Most Used Apps" list to showcase frequently accessed applications by the user.
• Integrate a feed of "New Releases from Marketplace" to highlight the latest additions to the marketplace.
• Feature "Hot Models" to draw attention to popular or trending AI models within the platform.
• #114
• #115
• #116
• Ensure that the Dashboard dynamically updates to reflect the user's interactions and preferences over time.
• Design the Dashboard UI to be intuitive and user-friendly, allowing for customization and personalization by the user.

Describe the bug
When running the installer again after Captain was already installed, the installer doesn't finish. Instead it shows the message "Can't uninstall old application files" after some time.

Expected behavior
I can run the installer as many times as I want, especially when there is a new version of Captain and the setup completes.

Versions:

• OS: Windows 11
• Captain: v1.0.0-beta.9

Objective: Enhance the Vector Store within Captain to handle a diverse range of data types and enable efficient search and categorization.

Core Development Tasks for Vector Store Enhancement:

• #110
• #111
• #112
• #308
• #310
• #309
• #311
• #312
• #313
• #314

Doesn't tell me what it is doing while the spinning thing is happening for captioning. Bad.

Maybe it is being denied and the program is in lala land not handling a case like that?

Version: 1.0.0-beta.4

User Story

Title: Estimating Development Effort for RunPod GPU-Cloud Integration
As a developer
I want to estimate the development effort and identify potential technical challenges for integrating RunPod's GPU-cloud services
So that we can plan our development sprints accurately and allocate resources effectively

Acceptance Criteria

• Develop a detailed breakdown of the tasks required for the integration.
• Estimate the time and resources needed for each task.
• Identify potential technical challenges that could impact the development timeline.

Is your feature request related to a problem? Please describe.
When I go into live-painting and have not added my OpenAI API key yet, I can't generate a story.

Describe the solution you'd like
It should be visible to the user, that this feature requires the OpenAI API key.

Objective: Establish an offline-first, privacy-centric marketplace within Captain for users to discover and manage AI applications and tools.

Core Development Tasks:

• #101
• #102
• Design and implement an offline-capable UI for browsing marketplace offerings.
• Create a local recommendation engine based on user-installed apps and usage.
• Design an app submission protocol that respects privacy and does not require online accounts.
• #103
• Draft and integrate a privacy policy for the marketplace that outlines the offline-first, no-tracking approach.
• #104
• #202

Describe the bug

When manually captioning an image, it get's overwritten, by the polling mechanism. We have 2 options (save on change, which might still cause errors, or prevent the focused/currently being edited caption from being updated by the polling info)

Expected behavior

When editing an image, I do not want the caption to be updated from sidk.

Versions (please complete the following information):

• OS: Windows 11
• Captain: 1.0.0-beta.5

As a user,
I want to use different guidelines for gpt-vision,
so that I can choose from a list to cover my use case

I expect to have a collection of guidelines * example responses.
I want to be able to easily switch between them
It would be nice if I could also share them with a friend

Describe the bug

The progres count during captioning is wrong. The button shows 111/21

Expected behavior

progress count should be correct

Screenshots

Versions (please complete the following information):

• OS: Windows 11
• Captain 1.0.0-beta.5

Additional context

User Story

Title: Enhance File and Collection Access with Vector Store

As a user of Captain,

I want the vector store to be utilized for retrieving data related to files and collections,

So that I can have quick and relevant access to the items I need.

Acceptance Criteria (Behavior Outline)

Feature: Utilize Vector Store for Enhanced Data Retrieval

  Scenario: Accessing files and collections
    Given the user queries for files or collections within Captain
    When the query is processed
    Then the vector store is utilized to retrieve and display the most relevant results based on the query

  Scenario: Enhanced retrieval performance
    Given a large number of files and collections are stored within Captain
    When the user performs a search or access operation
    Then the vector store facilitates quick and efficient access to the requested items

Potential Subtasks

• Implement integration between the Explorer and the vector store for query processing.
• Design a relevance algorithm or use existing vector store capabilities to rank search and access results.
• Optimize the data structure within the vector store for efficient file and collection retrieval.
• Conduct performance testing to ensure quick access times, even with large datasets.

As a user,
I want to rename a project,
So that I can decide later how to name it

Currently it is only possible to set a name upon creation. It would be helpful to adjsut it any time later

User Story

Title: Storage and Retrieval of Generated Content in Vector Store

As a user of Captain,

I want the Vector Store to support the storage and retrieval of generated content such as images,

So that I can efficiently manage and access my generated content directly within the platform.

Acceptance Criteria (Behavior Outline)

Feature: Support for Generated Content in Vector Store

  Scenario: Storing generated content
    Given a user generates content like images within Captain
    When the user decides to store this content in the Vector Store
    Then the content is stored with appropriate metadata for easy retrieval

  Scenario: Retrieving stored generated content
    Given generated content is stored in the Vector Store
    When the user queries the Vector Store for their content
    Then the requested content is retrieved and displayed to the user

  Scenario: Updating stored generated content
    Given existing generated content stored in the Vector Store
    When the user updates this content or its metadata
    Then the Vector Store updates the content and metadata accordingly

Potential Subtasks

• Develop a data model within the Vector Store that supports generated content, including images, with metadata tagging.
• Implement upload and storage mechanisms for handling large files and various content types.
• Create retrieval functions within the Vector Store that allow for efficient searching and access to generated content.
• Ensure that the system supports version control and metadata updates for stored content.
• Test storage, retrieval, and update processes with various types of generated content to ensure performance and accuracy.

User Story

Title: Enhanced Data Filtering and Categorization in Vector Store

As a developer using Captain,

I want to have robust filters and categorization functionality for data stored in the Vector Store,

So that I can efficiently manage and retrieve diverse data types based on specific attributes and categories.

