Merge pull request #125 from andrewnguonly/1.0.10

PR for app version `1.0.10`
andrewnguonly · Mar 8, 2024 · 5c02dfd · 5c02dfd
2 parents 44c5a4a + 9ec1e3b
commit 5c02dfd
Show file tree

Hide file tree

Showing 15 changed files with 483 additions and 134 deletions.
diff --git a/README.md b/README.md
@@ -4,7 +4,7 @@
 
 A RAG LLM co-pilot for browsing the web, powered by local LLMs.
 
-![Screenshot of Lumos](./screenshots/lumos_screenshot_2.png)
+![Screenshot of Lumos](./screenshots/lumos_screenshot_3.png)
 
 This Chrome extension is powered by [Ollama](https://ollama.ai/). Inference is done on your local machine without any _external_ server support. However, due to security constraints in the Chrome extension platform, the app does rely on _local_ server support to run the LLM. This app is inspired by the [Chrome extension example](https://github.com/mlc-ai/web-llm/tree/main/examples/chrome-extension) provided by the [Web LLM project](https://webllm.mlc.ai/) and the [local LLM examples](https://js.langchain.com/docs/use_cases/question_answering/local_retrieval_qa) provided by [LangChain](https://github.com/langchain-ai/langchainjs).
 
@@ -98,6 +98,7 @@ Right-click on the extension icon and select `Options` to access the extension's
 - **Ollama Host**: Select desired host (defaults to `http://0.0.0.0:11434`)
 - **Vector Store TTL (minutes)**: Number of minutes to store a URL's content in the vector store cache.
 - **Content Parser Config**: Lumos's default content parser will extract all text content between a page's `<body></body>` tag. To customize the content parser, add an entry to the configuration.
+- **Enable/Disable Tools**: Enable or disable individual tools. If a tool is enabled, a custom prefix trigger (e.g. "calc:") can be specified to override the app's internal prompt classification mechanism.
 
 ### Content Parser Config
 
@@ -194,6 +195,13 @@ Alternatively, if content is highlighted on a page (e.g. highlighted text), that
 
 Note: Content that is highlighted will not be cached in the vector store cache. Each subsequent prompt containing highlighted content will generate new embeddings.
 
+## Shortcuts
+
+- `cmd + c`: Copy last message to clipboard.
+- `cmd + j`: Toggle `Disable content parsing` checkbox.
+- `cmd + k`: Clear all messages.
+- `cmd + ;`: Open/close Chat History panel.
+
 ## Multimodal
 
 Lumos supports multimodal models! Images that are present on the current page will be downloaded and bound to the model for prompting. See documentation and examples [here](./docs/multimodal.md).

diff --git a/docs/edge.md b/docs/edge.md
@@ -0,0 +1,16 @@
+# Steps to Install on Microsoft Edge
+
+**Warning**: Lumos was originally developed as a Chrome extension. Some features may not be compatible with Microsoft Edge.
+
+## Start Ollama Server
+
+Example:
+```
+OLLAMA_ORIGINS=chrome-extension://* ollama serve
+```
+
+Note: The host protocol is still `chrome-extension://` despite the extension running on Microsoft Edge.
+
+## Install Lumos
+
+Follow instructions to [Sideload an extension](https://learn.microsoft.com/en-us/microsoft-edge/extensions-chromium/getting-started/extension-sideloading) on Microsoft Edge.
diff --git a/docs/linux.md b/docs/linux.md
@@ -0,0 +1,24 @@
+# Linux
+
+## Add `OLLAMA_ORIGINS` Environment Variable to Ollama Startup Service
+
+1. Follow [Ollama's instructions for adding it as a startup service](https://github.com/ollama/ollama/blob/main/docs/linux.md#adding-ollama-as-a-startup-service-recommended).
+2. In the service file, set `Environment` to `"OLLAMA_ORIGINS=chrome-extension://*"`.
+
+Example service file:
+```
+[Unit]
+Description=Ollama Service
+After=network-online.target
+
+[Service]
+Environment="OLLAMA_ORIGINS=chrome-extension://*"
+ExecStart=/usr/bin/ollama serve
+User=ollama
+Group=ollama
+Restart=always
+RestartSec=3
+
+[Install]
+WantedBy=default.target
+```
diff --git a/manifest.json b/manifest.json
@@ -1,5 +1,5 @@
 {
-    "version": "1.0.9",
+    "version": "1.0.10",
     "manifest_version": 3,
     "name": "Lumos",
     "description": "An LLM co-pilot for browsing the web, powered by local LLMs. Your prompts never leave the browser.",

diff --git a/package-lock.json b/package-lock.json
diff --git a/package.json b/package.json
@@ -1,6 +1,6 @@
 {
   "name": "lumos",
-  "version": "1.0.9",
+  "version": "1.0.10",
   "private": true,
   "dependencies": {
     "@chatscope/chat-ui-kit-react": "^1.10.1",
@@ -28,6 +28,7 @@
     "react-syntax-highlighter": "^15.5.0",
     "ts-loader": "^9.5.0",
     "typescript": "^4.9.5",
+    "uuid": "^9.0.1",
     "web-vitals": "^2.1.4"
   },
   "scripts": {

diff --git a/screenshots/lumos_screenshot_2.png b/screenshots/lumos_screenshot_2.png
diff --git a/screenshots/lumos_screenshot_3.png b/screenshots/lumos_screenshot_3.png