# How to Build TrendRadar from Source: Complete Docker Build Guide > Build TrendRadar from source using Docker Compose. Follow our guide to compile Python dependencies and create custom Docker images for news crawling and AI analysis. Get started today. - Repository: [sansan/TrendRadar](https://github.com/sansan0/TrendRadar) - Tags: how-to-guide - Published: 2026-04-22 --- **Build TrendRadar locally using Docker Compose with a multi-stage Dockerfile that compiles Python 3.12 dependencies via uv and produces two images: `wantcat/trendradar` for news crawling and `wantcat/trendradar-mcp` for AI analysis.** TrendRadar is an open-source news aggregation and push-notification service distributed as containerized images. This guide walks through the complete **TrendRadar build steps** from the official repository, covering cloning, multi-stage Docker builds, architecture targeting, and deployment. --- ## Prerequisites Before building TrendRadar, ensure you have: - **Docker** 20.10+ with Buildx support - **Docker Compose** plugin (v2.0+) - **Git** for cloning the repository The build process uses **uv** (a fast Python package installer) inside the container, so no local Python installation is required. --- ## Step 1: Clone the TrendRadar Repository Retrieve the complete source tree including Docker assets, configuration templates, and the `trendradar` Python package: ```bash git clone https://github.com/sansan0/TrendRadar.git cd TrendRadar ``` The repository structure includes: - `docker/` — Build and runtime configurations - `config/` — Runtime settings and keyword filters - `trendradar/` — Core Python source code - [`pyproject.toml`](https://github.com/sansan0/TrendRadar/blob/main/pyproject.toml) and `uv.lock` — Locked dependency manifests --- ## Step 2: Select the Build-Ready Compose File TrendRadar provides two compose configurations. Switch from the runtime version (pulls pre-built images) to the build version (compiles locally): ```bash cd docker cp docker-compose-build.yml docker-compose.yml ``` The [`docker-compose-build.yml`](https://github.com/sansan0/TrendRadar/blob/main/docker-compose-build.yml) file defines build contexts for both services with explicit `build:` sections targeting [`docker/Dockerfile`](https://github.com/sansan0/TrendRadar/blob/master/docker/Dockerfile). --- ## Step 3: (Optional) Set Target Architecture Control which **supercronic** binary (internal job scheduler) is downloaded by setting the `DOCKER_ARCH` environment variable: ```bash # For Apple Silicon, Raspberry Pi, or other ARM64 devices export DOCKER_ARCH=arm64 # For standard x86_64 servers (default if unset) export DOCKER_ARCH=amd64 ``` This variable is consumed in the Dockerfile at lines 13-22, which conditionally downloads the appropriate supercronic release. --- ## Step 4: Build the TrendRadar Docker Images Execute the multi-stage build. You can build both services or target them individually: ```bash # Build both images simultaneously docker compose build # Or build specific services docker compose build trendradar docker compose build trendradar-mcp ``` The multi-stage build in `docker/Dockerfile` performs: 1. **Stage 1** — Download supercronic for the target architecture 2. **Stage 2** — Copy the `uv` binary, install locked Python dependencies via `uv sync --locked` 3. **Stage 3** — Copy project source code and install the `trendradar` package itself The `uv.lock` file ensures **reproducible builds** with exact dependency versions. --- ## Step 5: Deploy and Run the Containers Start the services in detached mode: ```bash # Run both crawler and AI analysis server docker compose up -d # Or run only the news crawler docker compose up -d trendradar # Or run only the MCP AI server docker compose up -d trendradar-mcp ``` Service behavior: - **trendradar** — Launches `supercronic` via [`docker/entrypoint.sh`](https://github.com/sansan0/TrendRadar/blob/main/docker/entrypoint.sh) to run the scheduled news crawler - **trendradar-mcp** — Opens an HTTP API on `127.0.0.1:3333` for