Sam Allardyce walks into the video analysis room.. “Morning, I was watching the last Arsenal match and there were a few situations I saw that I thought they would hurt us on Saturday. I want you to find me all of these situations from their last 5 games and find the times in our last 5 matches that we have faced similar situations. Let me have it this afternoon. Thanks” …….“Gaffer, that is going…” Big Sam slams the door “… to take hours… of work”.

Based on work by Joe Mulberry

Demo: https://pydata-situsearch.herokuapp.com/

To get started with this repository make sure you clone it using git clone https://github.com/teamtv/pydata-situsearch.git.

After making a local clone you it's preferred to create a virtual environment to keep you machine clean.

Then run:

pip install -r requirements.txt
pip install -r requirements-local.txt

For this demo we use open tracking data provided by Metrica Sports. You can find more information about their datasets here: https://github.com/metrica-sports/sample-data

cd data/raw
wget "https://github.com/metrica-sports/sample-data/blob/master/data/Sample_Game_1/Sample_Game_1_RawTrackingData_Home_Team.csv?raw=true" -O sample1_home.csv
wget "https://github.com/metrica-sports/sample-data/blob/master/data/Sample_Game_1/Sample_Game_1_RawTrackingData_Away_Team.csv?raw=true" -O sample1_away.csv

This project is able to run on your local machine without any additional dependencies like servers, daemons, brokers, etc.

You can fill your local repository using the command line tool. The tool will parse the Metrica files and write the serialized files on your local disk.

cd src
python cmdline.py fill-local-repository

Open src/flask_api/__init__.py and make sure you use the local repository.

You can either use gunicorn or debug werkzeug server. To start the debug werkzeug server run:

cd src
python run_app.py

Learn more about the code itself

In favor of sharing knowledge about the domain between developers and domain experts, we create domain models. These models helps us to have a common language.

We try to use names that teach others about the domain. In the case of SituSearch we can identity three core tracking models:

1. TrackingDataset
2. Frame
3. Point

For the search domain there are two models:

1. ResultSet
2. Result

Next to the models there are domain services. These services provide the core of application.

In the tracking domain we have some interfaces:

1. Matcher
2. ReferenceMatcher And one implementation: MunkresMatcher

The search domain contains one service: SearchEngine

To make it possible to interact with the outside world we need an application layer. This layer provides an easy way to use the domain layers. In this project we only have a SearchApplicationService.

Next to the application layer there is the infrastructure layer. This layer makes it possible to connect to external systems like database or file systems. The infrastructure layer contains some more types of components.

For parsing Metrica data we can use the MetricaParser

For converting a TrackingDataset to json (to send to the client) we can use the DatasetToJson.

There are two repositories in this project:

1. LocalRepository
2. S3Repository Both implement the same Repository interface

We would like to expose the frames of dataset, and the search results to the outside world. We use flask to do so.

The flask app can be found here

Make sure you have an heroku account. You can sign up for free here

Install the heroku toolbelt to use their CLI. Read more about how to link your local repository with their build platform.

To deploy the application just do git push heroku master

Checkout the logging: heroku logs --tail

Validate you api is working: curl https://pydata-situsearch.herokuapp.com/datasets/test/frames | less

Make sure you have configured your aws keys: AWS_ACCESS_KEY_ID and AWS_SECRET_ACCESS_KEY

Normally a React frontend requires a lot of building tools. For this project we decides to do use babel within the browser. This makes the deployment way easier. Keep in mind you should use a tool like create-react-app when you want to build a real application.

The entire frontend can be found here.