Installation

Setting up and running locally

---

Tested with following configurations:

• Python 2.7.12, R i386 3.4.3 (32-bit), Linux 4.10.0-37-generic x86_64
• Python 2.7.12, R i386 3.4.3 (32-bit), Windows 10 Home (version 1709, build 16299.248)

---

Installing Dependencies

• Download SUMO & install required dependencies
• Download oeda branch of CrowdNav & install required dependencies
  • Resource to install setuptools
• Download and install ElasticSearch 6.2.4 and follow these instructions according to ElasticSearch 6.2.4. Installation can be verified by sending a GET request to (or navigating via web browser) http://localhost:9200/
• Download and install R.
  • For Linux see [this tutorial](https://www.digitalocean.com/community/tutorials/how-to-install-r-on-ubuntu-16-04-
• Download and install Node.js
• Run Zookeeper & Kafka
  • Create a host entry by adding 127.0.0.1 Kafka to /etc/hosts file
• Run CrowdNav and start simulation via GUI
• Installing R for Windows:
  • extend PATH variable with R's home. See these instructions. This and this might also be helpful.
  • Could not load the Visual C++ component fix
• Installing R for Linux:
  • Follow these instructions,

Installing OEDA

• Download master branch of OEDA

• Download and follow instructions in https://github.com/alinaciuysal/mlrMBO-API/

• In OEDA/Backend folder:

  • Download and install following libraries if your OS is Windows: (as they are around 80 MBs in total, we have not included them in the repository), also see and C++ compiler for Windows

    • Install pip and run pip install Tempita
    • Download following wheels and install them by running pip install wheel_name.whl
      • numpy‑1.13.3+mkl‑cp27‑cp27m Most libraries depend on this, so install this first.
      • scipy‑1.0.0‑cp27‑cp27m
      • scikit_learn‑0.19.1‑cp27‑cp27m
      • matplotlib‑2.2.0‑cp27‑cp27m
      • rpy2‑2.7.8‑cp27‑none
      • pandas‑0.22.0‑cp27‑cp27m
      • Cython‑0.27.3‑cp27‑cp27m

  • Run python setup.py install

  • Change /oeda/databases/user_db_config.json according to your preferences

  • Change /oeda/databases/experiment_db_config.json according to your preferences, e.g. if you would like to insert experiment data to another host/port

  • Run python server.py

  • In another terminal, navigate to tests/http-test-server and run node app.js. Make sure localhost:3003 is not being used beforehand

• In OEDA/Frontend folder:

  • Run npm install -g @angular/cli
  • Run npm install
  • Run serve.sh for Linux or serve.bat for Windows and navigate to localhost:4200

Running via Docker

---

Under development