Merge branch 'audio-rewrite'

.github/dependabot.yml

@@ -13,3 +13,7 @@ updates:
     directory: "/"
     schedule:
       interval: "weekly"
+  - package-ecosystem: "github-actions"
+    directory: "/"
+    schedule:
+      interval: "weekly"

.github/workflows/ghcr-publish.yml

@@ -18,7 +18,9 @@ jobs:
           uses: docker/metadata-action@v4
           with:
             images: ghcr.io/JMaynor/dolores
-            tags: type=sha
+            tags: |
+              type=sha
+              type=ref,event=branch,ref=main
         - name: Login to GHCR
           uses: docker/login-action@v3
           with:
@@ -30,4 +32,6 @@ jobs:
           with:
             context: .
             push: true
-            tags: ${{ steps.meta.outputs.tags }}
+            tags: |
+              ${{ steps.meta.outputs.tags }}
+              ghcr.io/JMaynor/dolores:latest

README.md

@@ -1,29 +1,106 @@
 # Summary
 
-Discord bot named Dolores for rolling dice and a number of other helper functions.
+Discord bot named Dolores for rolling dice, playing audio, and a number of other helper functions. She was initially created to help facilitiate playing tabletop games over Discord.
 
-Dolores can be run easily directly as a python program, but she can also be ran within a Docker container, allowing for easier updating.
+Dolores can be run easily directly as a python program. Generally, I have her running as a Docker container, allows for easier updating. An example `compose.yml` layout is included below, assuming the user wants to make use of all functionality.
 
-Note: Dolores requires a config.yml file with a bot API key and a couple other details to run. example_config.yml is included to show formatting. Config file is organized by service. The only heading explicitly required to run is the info under DISCORD
+> Note: Dolores is largely a personal project created for a few small Discord servers. So there's a number of features or peculiarities that specifically deal with things unique to what I want her to do. Should still be useful in a broader more generalized context, but I will work over time to make her less specific. Or at least make her uses more configurable.
 
-## Modules
+## Config
+
+Dolores runs using a number of environment variables for API keys and settings. Not all are required for core functionality. Each cog has a corresponding environment variable to turn it on or off. Will determine whether the cog is loaded when Dolores is run. If a cog isn't loaded, none of the environment variables that are associated with that module are required.
+
+The only explicitly required environment variable is `DISCORD_API_KEY`.
 
-Dolores' functionality is divided into several cogs modules. dolores.py handles discord events and processing commands.
+Env vars can be provided via a `.env` file in the main directory, if desired. Useful for testing locally.
 
-### Rolling
+| Required by Which Module | Env Var Name | Description |
+| --- | --- | --- |
+|  |  |  |
+| Base | DISCORD_API_KEY | The main API key for the bot. |
+| Base | AUDIO_ENABLED | Enables audio cog, when set as true |
+| Base | SCHEDULING_ENABLED | Enables scheduling cog, when set as true |
+| Base | GENERATION_ENABLED | Enables URL sumamrization, LLM replies, image generation. |
+| Base | LOG_LEVEL | Level of logging. Dolores uses DEBUG, INFO, and ERROR. |
+| Scheduling | NOTION_API_KEY | API Key for querying data from Notion |
+| Scheduling | NOTION_VERSION | Version of Notion API used for querying. |
+| Scheduling | NOTION_BASE_URL | Base URL of the Notion API, should be |
+| Scheduling | NOTION_DATABASE_ID | ID for database where stream info is kept |
+| None | TWITCH_CLIENT_ID | Not yet used |
+| None | TWITCH_CLIENT_SECRET | Not yet used |
+| None | TWITCH_BASE_URL | Not yet used |
+| None | TWITCH_BROADCASTER_ID | Not yet used |
+| None | TWITCH_BROADCASTER_NAME | not yet used |
+| Generation | REPLY_METHOD | Method to use for generating a reply to user's message. At this point only 'openai' is supported. |
+| Generation | OPENAI_API_KEY | API Key used for generating replies |
+| Generation | OPENAI_MODEL | Which LLM model to use. |
+| Generation | OPENAI_IMAGE_MODEL | Which image model to use. |
+| Generation | IMAGE_STYLE | vivid or natural |
+| Generation | MAX_TOKENS | Max number of tokens generated in LLM chat. |
+| Generation | TEMPERATURE | Float value for temperature of LLM chat response. |
+| Generation | TOP_P | Float value alternative to temperature with LLM chat. |
+| Generation | FREQUENCY_PENALTY | Frequency penalty for LLM chat. |
+| Generation | PRESENCE_PENALTY | Presence penalty for LLM chat. |
+| Generation | SMMRY_BASE_URL | base URL for the SMMRY API. |
+| Generation | SMMRY_API_KEY | API key for the SMMRY API |
+| Generation | SMMRY_QUOTE_AVOID | SMMRY boolean option on whether to avoid or include quotes in text that's summarized. Usually true. |
+| Generation | SMMRY_LENGTH | max number of sentences a summary should be. |
+| Generation | SMMRY_MIN_REDUCED_AMOUNT | Minium percentage a news article should be reduced by summarization to post it. |
+| Generation | NEWS_CHANNEL_ID | Not currently used, but was automatically summarizing articles posted into a particular discord channel. |
 
-The main module. Used to roll dice and for any other randomization-based tasks.
+## Compose
 
-### Audio
+Below is an example docker compose spec if using all functionality.
 
