diff --git a/ui/.eslintrc.json b/.eslintrc.json similarity index 100% rename from ui/.eslintrc.json rename to .eslintrc.json diff --git a/.github/workflows/docker-build.yaml b/.github/workflows/docker-build.yaml index f658c29..29f7987 100644 --- a/.github/workflows/docker-build.yaml +++ b/.github/workflows/docker-build.yaml @@ -8,18 +8,12 @@ on: types: [published] jobs: - build-and-push: + build-amd64: runs-on: ubuntu-latest - strategy: - matrix: - service: [backend, app] steps: - name: Checkout code uses: actions/checkout@v3 - - name: Set up QEMU - uses: docker/setup-qemu-action@v2 - - name: Set up Docker Buildx uses: docker/setup-buildx-action@v2 with: @@ -36,38 +30,109 @@ jobs: id: version run: echo "RELEASE_VERSION=${GITHUB_REF#refs/tags/}" >> $GITHUB_ENV - - name: Build and push Docker image for ${{ matrix.service }} + - name: Build and push AMD64 Docker image if: github.ref == 'refs/heads/master' && github.event_name == 'push' run: | - docker buildx create --use - if [[ "${{ matrix.service }}" == "backend" ]]; then \ - DOCKERFILE=backend.dockerfile; \ - IMAGE_NAME=perplexica-backend; \ - else \ - DOCKERFILE=app.dockerfile; \ - IMAGE_NAME=perplexica-frontend; \ - fi - docker buildx build --platform linux/amd64,linux/arm64 \ - --cache-from=type=registry,ref=itzcrazykns1337/${IMAGE_NAME}:main \ + DOCKERFILE=app.dockerfile + IMAGE_NAME=perplexica + docker buildx build --platform linux/amd64 \ + --cache-from=type=registry,ref=itzcrazykns1337/${IMAGE_NAME}:amd64 \ --cache-to=type=inline \ + --provenance false \ -f $DOCKERFILE \ - -t itzcrazykns1337/${IMAGE_NAME}:main \ + -t itzcrazykns1337/${IMAGE_NAME}:amd64 \ --push . - - name: Build and push release Docker image for ${{ matrix.service }} + - name: Build and push AMD64 release Docker image if: github.event_name == 'release' run: | - docker buildx create --use - if [[ "${{ matrix.service }}" == "backend" ]]; then \ - DOCKERFILE=backend.dockerfile; \ - IMAGE_NAME=perplexica-backend; \ - else \ - DOCKERFILE=app.dockerfile; \ - IMAGE_NAME=perplexica-frontend; \ - fi - docker buildx build --platform linux/amd64,linux/arm64 \ - --cache-from=type=registry,ref=itzcrazykns1337/${IMAGE_NAME}:${{ env.RELEASE_VERSION }} \ + DOCKERFILE=app.dockerfile + IMAGE_NAME=perplexica + docker buildx build --platform linux/amd64 \ + --cache-from=type=registry,ref=itzcrazykns1337/${IMAGE_NAME}:${{ env.RELEASE_VERSION }}-amd64 \ --cache-to=type=inline \ + --provenance false \ -f $DOCKERFILE \ - -t itzcrazykns1337/${IMAGE_NAME}:${{ env.RELEASE_VERSION }} \ + -t itzcrazykns1337/${IMAGE_NAME}:${{ env.RELEASE_VERSION }}-amd64 \ --push . + + build-arm64: + runs-on: ubuntu-24.04-arm + steps: + - name: Checkout code + uses: actions/checkout@v3 + + - name: Set up Docker Buildx + uses: docker/setup-buildx-action@v2 + with: + install: true + + - name: Log in to DockerHub + uses: docker/login-action@v2 + with: + username: ${{ secrets.DOCKER_USERNAME }} + password: ${{ secrets.DOCKER_PASSWORD }} + + - name: Extract version from release tag + if: github.event_name == 'release' + id: version + run: echo "RELEASE_VERSION=${GITHUB_REF#refs/tags/}" >> $GITHUB_ENV + + - name: Build and push ARM64 Docker image + if: github.ref == 'refs/heads/master' && github.event_name == 'push' + run: | + DOCKERFILE=app.dockerfile + IMAGE_NAME=perplexica + docker buildx build --platform linux/arm64 \ + --cache-from=type=registry,ref=itzcrazykns1337/${IMAGE_NAME}:arm64 \ + --cache-to=type=inline \ + --provenance false \ + -f $DOCKERFILE \ + -t itzcrazykns1337/${IMAGE_NAME}:arm64 \ + --push . + + - name: Build and push ARM64 release Docker image + if: github.event_name == 'release' + run: | + DOCKERFILE=app.dockerfile + IMAGE_NAME=perplexica + docker buildx build --platform linux/arm64 \ + --cache-from=type=registry,ref=itzcrazykns1337/${IMAGE_NAME}:${{ env.RELEASE_VERSION }}-arm64 \ + --cache-to=type=inline \ + --provenance false \ + -f $DOCKERFILE \ + -t itzcrazykns1337/${IMAGE_NAME}:${{ env.RELEASE_VERSION }}-arm64 \ + --push . + + manifest: + needs: [build-amd64, build-arm64] + runs-on: ubuntu-latest + steps: + - name: Log in to DockerHub + uses: docker/login-action@v2 + with: + username: ${{ secrets.DOCKER_USERNAME }} + password: ${{ secrets.DOCKER_PASSWORD }} + + - name: Extract version from release tag + if: github.event_name == 'release' + id: version + run: echo "RELEASE_VERSION=${GITHUB_REF#refs/tags/}" >> $GITHUB_ENV + + - name: Create and push multi-arch manifest for main + if: github.ref == 'refs/heads/master' && github.event_name == 'push' + run: | + IMAGE_NAME=perplexica + docker manifest create itzcrazykns1337/${IMAGE_NAME}:main \ + --amend itzcrazykns1337/${IMAGE_NAME}:amd64 \ + --amend itzcrazykns1337/${IMAGE_NAME}:arm64 + docker manifest push itzcrazykns1337/${IMAGE_NAME}:main + + - name: Create and push multi-arch manifest for releases + if: github.event_name == 'release' + run: | + IMAGE_NAME=perplexica + docker manifest create itzcrazykns1337/${IMAGE_NAME}:${{ env.RELEASE_VERSION }} \ + --amend itzcrazykns1337/${IMAGE_NAME}:${{ env.RELEASE_VERSION }}-amd64 \ + --amend itzcrazykns1337/${IMAGE_NAME}:${{ env.RELEASE_VERSION }}-arm64 + docker manifest push itzcrazykns1337/${IMAGE_NAME}:${{ env.RELEASE_VERSION }} diff --git a/.gitignore b/.gitignore index 8391d19..c95173d 100644 --- a/.gitignore +++ b/.gitignore @@ -4,9 +4,9 @@ npm-debug.log yarn-error.log # Build output -/.next/ -/out/ -/dist/ +.next/ +out/ +dist/ # IDE/Editor specific .vscode/ diff --git a/.prettierrc.js b/.prettierrc.js index 1937ff1..8ca480f 100644 --- a/.prettierrc.js +++ b/.prettierrc.js @@ -6,7 +6,6 @@ const config = { endOfLine: 'auto', singleQuote: true, tabWidth: 2, - semi: true, }; module.exports = config; diff --git a/CONTRIBUTING.md b/CONTRIBUTING.md index b16eccf..7bbb280 100644 --- a/CONTRIBUTING.md +++ b/CONTRIBUTING.md @@ -1,32 +1,43 @@ # How to Contribute to Perplexica -Hey there, thanks for deciding to contribute to Perplexica. Anything you help with will support the development of Perplexica and will make it better. Let's walk you through the key aspects to ensure your contributions are effective and in harmony with the project's setup. +Thanks for your interest in contributing to Perplexica! Your help makes this project better. This guide explains how to contribute effectively. + +Perplexica is a modern AI chat application with advanced search capabilities. ## Project Structure -Perplexica's design consists of two main domains: +Perplexica's codebase is organized as follows: -- **Frontend (`ui` directory)**: This is a Next.js application holding all user interface components. It's a self-contained environment that manages everything the user interacts with. -- **Backend (root and `src` directory)**: The backend logic is situated in the `src` folder, but the root directory holds the main `package.json` for backend dependency management. - - All of the focus modes are created using the Meta Search Agent class present in `src/search/metaSearchAgent.ts`. The main logic behind Perplexica lies there. +- **UI Components and Pages**: + - **Components (`src/components`)**: Reusable UI components. + - **Pages and Routes (`src/app`)**: Next.js app directory structure with page components. + - Main app routes include: home (`/`), chat (`/c`), discover (`/discover`), library (`/library`), and settings (`/settings`). + - **API Routes (`src/app/api`)**: API endpoints implemented with Next.js API routes. + - `/api/chat`: Handles chat interactions. + - `/api/search`: Provides direct access to Perplexica's search capabilities. + - Other endpoints for models, files, and suggestions. +- **Backend Logic (`src/lib`)**: Contains all the backend functionality including search, database, and API logic. + - The search functionality is present inside `src/lib/search` directory. + - All of the focus modes are implemented using the Meta Search Agent class in `src/lib/search/metaSearchAgent.ts`. + - Database functionality is in `src/lib/db`. + - Chat model and embedding model providers are managed in `src/lib/providers`. + - Prompt templates and LLM chain definitions are in `src/lib/prompts` and `src/lib/chains` respectively. + +## API Documentation + +Perplexica exposes several API endpoints for programmatic access, including: + +- **Search API**: Access Perplexica's advanced search capabilities directly via the `/api/search` endpoint. For detailed documentation, see `docs/api/search.md`. ## Setting Up Your Environment Before diving