Acceptance Criteria (Behavior Outline)

Feature: Enhanced Filtering and Categorization in Vector Store

  Scenario: Categorizing stored data
    Given a developer stores new data in the Vector Store
    When they specify categories and attributes for the data
    Then the Vector Store organizes the data according to the provided categories and attributes

  Scenario: Filtering data by category
    Given multiple categories of data are stored in the Vector Store
    When a user queries the Vector Store for a specific category
    Then the Vector Store retrieves only data that matches the specified category

  Scenario: Applying filters to data retrieval
    Given data with various attributes is stored in the Vector Store
    When a user applies filters based on those attributes
    Then the Vector Store returns only data that matches the specified filters

  Scenario: Updating data categorization and attributes
    Given existing data categorized in the Vector Store
    When a developer updates the data with new categories or attributes
    Then the Vector Store re-categorizes and updates the data accordingly

Potential Subtasks

• Implement a tagging system within the Vector Store for categorizing data upon entry.
• Develop a query interface that allows users to filter data based on categories and attributes.
• Create a dynamic update mechanism for changing categories and attributes of stored data.
• Ensure the user interface for data management supports easy categorization and application of filters.
• Document best practices for developers on how to use the new filtering and categorization functionalities.

As a user,
I want to use the app in my language,
So that I can better understand the interface.

Please offer some common languages like German, French, Spanish, Italian ...

This is also a great opportunity for Community support.

The app is already preconfigured with next-i18next :)

The progress in development seems quite promising, and our prototype has demonstrated its potential effectively. In order to avoid accruing technical debt and to enhance stability, it's important that we initiate the creation of a testing environment for behavior and end-to-end tests.

Here are some useful resources:

• Playwright's documentation on Electron testing: Playwright for Electron.
• An article about using Cypress for testing Electron.js applications: Testing Electron.js with Cypress.
• Electron's guide on automated testing: Electron Automated Testing.

I have a strong preference for Playwright, as it aligns well with our existing core technology stack.

User Story

Title: Technical Feasibility Study for Integrating RunPod's GPU-Cloud Services
As a developer
I want to conduct a technical feasibility study for integrating RunPod's GPU-cloud services
So that we can ensure compatibility and identify potential technical challenges or requirements before committing to integration

Acceptance Criteria

• Identify and document the technical requirements for integrating RunPod's GPU-cloud services with our existing systems.
• Evaluate our current infrastructure and software stack for compatibility with RunPod's services.
• Identify any potential technical challenges or limitations that could arise during integration.
• Assess the need for additional tools, libraries, or frameworks to facilitate integration.
• Develop a preliminary integration plan outlining the steps, resources, and timeline.

Potential Subtasks

• #183

As a user,
I want to be informed about hotkeys,
So that I know my options.

I expect to see an info icon or similar which shows the available hotkeys as legend

Is your feature request related to a problem? Please describe. A clear and concise description
of what the problem is. Ex. I'm always frustrated when [...]
There is no way to prefix or suffix captions

Describe the solution you'd like A clear and concise description of what you want to happen.
I want to be able to prefix or suffix all captions of a dataset (e.g. a trigger word or style name)

Describe alternatives you've considered A clear and concise description of any alternative
solutions or features you've considered.

currently only manual editing is an option, and it does not scale

Additional context Add any other context or screenshots about the feature request here.

User Story

Title: Language-Specific Data Handling in Vector Store

As a developer using Captain,

I want the Vector Store to handle data in a language-specific manner,

So that I can ensure improved localization and relevance of content for users in different linguistic regions.

Acceptance Criteria (Behavior Outline)

Feature: Language-Specific Data Handling in Vector Store

  Scenario: Storing language-specific data
    Given a developer wants to store data in the Vector Store
    When they specify the language of the data
    Then the Vector Store categorizes and stores the data according to its language

  Scenario: Retrieving language-specific data
    Given data in multiple languages is stored in the Vector Store
    When a user queries the Vector Store and specifies a language
    Then only data relevant to the specified language is retrieved

  Scenario: Updating language-specific data entries
    Given existing language-specific data in the Vector Store
    When a developer updates the data with language-specific modifications
    Then the Vector Store updates the data while retaining its language-specific categorization

  Scenario: Language fallback mechanism
    Given a user queries the Vector Store in a language with limited data available
    When there is not enough data in the specified language
    Then the Vector Store offers data in a fallback language with a notification about the language substitution

Potential Subtasks

• Develop a schema for the Vector Store that includes language as a key attribute for data.
• Implement functionality to tag and categorize incoming data by language during the storage process.
• Create a query system in the Vector Store that filters results based on the specified language.
• Design a language fallback mechanism for scenarios where requested language-specific data is sparse or unavailable.
• Update documentation for the Vector Store to guide developers on how to utilize language-specific data handling features.

User Story

Title: Implement Search Functionality in Explorer

As a user of Captain,

I want to search within the Explorer using the vector store,

So that I can find my files and content quickly and accurately.

Acceptance Criteria (Behavior Outline)

Feature: Explorer Search Functionality

  Scenario: User performs a search in the Explorer
    Given the user is in the Explorer
    When the user enters a search query
    Then the Explorer uses the vector store to retrieve and display relevant search results

  Scenario: Viewing search results
    Given the user has entered a search query in the Explorer
    When the search results are displayed
    Then the results include files and content that match the query based on vector store analysis

Potential Subtasks

• Design and integrate a search bar in the Explorer interface.
• Develop the backend logic to process search queries using the vector store.
• Implement a results display that ranks content based on relevance to the search query.
• Test the search functionality with various types of content to ensure accuracy and performance.
  ficantly improving user experience and productivity.