This AMI_Project is the Project of Group_7.
In order to run the function without causing any error, you should garantee that you are in the university internet environment, and the programm will get the token and doing the forecast automatically.
If you are not in university or in a studentenwerk, please commend line 23 and 24 in Subfuntions.py and line 171 and 172 in Subfunction_LSTM.py,
then type in the newest token and it should be fine.
In designing the web page, our group has already come up with the plotting value function (1h, 1d, and 1week) and also plotting the whole trend for 1 week, so there is no need for client.py. If you use client.py, it may cause error.
This project is released under the MIT License
We will upload documents in the moodle.
We use the followed dataset to train and test our model.
- Electricity price data from Montel
- Electricity production and consumption from Smard
- Weather data from DWD
git clone https://gitlab.ldv.ei.tum.de/ami2021/group07.git
We use python3.7 and Pytorch 1.8.1, and all of packages we used are in requirements.txt.
Before using dockerfile to build a new image, make sure you are in uni or studentenwerk. Otherwise, you need to modify the token code according to the beginning of the readme, and then build the dockerfile on your computer to install our project:
docker build -t image_name
Or you can easily use the following command to pull the image we created (make sure you are using uni or studentenwerk network)
docker pull eliasliu/electricity
After building a docker image, you can use the followed code to run our model
docker run -it -p 8888:8888 image_name
Then the model will run, and you should open your browser, run localhost:8888 in your browser (don't copy and use the links generated in docker container)