into coding, setting up your local environment is key. Here's what you need to do: -### Backend - 1. In the root directory, locate the `sample.config.toml` file. -2. Rename it to `config.toml` and fill in the necessary configuration fields specific to the backend. -3. Run `npm install` to install dependencies. -4. Run `npm run db:push` to set up the local sqlite. -5. Use `npm run dev` to start the backend in development mode. - -### Frontend - -1. Navigate to the `ui` folder and repeat the process of renaming `.env.example` to `.env`, making sure to provide the frontend-specific variables. -2. Execute `npm install` within the `ui` directory to get the frontend dependencies ready. -3. Launch the frontend development server with `npm run dev`. +2. Rename it to `config.toml` and fill in the necessary configuration fields. +3. Run `npm install` to install all dependencies. +4. Run `npm run db:push` to set up the local sqlite database. +5. Use `npm run dev` to start the application in development mode. **Please note**: Docker configurations are present for setting up production environments, whereas `npm run dev` is used for development purposes. diff --git a/README.md b/README.md index 291ea19..18c9f84 100644 --- a/README.md +++ b/README.md @@ -109,14 +109,13 @@ There are mainly 2 ways of installing Perplexica - With Docker, Without Docker. 1. Install SearXNG and allow `JSON` format in the SearXNG settings. 2. Clone the repository and rename the `sample.config.toml` file to `config.toml` in the root directory. Ensure you complete all required fields in this file. -3. Rename the `.env.example` file to `.env` in the `ui` folder and fill in all necessary fields. -4. After populating the configuration and environment files, run `npm i` in both the `ui` folder and the root directory. -5. Install the dependencies and then execute `npm run build` in both the `ui` folder and the root directory. -6. Finally, start both the frontend and the backend by running `npm run start` in both the `ui` folder and the root directory. +3. After populating the configuration run `npm i`. +4. Install the dependencies and then execute `npm run build`. +5. Finally, start the app by running `npm rum start` **Note**: Using Docker is recommended as it simplifies the setup process, especially for managing environment variables and dependencies. -See the [installation documentation](https://github.com/ItzCrazyKns/Perplexica/tree/master/docs/installation) for more information like exposing it your network, etc. +See the [installation documentation](https://github.com/ItzCrazyKns/Perplexica/tree/master/docs/installation) for more information like updating, etc. ### Ollama Connection Errors @@ -154,12 +153,13 @@ For more details, check out the full documentation [here](https://github.com/Itz ## Expose Perplexica to network -You can access Perplexica over your home network by following our networking guide [here](https://github.com/ItzCrazyKns/Perplexica/blob/master/docs/installation/NETWORKING.md). +Perplexica runs on Next.js and handles all API requests. It works right away on the same network and stays accessible even with port forwarding. ## One-Click Deployment [![Deploy to Sealos](https://raw.githubusercontent.com/labring-actions/templates/main/Deploy-on-Sealos.svg)](https://usw.sealos.io/?openapp=system-template%3FtemplateName%3Dperplexica) [![Deploy to RepoCloud](https://d16t0pc4846x52.cloudfront.net/deploylobe.svg)](https://repocloud.io/details/?app_id=267) +[![Run on ClawCloud](https://raw.githubusercontent.com/ClawCloud/Run-Template/refs/heads/main/Run-on-ClawCloud.svg)](https://template.run.claw.cloud/?referralCode=U11MRQ8U9RM4&openapp=system-fastdeploy%3FtemplateName%3Dperplexica) ## Upcoming Features diff --git a/app.dockerfile b/app.dockerfile index 488e64b..3433288 100644 --- a/app.dockerfile +++ b/app.dockerfile @@ -1,15 +1,27 @@ -FROM node:20.18.0-alpine - -ARG NEXT_PUBLIC_WS_URL=ws://127.0.0.1:3001 -ARG NEXT_PUBLIC_API_URL=http://127.0.0.1:3001/api -ENV NEXT_PUBLIC_WS_URL=${NEXT_PUBLIC_WS_URL} -ENV NEXT_PUBLIC_API_URL=${NEXT_PUBLIC_API_URL} +FROM node:20.18.0-slim AS builder WORKDIR /home/perplexica -COPY ui /home/perplexica/ +COPY package.json yarn.lock ./ +RUN yarn install --frozen-lockfile --network-timeout 600000 -RUN yarn install --frozen-lockfile +COPY tsconfig.json next.config.mjs next-env.d.ts postcss.config.js drizzle.config.ts tailwind.config.ts ./ +COPY src ./src +COPY public ./public + +RUN mkdir -p /home/perplexica/data RUN yarn build -CMD ["yarn", "start"] \ No newline at end of file +FROM node:20.18.0-slim + +WORKDIR /home/perplexica + +COPY --from=builder /home/perplexica/public ./public +COPY --from=builder /home/perplexica/.next/static ./public/_next/static + +COPY --from=builder /home/perplexica/.next/standalone ./ +COPY --from=builder /home/perplexica/data ./data + +RUN mkdir /home/perplexica/uploads + +CMD ["node", "server.js"] \ No newline at end of file diff --git a/backend.dockerfile b/backend.dockerfile deleted file mode 100644 index b6ab95a..0000000 --- a/backend.dockerfile +++ /dev/null @@ -1,17 +0,0 @@ -FROM node:18-slim - -WORKDIR /home/perplexica - -COPY src /home/perplexica/src -COPY tsconfig.json /home/perplexica/ -COPY drizzle.config.ts /home/perplexica/ -COPY package.json /home/perplexica/ -COPY yarn.lock /home/perplexica/ - -RUN mkdir /home/perplexica/data -RUN mkdir /home/perplexica/uploads - -RUN yarn install --frozen-lockfile --network-timeout 600000 -RUN yarn build - -CMD ["yarn", "start"] \ No newline at end of file diff --git a/docker-compose.yaml b/docker-compose.yaml index a0e1d73..b702b4e 100644 --- a/docker-compose.yaml +++ b/docker-compose.yaml @@ -9,41 +9,21 @@ services: - perplexica-network restart: unless-stopped - perplexica-backend: - build: - context: . - dockerfile: backend.dockerfile - image: itzcrazykns1337/perplexica-backend:main - environment: - - SEARXNG_API_URL=http://searxng:8080 - depends_on: - - searxng - ports: - - 3001:3001 - volumes: - - backend-dbstore:/home/perplexica/data - - uploads:/home/perplexica/uploads - - ./config.toml:/home/perplexica/config.toml - extra_hosts: - - 'host.docker.internal:host-gateway' - networks: - - perplexica-network - restart: unless-stopped - - perplexica-frontend: + app: + image: itzcrazykns1337/perplexica:main build: context: . dockerfile: app.dockerfile - args: - - NEXT_PUBLIC_API_URL=http://127.0.0.1:3001/api - - NEXT_PUBLIC_WS_URL=ws://127.0.0.1:3001 - image: itzcrazykns1337/perplexica-frontend:main - depends_on: - - perplexica-backend + environment: + - SEARXNG_API_URL=http://searxng:8080 ports: - 3000:3000 networks: - perplexica-network + volumes: + - backend-dbstore:/home/perplexica/data + - uploads:/home/perplexica/uploads + - ./config.toml:/home/perplexica/config.toml restart: unless-stopped networks: diff --git a/docs/API/SEARCH.md b/docs/API/SEARCH.md index 9405bc5..b67b62b 100644 --- a/docs/API/SEARCH.md +++ b/docs/API/SEARCH.md @@ -6,9 +6,9 @@ Perplexica’s Search API makes it easy to use our AI-powered search engine. You ## Endpoint -### **POST** `http://localhost:3001/api/search` +### **POST** `http://localhost:3000/api/search` -**Note**: Replace `3001` with any other port if you've changed the default PORT +**Note**: Replace `3000` with any other port if you've changed the default PORT ### Request @@ -20,11 +20,11 @@ The API accepts a JSON object in the request body, where you define the focus mo { "chatModel": { "provider": "openai", - "model": "gpt-4o-mini" + "name": "gpt-4o-mini" }, "embeddingModel": { "provider": "openai", - "model": "text-embedding-3-large" + "name": "text-embedding-3-large" }, "optimizationMode": "speed", "focusMode": "webSearch", @@ -32,24 +32,26 @@ The API accepts a JSON object in the request body, where you define the focus mo "history": [ ["human", "Hi, how are you?"], ["assistant", "I am doing well, how can I help you today?"] - ] + ], + "systemInstructions": "Focus on providing technical details about Perplexica's architecture.", + "stream": false } ``` ### Request Parameters -- **`chatModel`** (object, optional): Defines the chat model to be used for the query. For model details you can send a GET request at `http://localhost:3001/api/models`. Make sure to use the key value (For example "gpt-4o-mini" instead of the display name "GPT 4 omni mini"). +- **`chatModel`** (object, optional): Defines the chat model to be used for the query. For model details you can send a GET request at `http://localhost:3000/api/models`. Make sure to use the key value (For example "gpt-4o-mini" instead of the display name "GPT 4 omni mini"). - `provider`: Specifies the provider for the chat model (e.g., `openai`, `ollama`). - - `model`: The specific model from the chosen provider (e.g., `gpt-4o-mini`). + - `name`: The specific model from the chosen provider (e.g., `gpt-4o-mini`). - Optional fields for custom OpenAI configuration: - `customOpenAIBaseURL`: If you’re using a custom OpenAI instance, provide the base URL. - `customOpenAIKey`: The API key for a custom OpenAI instance. -- **`embeddingModel`** (object, optional): Defines the embedding model for similarity-based searching. For model details you can send a GET request at `http://localhost:3001/api/models`. Make sure to use the key value (For example "text-embedding-3-large" instead of the display name "Text Embedding 3 Large"). +- **`embeddingModel`** (object, optional): Defines the embedding model for similarity-based searching. For model details you can send a GET request at `http://localhost:3000/api/models`. Make sure to use the key value (For example "text-embedding-3-large" instead of the display name "Text Embedding 3 Large"). - `provider`: The provider for the embedding model (e.g., `openai`). - - `model`: The specific embedding model (e.g., `text-embedding-3-large`). + - `name`: The specific embedding model (e.g., `text-embedding-3-large`). - **`focusMode`** (string, required): Specifies which focus mode to use. Available modes: @@ -62,6 +64,8 @@ The API accepts a JSON object in the request body, where you define the focus mo - **`query`** (string, required): The search query or question. +- **`systemInstructions`** (string, optional): Custom instructions provided by the user to guide the AI's response. These instructions are treated as user preferences and have lower priority than the system's core instructions. For example, you can specify a particular writing style, format, or focus area. + - **`history`** (array, optional): An array of message pairs representing the conversation history. Each pair consists of a role (either 'human' or 'assistant') and the message content. This allows the system to use the context of the conversation to refine results. Example: ```json @@ -71,11 +75,13 @@ The API accepts a JSON object in the request body, where you define the focus mo ] ``` +- **`stream`** (boolean, optional): When set to `true`, enables streaming responses. Default is `false`. + ### Response The response from the API includes both the final message and the sources used to generate that message. -#### Example Response +#### Standard Response (stream: false) ```json { @@ -100,6 +106,28 @@ The response from the API includes both the final message and the sources used t } ``` +#### Streaming Response (stream: true) + +When streaming is enabled, the API returns a stream of newline-delimited JSON objects. Each line contains a complete, valid JSON object. The response has Content-Type: application/json. + +Example of streamed response objects: + +``` +{"type":"init","data":"Stream connected"} +{"type":"sources","data":[{"pageContent":"...","metadata":{"title":"...","url":"..."}},...]} +{"type":"response","data":"Perplexica is an "} +{"type":"response","data":"innovative, open-source "} +{"type":"response","data":"AI-powered search engine..."} +{"type":"done"} +``` + +Clients should process each line as a separate JSON object. The different message types include: + +- **`init`**: Initial connection message +- **`sources`**: All sources used for the response +- **`response`**: Chunks of the generated answer text +- **`done`**: Indicates the stream is complete + ### Fields in the Response - **`message`** (string): The search result, generated based on the query and focus mode. diff --git a/docs/architecture/WORKING.md b/docs/architecture/WORKING.md index 75b20fd..6bad4f9 100644 --- a/docs/architecture/WORKING.md +++ b/docs/architecture/WORKING.md @@ -4,7 +4,7 @@ Curious about how Perplexica works? Don't worry, we'll cover it here. Before we We'll understand how Perplexica works by taking an example of a scenario where a user asks: "How does an A.C. work?". We'll break down the process into steps to make it easier to understand. The steps are as follows: -1. The message is sent via WS to the backend server where it invokes the chain. The chain will depend on your focus mode. For this example, let's assume we use the "webSearch" focus mode. +1. The message is sent to the `/api/chat` route where it invokes the chain. The chain will depend on your focus mode. For this example, let's assume we use the "webSearch" focus mode. 2. The chain is now invoked; first, the message is passed to another chain where it first predicts (using the chat history and the question) whether there is a need for sources and searching the web. If there is, it will generate a query (in accordance with the chat history) for searching the web that we'll take up later. If not, the chain will end there, and then the answer generator chain, also known as the response generator, will be started. 3. The query returned by the first chain is passed to SearXNG to search the web for information. 4. After the information is retrieved, it is based on keyword-based search. We then convert the information into embeddings and the query as well, then we perform a similarity search to find the most relevant sources to answer the query. diff --git a/docs/installation/NETWORKING.md b/docs/installation/NETWORKING.md deleted file mode 100644 index ae39e3f..0000000 --- a/docs/installation/NETWORKING.md +++ /dev/null @@ -1,109 +0,0 @@ -# Expose Perplexica to a network - -This guide will show you how to make Perplexica available over a network. Follow these steps to allow computers on the same network to interact with Perplexica. Choose the instructions that match the operating system you are using. - -## Windows - -1. Open PowerShell as Administrator - -2. Navigate to the directory containing the `docker-compose.yaml` file - -3. Stop and remove the existing Perplexica containers and images: - - ```bash - docker compose down --rmi all - ``` - -4. Open the `docker-compose.yaml` file in a text editor like Notepad++ - -5. Replace `127.0.0.1` with the IP address of the server Perplexica is running on in these two lines: - - ```bash - args: - - NEXT_PUBLIC_API_URL=http://127.0.0.1:3001/api - - NEXT_PUBLIC_WS_URL=ws://127.0.0.1:3001 - ``` - -6. Save and close the `docker-compose.yaml` file - -7. Rebuild and restart the Perplexica container: - - ```bash - docker compose up -d --build - ``` - -## macOS - -1. Open the Terminal application - -2. Navigate to the directory with the `docker-compose.yaml` file: - - ```bash - cd /path/to/docker-compose.yaml - ``` - -3. Stop and remove existing containers and images: - - ```bash - docker compose down --rmi all - ``` - -4. Open `docker-compose.yaml` in a text editor like Sublime Text: - - ```bash - nano docker-compose.yaml - ``` - -5. Replace `127.0.0.1` with the server IP in these lines: - - ```bash - args: - - NEXT_PUBLIC_API_URL=http://127.0.0.1:3001/api - - NEXT_PUBLIC_WS_URL=ws://127.0.0.1:3001 - ``` - -6. Save and exit the editor - -7. Rebuild and restart Perplexica: - - ```bash - docker compose up -d --build - ``` - -## Linux - -1. Open the terminal - -2. Navigate to the `docker-compose.yaml` directory: - - ```bash - cd /path/to/docker-compose.yaml - ``` - -3. Stop and remove containers and images: - - ```bash - docker compose down --rmi all - ``` - -4. Edit `docker-compose.yaml`: - - ```bash - nano docker-compose.yaml - ``` - -5. Replace `127.0.0.1` with the server IP: - - ```bash - args: - - NEXT_PUBLIC_API_URL=http://127.0.0.1:3001/api - - NEXT_PUBLIC_WS_URL=ws://127.0.0.1:3001 - ``` - -6. Save and exit the editor - -7. Rebuild and restart Perplexica: - - ```bash - docker compose up -d --build - ``` diff --git a/docs/installation/UPDATING.md b/docs/installation/UPDATING.md index a4651df..972142f 100644 --- a/docs/installation/UPDATING.md +++ b/docs/installation/UPDATING.md @@ -39,11 +39,8 @@ To update Perplexica to the latest version, follow these steps: 2. Navigate to the project directory. 