AI-powered content analysis --- ## Step 6: Verify and Manage the Build Confirm successful deployment and interact with running services: ```bash # Check container status docker compose ps # View crawler logs in real-time docker logs -f trendradar # Check service health status docker exec -it trendradar python manage.py status # Trigger manual crawl immediately docker exec -it trendradar python manage.py run ``` The [`docker/manage.py`](https://github.com/sansan0/TrendRadar/blob/main/docker/manage.py) script provides a CLI interface for common operational tasks without restarting containers. --- ## Key Build Files Reference | File | Purpose | Location | |------|---------|----------| | `docker/Dockerfile` | Multi-stage build definition with supercronic, uv, and package installation | [`docker/Dockerfile`](https://github.com/sansan0/TrendRadar/blob/master/docker/Dockerfile) | | [`docker/docker-compose-build.yml`](https://github.com/sansan0/TrendRadar/blob/main/docker/docker-compose-build.yml) | Compose configuration for local builds | [[`docker/docker-compose-build.yml`](https://github.com/sansan0/TrendRadar/blob/main/docker/docker-compose-build.yml)](https://github.com/sansan0/TrendRadar/blob/master/docker/docker-compose-build.yml) | | [`docker/entrypoint.sh`](https://github.com/sansan0/TrendRadar/blob/main/docker/entrypoint.sh) | Container entrypoint launching supercronic | [[`docker/entrypoint.sh`](https://github.com/sansan0/TrendRadar/blob/main/docker/entrypoint.sh)](https://github.com/sansan0/TrendRadar/blob/master/docker/entrypoint.sh) | | [`docker/manage.py`](https://github.com/sansan0/TrendRadar/blob/main/docker/manage.py) | Operational CLI for status, manual runs, and logs | [[`docker/manage.py`](https://github.com/sansan0/TrendRadar/blob/main/docker/manage.py)](https://github.com/sansan0/TrendRadar/blob/master/docker/manage.py) | | [`pyproject.toml`](https://github.com/sansan0/TrendRadar/blob/main/pyproject.toml) / `uv.lock` | Locked Python dependencies for reproducible builds | [[`pyproject.toml`](https://github.com/sansan0/TrendRadar/blob/main/pyproject.toml)](https://github.com/sansan0/TrendRadar/blob/master/pyproject.toml) | --- ## Summary - **Clone** the repository from `https://github.com/sansan0/TrendRadar.git` - **Switch** to the build compose file: `cp docker/docker-compose-build.yml docker/docker-compose.yml` - **Optionally set** `DOCKER_ARCH=arm64` for ARM64 builds - **Build** with `docker compose build` (multi-stage: supercronic, uv deps, package install) - **Deploy** with `docker compose up -d` and verify with `docker logs` and [`manage.py`](https://github.com/sansan0/TrendRadar/blob/main/manage.py) The entire build process is containerized, requiring only Docker and producing reproducible images for both the news crawler and optional AI analysis server. --- ## Frequently Asked Questions ### What is the minimal command to build and run TrendRadar? Clone, switch compose files, build, and start: ```bash git clone https://github.com/sansan0/TrendRadar.git && cd TrendRadar cd docker && cp docker-compose-build.yml docker-compose.yml docker compose build docker compose up -d ``` ### Can I build TrendRadar for ARM64 devices like Raspberry Pi? Yes. Set `export DOCKER_ARCH=arm64` before building. The Dockerfile automatically downloads the ARM64 version of supercronic. Omit this variable for default AMD64 builds. ### How do I verify the build succeeded without running the full service? Check image creation with `docker images | grep trendradar`, then inspect a single layer: `docker run --rm wantcat/trendradar python -c "import trendradar; print('OK')"`. The container should exit cleanly with confirmation output. ### What is the difference between the trendradar and trendradar-mcp images? The `trendradar` image runs the scheduled news crawler and push-notification service using supercronic. The `trendradar-mcp` image (built from the same Dockerfile with a different target) provides an optional AI analysis server exposing an HTTP API on port 3333.