FeedbackFlow

Overview

FeedbackFlow is a set of tools designed to collect and consolidate feedback from GitHub issues, discussions, and YouTube comments into a machine-readable format (JSON). With dedicated components for different platforms, FeedbackFlow simplifies the process of centralizing feedback for analysis and decision-making.

Components

GitHub Feedback Collector (ghdump)

This tool retrieves data from GitHub issues and discussions, including associated comments. The collected information is saved as JSON files, enabling easy integration with other tools.

YouTube Comment Collector (ytdump)

This tool gathers comments from specified YouTube videos or playlists and exports them into JSON format. Input sources can be configured through a file or directly via command-line arguments.

Getting Started

Prerequisites

To use FeedbackFlow, ensure you have the following:

• .NET 9.0 SDK installed
• A valid GitHub API Key
• A valid YouTube API Key

Installation

1. Clone the repository:

   git clone https://github.com/davidfowl/feedbackflow.git
   cd feedbackflow

2. Restore dependencies:

   dotnet restore

3. Build the project:

   dotnet build

Usage

GitHub Feedback Collector (ghdump)

To collect feedback from GitHub:

./ghdump -r <repository>

YouTube Comment Collector (ytdump)

To gather comments from YouTube:

./ytdump -v <video-id> -o <output-file.json>

Azure Functions Configuration

Local Development Setup

To run the Azure Functions project locally, you'll need to configure the local.settings.json file in the feedbackfunctions directory. Create the file with the following structure:

{
  "IsEncrypted": false,
  "Values": {
    "AzureWebJobsStorage": "UseDevelopmentStorage=true",
    "FUNCTIONS_WORKER_RUNTIME": "dotnet-isolated",
    "GitHub:AccessToken": "your_github_pat_here",
    "YouTube:ApiKey": "your_youtube_api_key_here",
    "Azure:OpenAI:Endpoint": "your_azure_openai_endpoint",
    "Azure:OpenAI:ApiKey": "your_azure_openai_key",
    "Azure:OpenAI:Deployment": "your_model_deployment_name"
  }
}

Required API Keys and Configuration

1. GitHub Personal Access Token (PAT)

   • Create a GitHub PAT with repo scope
   • Set it in GitHub:AccessToken

2. YouTube API Key

   • Create a project in Google Cloud Console
   • Enable YouTube Data API v3
   • Create API credentials
   • Set the key in YouTube:ApiKey

3. Azure OpenAI Configuration

   • Create an Azure OpenAI resource
   • Set the endpoint URL in Azure:OpenAI:Endpoint
   • Set the API key in Azure:OpenAI:ApiKey
   • Deploy a model and set its name in Azure:OpenAI:Deployment

4. Azure Storage Emulator

   • Install Azure Storage Emulator for local development
   • The default connection string is already set in AzureWebJobsStorage

Running the Functions

After configuring the settings:

cd feedbackfunctions
func start

Note: Keep your API keys and tokens secure and never commit them to source control.

Azure Functions local.settings.json

You will need the following:

{
   "IsEncrypted": false,
   "Values": {
      "AzureWebJobsStorage": "UseDevelopmentStorage=true",
      "FUNCTIONS_WORKER_RUNTIME": "dotnet-isolated"
   },
   "YouTube:ApiKey": "YOUR_API_KEY_HERE",
   "Reddit:ClientId": "YOUR_REDDIT_CLIENT_ID",
   "Reddit:ClientSecret": "YOUR_REDDIT_CLIENT_SECRET", 
   "GitHub:AccessToken": "YOUR_ACCESS_TOKEN_HERE",
   "Azure:OpenAI:Endpoint": "YOUR_AZURE_OPENAI_ENDPOINT",
   "Azure:OpenAI:ApiKey": "YOUR_AZURE_OPENAI_API_KEY",
   "Azure:OpenAI:Deployment": "YOUR_DEPLOYMENT_NAME",
   "Twitter:BearerToken": "YOUR_TWITTER_BEARER_TOKEN",
   "BlueSky:Username": "YOUR_BLUESKY_USERNAME",
   "BlueSky:AppPassword": "YOUR_BLUESKY_APP_PASSWORD"
}

License

This project is licensed under the MIT License. See the LICENSE file for details.