3. Check for changes in the configuration files. If the `sample.config.toml` file contains new fields, delete your existing `config.toml` file, rename `sample.config.toml` to `config.toml`, and update the configuration accordingly. - -4. Execute `npm i` in both the `ui` folder and the root directory. - -5. Once the packages are updated, execute `npm run build` in both the `ui` folder and the root directory. - -6. Finally, start both the frontend and the backend by running `npm run start` in both the `ui` folder and the root directory. +4. After populating the configuration run `npm i`. +5. Install the dependencies and then execute `npm run build`. +6. Finally, start the app by running `npm rum start` --- diff --git a/drizzle.config.ts b/drizzle.config.ts index 9ac3ec5..58de9e0 100644 --- a/drizzle.config.ts +++ b/drizzle.config.ts @@ -2,7 +2,7 @@ import { defineConfig } from 'drizzle-kit'; export default defineConfig({ dialect: 'sqlite', - schema: './src/db/schema.ts', + schema: './src/lib/db/schema.ts', out: './drizzle', dbCredentials: { url: './data/db.sqlite', diff --git a/next-env.d.ts b/next-env.d.ts new file mode 100644 index 0000000..1b3be08 --- /dev/null +++ b/next-env.d.ts @@ -0,0 +1,5 @@ +/// +/// + +// NOTE: This file should not be edited +// see https://nextjs.org/docs/app/api-reference/config/typescript for more information. diff --git a/ui/next.config.mjs b/next.config.mjs similarity index 75% rename from ui/next.config.mjs rename to next.config.mjs index c3f2e1a..2300ff4 100644 --- a/ui/next.config.mjs +++ b/next.config.mjs @@ -1,5 +1,6 @@ /** @type {import('next').NextConfig} */ const nextConfig = { + output: 'standalone', images: { remotePatterns: [ { @@ -7,6 +8,7 @@ const nextConfig = { }, ], }, + serverExternalPackages: ['pdf-parse'], }; export default nextConfig; diff --git a/package.json b/package.json index 6041377..e68410f 100644 --- a/package.json +++ b/package.json @@ -1,53 +1,65 @@ { - "name": "perplexica-backend", - "version": "1.10.0-rc3", + "name": "perplexica-frontend", + "version": "1.10.2", "license": "MIT", "author": "ItzCrazyKns", "scripts": { - "start": "npm run db:push && node dist/app.js", - "build": "tsc", - "dev": "nodemon --ignore uploads/ src/app.ts ", - "db:push": "drizzle-kit push sqlite", - "format": "prettier . --check", - "format:write": "prettier . --write" - }, - "devDependencies": { - "@types/better-sqlite3": "^7.6.10", - "@types/cors": "^2.8.17", - "@types/express": "^4.17.21", - "@types/html-to-text": "^9.0.4", - "@types/multer": "^1.4.12", - "@types/pdf-parse": "^1.1.4", - "@types/readable-stream": "^4.0.11", - "@types/ws": "^8.5.12", - "drizzle-kit": "^0.22.7", - "nodemon": "^3.1.0", - "prettier": "^3.2.5", - "ts-node": "^10.9.2", - "typescript": "^5.4.3" + "dev": "next dev", + "build": "npm run db:push && next build", + "start": "next start", + "lint": "next lint", + "format:write": "prettier . --write", + "db:push": "drizzle-kit push" }, "dependencies": { + "@headlessui/react": "^2.2.0", "@iarna/toml": "^2.2.5", - "@langchain/anthropic": "^0.2.3", - "@langchain/community": "^0.2.16", + "@icons-pack/react-simple-icons": "^12.3.0", + "@langchain/anthropic": "^0.3.15", + "@langchain/community": "^0.3.36", + "@langchain/core": "^0.3.42", + "@langchain/google-genai": "^0.1.12", "@langchain/openai": "^0.0.25", - "@langchain/google-genai": "^0.0.23", - "@xenova/transformers": "^2.17.1", - "axios": "^1.6.8", - "better-sqlite3": "^11.0.0", + "@langchain/textsplitters": "^0.1.0", + "@tailwindcss/typography": "^0.5.12", + "@xenova/transformers": "^2.17.2", + "axios": "^1.8.3", + "better-sqlite3": "^11.9.1", + "clsx": "^2.1.0", "compute-cosine-similarity": "^1.1.0", "compute-dot": "^1.1.0", - "cors": "^2.8.5", - "dotenv": "^16.4.5", - "drizzle-orm": "^0.31.2", - "express": "^4.19.2", + "drizzle-orm": "^0.40.1", "html-to-text": "^9.0.5", "langchain": "^0.1.30", - "mammoth": "^1.8.0", - "multer": "^1.4.5-lts.1", + "lucide-react": "^0.363.0", + "markdown-to-jsx": "^7.7.2", + "next": "^15.2.2", + "next-themes": "^0.3.0", "pdf-parse": "^1.1.1", - "winston": "^3.13.0", - "ws": "^8.17.1", + "react": "^18", + "react-dom": "^18", + "react-text-to-speech": "^0.14.5", + "react-textarea-autosize": "^8.5.3", + "sonner": "^1.4.41", + "tailwind-merge": "^2.2.2", + "winston": "^3.17.0", + "yet-another-react-lightbox": "^3.17.2", "zod": "^3.22.4" + }, + "devDependencies": { + "@types/better-sqlite3": "^7.6.12", + "@types/html-to-text": "^9.0.4", + "@types/node": "^20", + "@types/pdf-parse": "^1.1.4", + "@types/react": "^18", + "@types/react-dom": "^18", + "autoprefixer": "^10.0.1", + "drizzle-kit": "^0.30.5", + "eslint": "^8", + "eslint-config-next": "14.1.4", + "postcss": "^8", + "prettier": "^3.2.5", + "tailwindcss": "^3.3.0", + "typescript": "^5" } } diff --git a/ui/postcss.config.js b/postcss.config.js similarity index 100% rename from ui/postcss.config.js rename to postcss.config.js diff --git a/ui/public/next.svg b/public/next.svg similarity index 100% rename from ui/public/next.svg rename to public/next.svg diff --git a/ui/public/vercel.svg b/public/vercel.svg similarity index 100% rename from ui/public/vercel.svg rename to public/vercel.svg diff --git a/sample.config.toml b/sample.config.toml index 7b09d67..1db2125 100644 --- a/sample.config.toml +++ b/sample.config.toml @@ -1,5 +1,4 @@ [GENERAL] -PORT = 3001 # Port to run the server on SIMILARITY_MEASURE = "cosine" # "cosine" or "dot" KEEP_ALIVE = "5m" # How long to keep Ollama models loaded into memory. (Instead of using -1 use "-1m") @@ -18,9 +17,16 @@ API_KEY = "" [MODELS.CUSTOM_OPENAI] API_KEY = "" API_URL = "" +MODEL_NAME = "" [MODELS.OLLAMA] API_URL = "" # Ollama API URL - http://host.docker.internal:11434 +[MODELS.DEEPSEEK] +API_KEY = "" + +[MODELS.LM_STUDIO] +API_URL = "" # LM Studio API URL - http://host.docker.internal:1234 + [API_ENDPOINTS] -SEARXNG = "http://localhost:32768" # SearxNG API URL \ No newline at end of file +SEARXNG = "" # SearxNG API URL - http://localhost:32768 diff --git a/src/app.ts b/src/app.ts deleted file mode 100644 index 96b3a0c..0000000 --- a/src/app.ts +++ /dev/null @@ -1,38 +0,0 @@ -import { startWebSocketServer } from './websocket'; -import express from 'express'; -import cors from 'cors'; -import http from 'http'; -import routes from './routes'; -import { getPort } from './config'; -import logger from './utils/logger'; - -const port = getPort(); - -const app = express(); -const server = http.createServer(app); - -const corsOptions = { - origin: '*', -}; - -app.use(cors(corsOptions)); -app.use(express.json()); - -app.use('/api', routes); -app.get('/api', (_, res) => { - res.status(200).json({ status: 'ok' }); -}); - -server.listen(port, () => { - logger.info(`Server is running on port ${port}`); -}); - -startWebSocketServer(server); - -process.on('uncaughtException', (err, origin) => { - logger.error(`Uncaught Exception at ${origin}: ${err}`); -}); - -process.on('unhandledRejection', (reason, promise) => { - logger.error(`Unhandled Rejection at: ${promise}, reason: ${reason}`); -}); diff --git a/src/app/api/chat/route.ts b/src/app/api/chat/route.ts new file mode 100644 index 0000000..e566edb --- /dev/null +++ b/src/app/api/chat/route.ts @@ -0,0 +1,306 @@ +import prompts from '@/lib/prompts'; +import MetaSearchAgent from '@/lib/search/metaSearchAgent'; +import crypto from 'crypto'; +import { AIMessage, BaseMessage, HumanMessage } from '@langchain/core/messages'; +import { EventEmitter } from 'stream'; +import { + chatModelProviders, + embeddingModelProviders, + getAvailableChatModelProviders, + getAvailableEmbeddingModelProviders, +} from '@/lib/providers'; +import db from '@/lib/db'; +import { chats, messages as messagesSchema } from '@/lib/db/schema'; +import { and, eq, gt } from 'drizzle-orm'; +import { getFileDetails } from '@/lib/utils/files'; +import { BaseChatModel } from '@langchain/core/language_models/chat_models'; +import { ChatOpenAI } from '@langchain/openai'; +import { + getCustomOpenaiApiKey, + getCustomOpenaiApiUrl, + getCustomOpenaiModelName, +} from '@/lib/config'; +import { searchHandlers } from '@/lib/search'; + +export const runtime = 'nodejs'; +export const dynamic = 'force-dynamic'; + +type Message = { + messageId: string; + chatId: string; + content: string; +}; + +type ChatModel = { + provider: string; + name: string; +}; + +type EmbeddingModel = { + provider: string; + name: string; +}; + +type Body = { + message: Message; + optimizationMode: 'speed' | 'balanced' | 'quality'; + focusMode: string; + history: Array<[string, string]>; + files: Array; + chatModel: ChatModel; + embeddingModel: EmbeddingModel; + systemInstructions: string; +}; + +const handleEmitterEvents = async ( + stream: EventEmitter, + writer: WritableStreamDefaultWriter, + encoder: TextEncoder, + aiMessageId: string, + chatId: string, +) => { + let recievedMessage = ''; + let sources: any[] = []; + + stream.on('data', (data) => { + const parsedData = JSON.parse(data); + if (parsedData.type === 'response') { + writer.write( + encoder.encode( + JSON.stringify({ + type: 'message', + data: parsedData.data, + messageId: aiMessageId, + }) + '\n', + ), + ); + + recievedMessage += parsedData.data; + } else if (parsedData.type === 'sources') { + writer.write( + encoder.encode( + JSON.stringify({ + type: 'sources', + data: parsedData.data, + messageId: aiMessageId, + }) + '\n', + ), + ); + + sources = parsedData.data; + } + }); + stream.on('end', () => { + writer.write( + encoder.encode( + JSON.stringify({ + type: 'messageEnd', + messageId: aiMessageId, + }) + '\n', + ), + ); + writer.close(); + + db.insert(messagesSchema) + .values({ + content: recievedMessage, + chatId: chatId, + messageId: aiMessageId, + role: 'assistant', + metadata: JSON.stringify({ + createdAt: new Date(), + ...