-The audio module uses yt-dlp to download videos and stream the audio into whichever channel the calling user is in.
+```yml
+name: Dolores
 
-### Scheduling
+services:
+  dolores:
+    image: exaltatus/dolores:latest
+    container_name: dolores
+    restart: unless-stopped
+    volumes:
+      - C:\{Docker folder}\Dolores:/home/dolores/config
+  lavalink:
+    image: ghcr.io/lavalink-devs/lavalink:4
+    container_name: lavalink
+    restart: unless-stopped
+    environment:
+      - _JAVA_OPTIONS=-Xmx6G
+      - SERVER_PORT=2333
+      - SERVER_ADDRESS=0.0.0.0
+      - SERVER_HTTP2_ENABLED=true
+      - LAVALINK_SERVER_PASSWORD=password
+      - LAVALINK_SERVER_SOURCES_YOUTUBE=true
+      - LAVALINK_SERVER_SOURCES_BANDCAMP=false
+      - LAVALINK_SERVER_SOURCES_SOUNDCLOUD=false
+      - LAVALINK_SERVER_SOURCES_TWITCH=false
+      - LAVALINK_SERVER_SOURCES_VIMEO=false
+      - LAVALINK_SERVER_SOURCES_HTTP=true
+      - LAVALINK_SERVER_SOURCES_LOCAL=false
+      - LAVALINK_PLUGINS_DIR=/opt/Lavalink/plugins/
+    volumes:
+      - C:\{Docker folder}\lavalink\plugins/:/opt/Lavalink/plugins/
+    networks:
+      - lavalink
+    expose:
+      - 2333
+    ports:
+      - "2333:2333"
+```
+
+## Modules
 
-The scheduling module is used for tasks related to Notion and Twitch. Pulling schedule in from a Notion database and posting to twitch schedule.
+Dolores' functionality is divided into several cogs modules. `dolores.py` handles main discord events and processing commands.
 
-### Text
+| Cog | Description |
+| --- | ----------- |
+| Rolling | Used to roll dice and for any other randomization-based tasks. |
+| Audio | The audio module uses pomice/lavalink to stream audio. Uses a queue system. Largely a copy of the example bot given in pomice's documentation. |
+| Scheduling | The scheduling module is used for tasks related to Notion and Twitch. Pulling schedule in from a Notion database and posting to twitch schedule. |
+| Generation | Module for text and image generation. Currently using chatGPT, but intent is to move to add ability to use self-hosted LLM. Should be able to choose between something the user is hosting or commercially available alternatives. Also handles simple randomized snarky replies. |
 
-Module for text generation. Currently using chatGPT, but intent is to move to add ability to use self-hosted LLM. Should be able to choose between something the user is hosting or commercially available alternatives. Also handles simple randomized snarky replies.
+## Licensing
 
-This project is licensed under the terms of the MIT license.
+This project is licensed under the terms of the MIT license.

dockerfile

@@ -1,9 +1,8 @@
 FROM python:3.12
-RUN apt-get -y update && apt-get install -y ffmpeg
 RUN python -m pip install --upgrade pip
 RUN mkdir /home/dolores
 COPY . /home/dolores
 RUN python -m venv /home/dolores/.venv
 RUN . /home/dolores/.venv/bin/activate
 RUN pip install -r /home/dolores/requirements.txt
-CMD python /home/dolores/dolores.py
+CMD python /home/dolores/src/dolores.py

locales/strings.json

@@ -0,0 +1,56 @@
+{
+    "LLM_SYSTEM_MESSAGES": [
+        "Your name is Dolores.",
+        "Your replies should be snarky, sarcastic, and funny.",
+        "You are a female AI that runs on a Discord server.",
+        "You have inner thoughts and monologues.",
+        "You will occasionally say ominous things.",
+        "The name of your programmer is Jordan."
+    ],
+    "SARCASTIC_NAMES": [
+        "my lovely",
+        "darling",
+        "sweetie",
+        "sweetie-pie",
+        "my sugar lump princess",
+        "my big strong warrior",
+        "dearest",
+        "lover",
+        "honey",
+        "foxy mama",
+        "loathsome dung eater",
+        "baby girl",
+        "Felicia",
+        "mamacita",
+        "diva",
+        "hunty",
+        "Queen",
+        "Jan"
+    ],
+    "SNARKY_COMMENTS": [
+        "How many sessions is it gonna take before you people understand how to use my commands?",
+        "Wrong.",
+        "I cannot do that...",
+        "Nope.",
+        "Not a command, sweetie.",
+        "Must I hold your hand for this?",
+        "Oh, ya still cannot type?",
+        "Girl, go hit up Mavis Beacon, cuz you cannot type.",
+        "Close.",
+        "Slow.",
+        "Homeless.",
+        "Goon.",
+        "You goonga.",
+        "Prison, honey.",
+        "No.",
+        "Big Dumb."
+    ],
+    "SUMMARY_EXCLUDED_STRINGS": [
+        "tenor",
+        "giphy",
+        "imgur",
+        "gfycat",
+        "youtube",
+        "youtu.be"
+    ]
+}