(sources && sources.length > 0 && { sources }), + }), + }) + .execute(); + }); + stream.on('error', (data) => { + const parsedData = JSON.parse(data); + writer.write( + encoder.encode( + JSON.stringify({ + type: 'error', + data: parsedData.data, + }), + ), + ); + writer.close(); + }); +}; + +const handleHistorySave = async ( + message: Message, + humanMessageId: string, + focusMode: string, + files: string[], +) => { + const chat = await db.query.chats.findFirst({ + where: eq(chats.id, message.chatId), + }); + + if (!chat) { + await db + .insert(chats) + .values({ + id: message.chatId, + title: message.content, + createdAt: new Date().toString(), + focusMode: focusMode, + files: files.map(getFileDetails), + }) + .execute(); + } + + const messageExists = await db.query.messages.findFirst({ + where: eq(messagesSchema.messageId, humanMessageId), + }); + + if (!messageExists) { + await db + .insert(messagesSchema) + .values({ + content: message.content, + chatId: message.chatId, + messageId: humanMessageId, + role: 'user', + metadata: JSON.stringify({ + createdAt: new Date(), + }), + }) + .execute(); + } else { + await db + .delete(messagesSchema) + .where( + and( + gt(messagesSchema.id, messageExists.id), + eq(messagesSchema.chatId, message.chatId), + ), + ) + .execute(); + } +}; + +export const POST = async (req: Request) => { + try { + const body = (await req.json()) as Body; + const { message } = body; + + if (message.content === '') { + return Response.json( + { + message: 'Please provide a message to process', + }, + { status: 400 }, + ); + } + + const [chatModelProviders, embeddingModelProviders] = await Promise.all([ + getAvailableChatModelProviders(), + getAvailableEmbeddingModelProviders(), + ]); + + const chatModelProvider = + chatModelProviders[ + body.chatModel?.provider || Object.keys(chatModelProviders)[0] + ]; + const chatModel = + chatModelProvider[ + body.chatModel?.name || Object.keys(chatModelProvider)[0] + ]; + + const embeddingProvider = + embeddingModelProviders[ + body.embeddingModel?.provider || Object.keys(embeddingModelProviders)[0] + ]; + const embeddingModel = + embeddingProvider[ + body.embeddingModel?.name || Object.keys(embeddingProvider)[0] + ]; + + let llm: BaseChatModel | undefined; + let embedding = embeddingModel.model; + + if (body.chatModel?.provider === 'custom_openai') { + llm = new ChatOpenAI({ + openAIApiKey: getCustomOpenaiApiKey(), + modelName: getCustomOpenaiModelName(), + temperature: 0.7, + configuration: { + baseURL: getCustomOpenaiApiUrl(), + }, + }) as unknown as BaseChatModel; + } else if (chatModelProvider && chatModel) { + llm = chatModel.model; + } + + if (!llm) { + return Response.json({ error: 'Invalid chat model' }, { status: 400 }); + } + + if (!embedding) { + return Response.json( + { error: 'Invalid embedding model' }, + { status: 400 }, + ); + } + + const humanMessageId = + message.messageId ?? crypto.randomBytes(7).toString('hex'); + const aiMessageId = crypto.randomBytes(7).toString('hex'); + + const history: BaseMessage[] = body.history.map((msg) => { + if (msg[0] === 'human') { + return new HumanMessage({ + content: msg[1], + }); + } else { + return new AIMessage({ + content: msg[1], + }); + } + }); + + const handler = searchHandlers[body.focusMode]; + + if (!handler) { + return Response.json( + { + message: 'Invalid focus mode', + }, + { status: 400 }, + ); + } + + const stream = await handler.searchAndAnswer( + message.content, + history, + llm, + embedding, + body.optimizationMode, + body.files, + body.systemInstructions, + ); + + const responseStream = new TransformStream(); + const writer = responseStream.writable.getWriter(); + const encoder = new TextEncoder(); + + handleEmitterEvents(stream, writer, encoder, aiMessageId, message.chatId); + handleHistorySave(message, humanMessageId, body.focusMode, body.files); + + return new Response(responseStream.readable, { + headers: { + 'Content-Type': 'text/event-stream', + Connection: 'keep-alive', + 'Cache-Control': 'no-cache, no-transform', + }, + }); + } catch (err) { + console.error('An error occurred while processing chat request:', err); + return Response.json( + { message: 'An error occurred while processing chat request' }, + { status: 500 }, + ); + } +}; diff --git a/src/app/api/chats/[id]/route.ts b/src/app/api/chats/[id]/route.ts new file mode 100644 index 0000000..6891454 --- /dev/null +++ b/src/app/api/chats/[id]/route.ts @@ -0,0 +1,69 @@ +import db from '@/lib/db'; +import { chats, messages } from '@/lib/db/schema'; +import { eq } from 'drizzle-orm'; + +export const GET = async ( + req: Request, + { params }: { params: Promise<{ id: string }> }, +) => { + try { + const { id } = await params; + + const chatExists = await db.query.chats.findFirst({ + where: eq(chats.id, id), + }); + + if (!chatExists) { + return Response.json({ message: 'Chat not found' }, { status: 404 }); + } + + const chatMessages = await db.query.messages.findMany({ + where: eq(messages.chatId, id), + }); + + return Response.json( + { + chat: chatExists, + messages: chatMessages, + }, + { status: 200 }, + ); + } catch (err) { + console.error('Error in getting chat by id: ', err); + return Response.json( + { message: 'An error has occurred.' }, + { status: 500 }, + ); + } +}; + +export const DELETE = async ( + req: Request, + { params }: { params: Promise<{ id: string }> }, +) => { + try { + const { id } = await params; + + const chatExists = await db.query.chats.findFirst({ + where: eq(chats.id, id), + }); + + if (!chatExists) { + return Response.json({ message: 'Chat not found' }, { status: 404 }); + } + + await db.delete(chats).where(eq(chats.id, id)).execute(); + await db.delete(messages).where(eq(messages.chatId, id)).execute(); + + return Response.json( + { message: 'Chat deleted successfully' }, + { status: 200 }, + ); + } catch (err) { + console.error('Error in deleting chat by id: ', err); + return Response.json( + { message: 'An error has occurred.' }, + { status: 500 }, + ); + } +}; diff --git a/src/app/api/chats/route.ts b/src/app/api/chats/route.ts new file mode 100644 index 0000000..986a192 --- /dev/null +++ b/src/app/api/chats/route.ts @@ -0,0 +1,15 @@ +import db from '@/lib/db'; + +export const GET = async (req: Request) => { + try { + let chats = await db.query.chats.findMany(); + chats = chats.reverse(); + return Response.json({ chats: chats }, { status: 200 }); + } catch (err) { + console.error('Error in getting chats: ', err); + return Response.json( + { message: 'An error has occurred.' }, + { status: 500 }, + ); + } +}; diff --git a/src/app/api/config/route.ts b/src/app/api/config/route.ts new file mode 100644 index 0000000..c1e5bbd --- /dev/null +++ b/src/app/api/config/route.ts @@ -0,0 +1,119 @@ +import { + getAnthropicApiKey, + getCustomOpenaiApiKey, + getCustomOpenaiApiUrl, + getCustomOpenaiModelName, + getGeminiApiKey, + getGroqApiKey, + getOllamaApiEndpoint, + getOpenaiApiKey, + getDeepseekApiKey, + getLMStudioApiEndpoint, + updateConfig, +} from '@/lib/config'; +import { + getAvailableChatModelProviders, + getAvailableEmbeddingModelProviders, +} from '@/lib/providers'; + +export const GET = async (req: Request) => { + try { + const config: Record = {}; + + const [chatModelProviders, embeddingModelProviders] = await Promise.all([ + getAvailableChatModelProviders(), + getAvailableEmbeddingModelProviders(), + ]); + + config['chatModelProviders'] = {}; + config['embeddingModelProviders'] = {}; + + for (const provider in chatModelProviders) { + config['chatModelProviders'][provider] = Object.keys( + chatModelProviders[provider], + ).map((model) => { + return { + name: model, + displayName: chatModelProviders[provider][model].displayName, + }; + }); + } + + for (const provider in embeddingModelProviders) { + config['embeddingModelProviders'][provider] = Object.keys( + embeddingModelProviders[provider], + ).map((model) => { + return { + name: model, + displayName: embeddingModelProviders[provider][model].displayName, + }; + }); + } + + config['openaiApiKey'] = getOpenaiApiKey(); + config['ollamaApiUrl'] = getOllamaApiEndpoint(); + config['lmStudioApiUrl'] = getLMStudioApiEndpoint(); + config['anthropicApiKey'] = getAnthropicApiKey(); + config['groqApiKey'] = getGroqApiKey(); + config['geminiApiKey'] = getGeminiApiKey(); + config['deepseekApiKey'] = getDeepseekApiKey(); + config['customOpenaiApiUrl'] = getCustomOpenaiApiUrl(); + config['customOpenaiApiKey'] = getCustomOpenaiApiKey(); + config['customOpenaiModelName'] = getCustomOpenaiModelName(); + + return Response.json({ ...config }, { status: 200 }); + } catch (err) { + console.error('An error occurred while getting config:', err); + return Response.json( + { message: 'An error occurred while getting config' }, + { status: 500 }, + ); + } +}; + +export const POST = async (req: Request) => { + try { + const config = await req.json(); + + const updatedConfig = { + MODELS: { + OPENAI: { + API_KEY: config.openaiApiKey, + }, + GROQ: { + API_KEY: config.groqApiKey, + }, + ANTHROPIC: { + API_KEY: config.anthropicApiKey, + }, + GEMINI: { + API_KEY: config.geminiApiKey, + }, + OLLAMA: { + API_URL: config.ollamaApiUrl, + }, + DEEPSEEK: { + API_KEY: config.deepseekApiKey, + }, + LM_STUDIO: { + API_URL: config.lmStudioApiUrl, + }, + CUSTOM_OPENAI: { + API_URL: config.customOpenaiApiUrl, + API_KEY: config.customOpenaiApiKey, + MODEL_NAME: config.customOpenaiModelName, + }, + }, + }; + + updateConfig(updatedConfig); + + return Response.json({ message: 'Config updated' }, { status: 200 }); + } catch (err) { + console.error('An error occurred while updating config:', err); + return Response.json( + { message: 'An error occurred while updating config' }, + { status: 500 }, + ); + } +}; diff --git a/src/app/api/discover/route.ts b/src/app/api/discover/route.ts new file mode 100644 index 0000000..8c1f470 --- /dev/null +++ b/src/app/api/discover/route.ts @@ -0,0 +1,61 @@ +import { searchSearxng } from '@/lib/searxng'; + +const articleWebsites = [ + 'yahoo.com', + 'www.exchangewire.com', + 'businessinsider.com', + /* 'wired.com', + 'mashable.com', + 'theverge.com', + 'gizmodo.com', + 'cnet.com', + 'venturebeat.com', */ +]; + +const topics = ['AI', 'tech']; /* TODO: Add UI to customize this */ + +export const GET = async (req: Request) => { + try { + const data = ( + await Promise.all([ + ...new Array(articleWebsites.length * topics.length) + .fill(0) + .map(async (_, i) => { + return ( + await searchSearxng( + `site:${articleWebsites[i % articleWebsites.length]} ${ + topics[i % topics.length] + }`, + { + engines: ['bing news'], + pageno: 1, + }, + ) + ).results; + }), + ]) + ) + .map((result) => result) + .flat() + .sort(() => Math.random() - 0.5); + + return Response.json( + { + blogs: data, + }, + { + status: 200, + }, + ); + } catch (err) { + console.error(`An error occurred in discover route: ${err}`); + return Response.json( + { + message: 'An error has occurred', + }, + { + status: 500, + }, + ); + } +}; diff --git a/src/app/api/images/route.ts b/src/app/api/images/route.ts new file mode 100644 index 0000000..db39d9f --- /dev/null +++ b/src/app/api/images/route.ts @@ -0,0 +1,83 @@ +import handleImageSearch from '@/lib/chains/imageSearchAgent'; +import { + getCustomOpenaiApiKey, + getCustomOpenaiApiUrl, + getCustomOpenaiModelName, +} from '@/lib/config'; +import { getAvailableChatModelProviders } from '@/lib/providers'; +import { BaseChatModel } from '@langchain/core/language_models/chat_models'; +import { AIMessage, BaseMessage, HumanMessage } from '@langchain/core/messages'; +import { ChatOpenAI } from '@langchain/openai'; + +interface ChatModel { + provider: string; + model: string; +} + +interface ImageSearchBody { + query: string; + chatHistory: any[]; + chatModel?: ChatModel; +} + +export const POST = async (req: Request) => { + try { + const body: ImageSearchBody = await req.json(); + + const chatHistory = body.chatHistory + .map((msg: any) => { + if (msg.role === 'user') { + return new HumanMessage(msg.content); + } else if (msg.role === 'assistant') { + return new AIMessage(msg.content); + } + }) + .filter((msg) => msg !== undefined) as BaseMessage[]; + + const chatModelProviders = await getAvailableChatModelProviders(); + + const chatModelProvider = + chatModelProviders[ + body.chatModel?.provider || Object.keys(chatModelProviders)[0] + ]; + const chatModel = + chatModelProvider[ + body.chatModel?.model || Object.keys(chatModelProvider)[0] + ]; + + let llm: BaseChatModel | undefined; + + if (body.chatModel?.provider === 'custom_openai') { + llm = new ChatOpenAI({ + openAIApiKey: getCustomOpenaiApiKey(), + modelName: getCustomOpenaiModelName(), + temperature: 0.7, + configuration: { + baseURL: getCustomOpenaiApiUrl(), + }, + }) as unknown as BaseChatModel; + } else if (chatModelProvider && chatModel) { + llm = chatModel.model; + } + + if (!llm) { + return Response.json({ error: 'Invalid chat model' }, { status: 400 }); + } + + const images = await handleImageSearch( + { + chat_history: chatHistory, + query: body.query, + }, + llm, + ); + + return Response.json({ images }, { status: 200 }); + } catch (err) { + console.error(`An error occurred while searching images: ${err}`); + return Response.json( + { message: 'An error occurred while searching images' }, + { status: 500 }, + ); + } +}; diff --git a/src/app/api/models/route.ts b/src/app/api/models/route.ts new file mode 100644 index 0000000..04a6949 --- /dev/null +++ b/src/app/api/models/route.ts @@ -0,0 +1,47 @@ +import { + getAvailableChatModelProviders, + getAvailableEmbeddingModelProviders, +} from '@/lib/providers'; + +export const GET = async (req: Request) => { + try { + const [chatModelProviders, embeddingModelProviders] = await Promise.all([ + getAvailableChatModelProviders(), + getAvailableEmbeddingModelProviders(), + ]); + + Object.keys(chatModelProviders).forEach((provider) => { + Object.keys(chatModelProviders[provider]).forEach((model) => { + delete (chatModelProviders[provider][model] as { model?: unknown }) + .model; + }); + }); + + Object.keys(embeddingModelProviders).forEach((provider) => { + Object.keys(embeddingModelProviders[provider]).forEach((model) => { + delete (embeddingModelProviders[provider][model] as { model?: unknown }) + .model; + }); + }); + + return Response.json( + { + chatModelProviders, + embeddingModelProviders, + }, + { + status: 200, + }, + ); + } catch (err) { + console.error('An error occurred while fetching models', err); + return Response.json( + { + message: 'An error has occurred.', + }, + { + status: 500, + }, + ); + } +}; diff --git a/src/app/api/search/route.ts b/src/app/api/search/route.ts new file mode 100644 index 0000000..970ec42 --- /dev/null +++ b/src/app/api/search/route.ts @@ -0,0 +1,270 @@ +import type { BaseChatModel } from '@langchain/core/language_models/chat_models'; +import type { Embeddings } from '@langchain/core/embeddings'; +import { ChatOpenAI } from '@langchain/openai'; +import { + getAvailableChatModelProviders, + getAvailableEmbeddingModelProviders, +} from '@/lib/providers'; +import { AIMessage, BaseMessage, HumanMessage } from '@langchain/core/messages'; +import { MetaSearchAgentType } from '@/lib/search/metaSearchAgent'; +import { + getCustomOpenaiApiKey, + getCustomOpenaiApiUrl, + getCustomOpenaiModelName, +} from '@/lib/config'; +import { searchHandlers } from '@/lib/search'; + +interface chatModel { + provider: string; + name: string; + customOpenAIKey?: string; + customOpenAIBaseURL?: string; +} + +interface embeddingModel { + provider: string; + name: string; +} + +interface ChatRequestBody { + optimizationMode: 'speed' | 'balanced'; + focusMode: string; + chatModel?: chatModel; + embeddingModel?: embeddingModel; + query: string; + history: Array<[string, string]>; + stream?: boolean; + systemInstructions?: string; +} + +export const POST = async (req: Request) => { + try { + const body: ChatRequestBody = await req.json(); + + if (!body.focusMode || !body.query) { + return Response.json( + { message: 'Missing focus mode or query' }, + { status: 400 }, + ); + } + + body.history = body.history || []; + body.optimizationMode = body.optimizationMode || 'balanced'; + body.stream = body.stream || false; + + const history: BaseMessage[] = body.history.map((msg) => { + return msg[0] === 'human' + ? new HumanMessage({ content: msg[1] }) + : new AIMessage({ content: msg[1] }); + }); + + const [chatModelProviders, embeddingModelProviders] = await Promise.all([ + getAvailableChatModelProviders(), + getAvailableEmbeddingModelProviders(), + ]); + + const chatModelProvider = + body.chatModel?.provider || Object.keys(chatModelProviders)[0]; + const chatModel = + body.chatModel?.name || + Object.keys(chatModelProviders[chatModelProvider])[0]; + + const embeddingModelProvider = + body.embeddingModel?.provider || Object.keys(embeddingModelProviders)[0]; + const embeddingModel = + body.embeddingModel?.name || + Object.keys(embeddingModelProviders[embeddingModelProvider])[0]; + + let llm: BaseChatModel | undefined; + let embeddings: Embeddings | undefined; + + if (body.chatModel?.provider === 'custom_openai') { + llm = new ChatOpenAI({ + modelName: body.chatModel?.name || getCustomOpenaiModelName(), + openAIApiKey: + body.chatModel?.customOpenAIKey || getCustomOpenaiApiKey(), + temperature: 0.7, + configuration: { + baseURL: + body.chatModel?.customOpenAIBaseURL || getCustomOpenaiApiUrl(), + }, + }) as unknown as BaseChatModel; + } else if ( + chatModelProviders[chatModelProvider] && + chatModelProviders[chatModelProvider][chatModel] + ) { + llm = chatModelProviders[chatModelProvider][chatModel] + .model as unknown as BaseChatModel | undefined; + } + + if ( + embeddingModelProviders[embeddingModelProvider] && + embeddingModelProviders[embeddingModelProvider][embeddingModel] + ) { + embeddings = embeddingModelProviders[embeddingModelProvider][ + embeddingModel + ].model as Embeddings | undefined; + } + + if (!llm || !embeddings) { + return Response.json( + { message: 'Invalid model selected' }, + { status: 400 }, + ); + } + + const searchHandler: MetaSearchAgentType = searchHandlers[body.focusMode]; + + if (!searchHandler) { + return Response.json({ message: 'Invalid focus mode' }, { status: 400 }); + } + + const emitter = await searchHandler.searchAndAnswer( + body.query, + history, + llm, + embeddings, + body.optimizationMode, + [], + body.systemInstructions || '', + ); + + if (!body.stream) { + return new Promise( + ( + resolve: (value: Response) => void, + reject: (value: Response) => void, + ) => { + let message = ''; + let sources: any[] = []; + + emitter.on('data', (data: string) => { + try { + const parsedData = JSON.parse(data); + if (parsedData.type === 'response') { + message += parsedData.data; + } else if (parsedData.type === 'sources') { + sources = parsedData.data; + } + } catch (error) { + reject( + Response.json( + { message: 'Error parsing data' }, + { status: 500 }, + ), + ); + } + }); + + emitter.on('end', () => { + resolve(Response.json({ message, sources }, { status: 200 })); + }); + + emitter.on('error', (error: any) => { + reject( + Response.json( + { message: 'Search error', error }, + { status: 500 }, + ), + ); + }); + }, + ); + } + + const encoder = new TextEncoder(); + + const abortController = new AbortController(); + const { signal } = abortController; + + const stream = new ReadableStream({ + start(controller) { + let sources: any[] = []; + + controller.enqueue( + encoder.encode( + JSON.stringify({ + type: 'init', + data: 'Stream connected', + }) + '\n', + ), + ); + + signal.addEventListener('abort', () => { + emitter.removeAllListeners(); + + try { + controller.close(); + } catch (error) {} + }); + + emitter.on('data', (data: string) => { + if (signal.aborted) return; + + try { + const parsedData = JSON.parse(data); + + if (parsedData.type === 'response') { + controller.enqueue( + encoder.encode( + JSON.stringify({ + type: 'response', + data: parsedData.data, + }) + '\n', + ), + ); + } else if (parsedData.type === 'sources') { + sources = parsedData.data; + controller.enqueue( + encoder.encode( + JSON.stringify({ + type: 'sources', + data: sources, + }) + '\n', + ), + ); + } + } catch (error) { + controller.error(error); + } + }); + + emitter.on('end', () => { + if (signal.aborted) return; + + controller.enqueue( + encoder.encode( + JSON.stringify({ + type: 'done', + }) + '\n', + ), + ); + controller.close(); + }); + + emitter.on('error', (error: any) => { + if (signal.aborted) return; + + controller.error(error); + }); + }, + cancel() { + abortController.abort(); + }, + }); + + return new Response(stream, { + headers: { + 'Content-Type': 'text/event-stream', + 'Cache-Control': 'no-cache, no-transform', + Connection: 'keep-alive', + }, + }); + } catch (err: any) { + console.error(`Error in getting search results: ${err.message}`); + return Response.json( + { message: 'An error has occurred.' }, + { status: 500 }, + ); + } +}; diff --git a/src/app/api/suggestions/route.ts b/src/app/api/suggestions/route.ts new file mode 100644 index 0000000..e92e5ec --- /dev/null +++ b/src/app/api/suggestions/route.ts @@ -0,0 +1,81 @@ +import generateSuggestions from '@/lib/chains/suggestionGeneratorAgent'; +import { + getCustomOpenaiApiKey, + getCustomOpenaiApiUrl, + getCustomOpenaiModelName, +} from '@/lib/config'; +import { getAvailableChatModelProviders } from '@/lib/providers'; +import { BaseChatModel } from '@langchain/core/language_models/chat_models'; +import { AIMessage, BaseMessage, HumanMessage } from '@langchain/core/messages'; +import { ChatOpenAI } from '@langchain/openai'; + +interface ChatModel { + provider: string; + model: string; +} + +interface SuggestionsGenerationBody { + chatHistory: any[]; + chatModel?: ChatModel; +} + +export const POST = async (req: Request) => { + try { + const body: SuggestionsGenerationBody = await req.json(); + + const chatHistory = body.chatHistory + .map((msg: any) => { + if (msg.role === 'user') { + return new HumanMessage(msg.content); + } else if (msg.role === 'assistant') { + return new AIMessage(msg.content); + } + }) + .filter((msg) => msg !== undefined) as BaseMessage[]; + + const chatModelProviders = await getAvailableChatModelProviders(); + + const chatModelProvider = + chatModelProviders[ + body.chatModel?.provider || Object.keys(chatModelProviders)[0] + ]; + const chatModel = + chatModelProvider[ + body.chatModel?.model || Object.keys(chatModelProvider)[0] + ]; + + let llm: BaseChatModel | undefined; + + if (body.chatModel?.provider === 'custom_openai') { + llm = new ChatOpenAI({ + openAIApiKey: getCustomOpenaiApiKey(), + modelName: getCustomOpenaiModelName(), + temperature: 0.7, + configuration: { + baseURL: getCustomOpenaiApiUrl(), + }, + }) as unknown as BaseChatModel; + } else if (chatModelProvider && chatModel) { + llm = chatModel.model; + } + + if (!llm) { + return Response.json({ error: 'Invalid chat model' }, { status: 400 }); + } + + const suggestions = await generateSuggestions( + { + chat_history: chatHistory, + }, + llm, + ); + + return Response.json({ suggestions }, { status: 200 }); + } catch (err) { + console.error(`An error occurred while generating suggestions: ${err}`); + return Response.json( + { message: 'An error occurred while generating suggestions' }, + { status: 500 }, + ); + } +}; diff --git a/src/app/api/uploads/route.ts b/src/app/api/uploads/route.ts new file mode 100644 index 0000000..9fbaf2d --- /dev/null +++ b/src/app/api/uploads/route.ts @@ -0,0 +1,134 @@ +import { NextResponse } from 'next/server'; +import fs from 'fs'; +import path from 'path'; +import crypto from 'crypto'; +import { getAvailableEmbeddingModelProviders } from '@/lib/providers'; +import { PDFLoader } from '@langchain/community/document_loaders/fs/pdf'; +import { DocxLoader } from '@langchain/community/document_loaders/fs/docx'; +import { RecursiveCharacterTextSplitter } from '@langchain/textsplitters'; +import { Document } from 'langchain/document'; + +interface FileRes { + fileName: string; + fileExtension: string; + fileId: string; +} + +const uploadDir = path.join(process.cwd(), 'uploads'); + +if (!fs.existsSync(uploadDir)) { + fs.mkdirSync(uploadDir, { recursive: true }); +} + +const splitter = new RecursiveCharacterTextSplitter({ + chunkSize: 500, + chunkOverlap: 100, +}); + +export async function POST(req: Request) { + try { + const formData = await req.formData(); + + const files = formData.getAll('files') as File[]; + const embedding_model = formData.get('embedding_model'); + const embedding_model_provider = formData.get('embedding_model_provider'); + + if (!embedding_model || !embedding_model_provider) { + return NextResponse.json( + { message: 'Missing embedding model or provider' }, + { status: 400 }, + ); + } + + const embeddingModels = await getAvailableEmbeddingModelProviders(); + const provider = + embedding_model_provider ?? Object.keys(embeddingModels)[0]; + const embeddingModel = + embedding_model ?? Object.keys(embeddingModels[provider as string])[0]; + + let embeddingsModel = + embeddingModels[provider as string]?.[embeddingModel as string]?.model; + if (!embeddingsModel) { + return NextResponse.json( + { message: 'Invalid embedding model selected' }, + { status: 400 }, + ); + } + + const processedFiles: FileRes[] = []; + + await Promise.all( + files.map(async (file: any) => { + const fileExtension = file.name.split('.').pop(); + if (!['pdf', 'docx', 'txt'].includes(fileExtension!)) { + return NextResponse.json( + { message: 'File type not supported' }, + { status: 400 }, + ); + } + + const uniqueFileName = `${crypto.randomBytes(16).toString('hex')}.${fileExtension}`; + const filePath = path.join(uploadDir, uniqueFileName); + + const buffer = Buffer.from(await file.arrayBuffer()); + fs.writeFileSync(filePath, new Uint8Array(buffer)); + + let docs: any[] = []; + if (fileExtension === 'pdf') { + const loader = new PDFLoader(filePath); + docs = await loader.load(); + } else if (fileExtension === 'docx') { + const loader = new DocxLoader(filePath); + docs = await loader.load(); + } else if (fileExtension === 'txt') { + const text = fs.readFileSync(filePath, 'utf-8'); + docs = [ + new Document({ pageContent: text, metadata: { title: file.name } }), + ]; + } + + const splitted = await splitter.splitDocuments(docs); + + const extractedDataPath = filePath.replace(/\.\w+$/, '-extracted.json'); + fs.writeFileSync( + extractedDataPath, + JSON.stringify({ + title: file.name, + contents: splitted.map((doc) => doc.pageContent), + }), + ); + + const embeddings = await embeddingsModel.embedDocuments( + splitted.map((doc) => doc.pageContent), + ); + const embeddingsDataPath = filePath.replace( + /\.\w+$/, + '-embeddings.json', + ); + fs.writeFileSync( + embeddingsDataPath, + JSON.stringify({ + title: file.name, + embeddings, + }), + ); + + processedFiles.push({ + fileName: file.name, + fileExtension: fileExtension, + fileId: uniqueFileName.replace(/\.\w+$/, ''), + }); + }), + ); + + return NextResponse.json({ + files: processedFiles, + }); + } catch (error) { + console.error('Error uploading file:', error); + return NextResponse.json( + { message: 'An error has occurred.' }, + { status: 500 }, + ); + } +} diff --git a/src/app/api/videos/route.ts b/src/app/api/videos/route.ts new file mode 100644 index 0000000..34ae7fd --- /dev/null +++ b/src/app/api/videos/route.ts @@ -0,0 +1,83 @@ +import handleVideoSearch from '@/lib/chains/videoSearchAgent'; +import { + getCustomOpenaiApiKey, + getCustomOpenaiApiUrl, + getCustomOpenaiModelName, +} from '@/lib/config'; +import { getAvailableChatModelProviders } from '@/lib/providers'; +import { BaseChatModel } from '@langchain/core/language_models/chat_models'; +import { AIMessage, BaseMessage, HumanMessage } from '@langchain/core/messages'; +import { ChatOpenAI } from '@langchain/openai'; + +interface ChatModel { + provider: string; + model: string; +} + +interface VideoSearchBody { + query: string; + chatHistory: any[]; + chatModel?: ChatModel; +} + +export const POST = async (req: Request) => { + try { + const body: VideoSearchBody = await req.json(); + + const chatHistory = body.chatHistory + .map((msg: any) => { + if (msg.role === 'user') { + return new HumanMessage(msg.content); + } else if (msg.role === 'assistant') { + return new AIMessage(msg.content); + } + }) + .filter((msg) => msg !== undefined) as BaseMessage[]; + + const chatModelProviders = await getAvailableChatModelProviders(); + + const chatModelProvider = + chatModelProviders[ + body.chatModel?.provider || Object.keys(chatModelProviders)[0] + ]; + const chatModel = + chatModelProvider[ + body.chatModel?.model || Object.keys(chatModelProvider)[0] + ]; + + let llm: BaseChatModel | undefined; + + if (body.chatModel?.provider === 'custom_openai') { + llm = new ChatOpenAI({ + openAIApiKey: getCustomOpenaiApiKey(), + modelName: getCustomOpenaiModelName(), + temperature: 0.7, + configuration: { + baseURL: getCustomOpenaiApiUrl(), + }, + }) as unknown as BaseChatModel; + } else if (chatModelProvider && chatModel) { + llm = chatModel.model; + } + + if (!llm) { + return Response.json({ error: 'Invalid chat model' }, { status: 400 }); + } + + const videos = await handleVideoSearch( + { + chat_history: chatHistory, + query: body.query, + }, + llm, + ); + + return Response.json({ videos }, { status: 200 }); + } catch (err) { + console.error(`An error occurred while searching videos: ${err}`); + return Response.json( + { message: 'An error occurred while searching videos' }, + { status: 500 }, + ); + } +}; diff --git a/src/app/c/[chatId]/page.tsx b/src/app/c/[chatId]/page.tsx new file mode 100644 index 0000000..aac125a --- /dev/null +++ b/src/app/c/[chatId]/page.tsx @@ -0,0 +1,9 @@ +import ChatWindow from '@/components/ChatWindow'; +import React from 'react'; + +const Page = ({ params }: { params: Promise<{ chatId: string }> }) => { + const { chatId } = React.use(params); + return ; +}; + +export default Page; diff --git a/ui/app/discover/page.tsx b/src/app/discover/page.tsx similarity index 98% rename from ui/app/discover/page.tsx rename to src/app/discover/page.tsx index eb94040..eb7de7f 100644 --- a/ui/app/discover/page.tsx +++ b/src/app/discover/page.tsx @@ -19,7 +19,7 @@ const Page = () => { useEffect(() => { const fetchData = async () => { try { - const res = await fetch(`${process.env.NEXT_PUBLIC_API_URL}/discover`, { + const res = await fetch(`/api/discover`, { method: 'GET', headers: { 'Content-Type': 'application/json', diff --git a/ui/app/favicon.ico b/src/app/favicon.ico similarity index 100% rename from ui/app/favicon.ico rename to src/app/favicon.ico diff --git a/ui/app/globals.css b/src/app/globals.css similarity index 100% rename from ui/app/globals.css rename to src/app/globals.css diff --git a/ui/app/layout.tsx b/src/app/layout.tsx similarity index 100% rename from ui/app/layout.tsx rename to src/app/layout.tsx diff --git a/ui/app/library/layout.tsx b/src/app/library/layout.tsx similarity index 100% rename from ui/app/library/layout.tsx rename to src/app/library/layout.tsx diff --git a/ui/app/library/page.tsx b/src/app/library/page.tsx similarity index 98% rename from ui/app/library/page.tsx rename to src/app/library/page.tsx index 379596c..9c40b2b 100644 --- a/ui/app/library/page.tsx +++ b/src/app/library/page.tsx @@ -21,7 +21,7 @@ const Page = () => { const fetchChats = async () => { setLoading(true); - const res = await fetch(`${process.env.NEXT_PUBLIC_API_URL}/chats`, { + const res = await fetch(`/api/chats`, { method: 'GET', headers: { 'Content-Type': 'application/json', diff --git a/ui/app/page.tsx b/src/app/page.tsx similarity index 100% rename from ui/app/page.tsx rename to src/app/page.tsx diff --git a/ui/app/settings/page.tsx b/src/app/settings/page.tsx similarity index 90% rename from ui/app/settings/page.tsx rename to src/app/settings/page.tsx index c659265..e97d189 100644 --- a/ui/app/settings/page.tsx +++ b/src/app/settings/page.tsx @@ -7,6 +7,7 @@ import { Switch } from '@headlessui/react'; import ThemeSwitcher from '@/components/theme/Switcher'; import { ImagesIcon, VideoIcon } from 'lucide-react'; import Link from 'next/link'; +import { PROVIDER_METADATA } from '@/lib/providers'; interface SettingsType { chatModelProviders: { @@ -20,6 +21,8 @@ interface SettingsType { anthropicApiKey: string; geminiApiKey: string; ollamaApiUrl: string; + lmStudioApiUrl: string; + deepseekApiKey: string; customOpenaiApiKey: string; customOpenaiApiUrl: string; customOpenaiModelName: string; @@ -55,6 +58,38 @@ const Input = ({ className, isSaving, onSave, ...restProps }: InputProps) => { ); }; +interface TextareaProps extends React.InputHTMLAttributes { + isSaving?: boolean; + onSave?: (value: string) => void; +} + +const Textarea = ({ + className, + isSaving, + onSave, + ...restProps +}: TextareaProps) => { + return ( +
+