From 748714d616123685477a8f22111c9c00e98b3547 Mon Sep 17 00:00:00 2001
From: Bluesheeptoken <louis.fruleux1@gmail.com>
Date: Tue, 12 Mar 2019 14:24:57 +0100
Subject: [PATCH] Updated README and added general guidelines to contribute

---
 Contribute.md | 46 +++++++++++++++++++++++++++++++++++++++++++
 README.md     | 54 ++++++++-------------------------------------------
 2 files changed, 54 insertions(+), 46 deletions(-)
 create mode 100644 Contribute.md

diff --git a/Contribute.md b/Contribute.md
new file mode 100644
index 0000000..4ee5a0f
--- /dev/null
+++ b/Contribute.md
@@ -0,0 +1,46 @@
+# Guidelines
+
+To compile, cython needs to be installed.
+
+## Tests
+### Run tests
+Pytest is used for tests
+
+`make test`
+
+### Generate coverage
+To generate coverage, you should use the coverage python module
+
+For the python code you can use `pytest --cov=cpt tests`
+
+## Linter
+pycodestyle and pylint are used for linter
+
+`make lint`
+
+## Sources
+
+## Data
+### Download files
+To download data, you will need to install lfs git extension
+
+## Profiling
+### Add metadata to metadata.json
+You should run `python generate_metadata.py <data_path> <datasetname>` from the data directory
+
+For instance, `python generate_metadata.py FIFA.dat partial_fifa`
+
+### Run profiling
+To run the profiling, you need to run the command `python profiling/profiling.py <mode> <data_path> <profile_path>`
+
+For instance, `python profiling/profiling.py train data/FIFA.dat profiling/sample_profiling.profile`
+
+The mode should be either train or predict
+
+The train profiles should be made with the full datasets, the predict profiles should be made with the partial datasets. The `predict` method is taking more time than the `train` method, so a smaller dataset is enough to profile `predict`
+
+### Read stats
+To read stats you need to use the [pstats](https://docs.python.org/3/library/profile.html) module in python. `python -m pstats <profile_path>`
+
+## Before pushing
+Make sure you ran `make test` and `make lint` before pushing
diff --git a/README.md b/README.md
index b1bd77e..a779989 100644
--- a/README.md
+++ b/README.md
@@ -1,5 +1,13 @@
 # CPT
 
+CPT is a cython open-source implementation of the Compact Prediction Tree algorithm using multithreading.
+
+This is an implementation of the following research papers
+
+http://www.philippe-fournier-viger.com/ADMA2013_Compact_Prediction_trees.pdf
+
+http://www.philippe-fournier-viger.com/spmf/PAKDD2015_Compact_Prediction_tree+.pdf
+
 ## Simple example
 
 You can test the model with the following code
@@ -49,49 +57,3 @@ unpickled_model = pickle.loads(dumped)
 
 print(model == unpickled_model)
 ```
-
-## Tests
-### Run tests
-Pytest is used for tests
-
-`make test`
-
-### Generate coverage
-To generate coverage, you should use the coverage python module
-
-For the python code you can use `pytest --cov=cpt tests`
-
-## Linter
-pycodestyle and pylint are used for linter
-
-`make lint`
-
-## Sources
-http://www.philippe-fournier-viger.com/ADMA2013_Compact_Prediction_trees.pdf
-
-http://www.philippe-fournier-viger.com/spmf/PAKDD2015_Compact_Prediction_tree+.pdf
-
-## Data
-### Download files
-To download data, you will need to install lfs git extension
-
-## Profiling
-### Add metadata to metadata.json
-You should run `python generate_metadata.py <data_path> <datasetname>` from the data directory
-
-For instance, `python generate_metadata.py FIFA.dat partial_fifa`
-
-### Run profiling
-To run the profiling, you need to run the command `python profiling/profiling.py <mode> <data_path> <profile_path>`
-
-For instance, `python profiling/profiling.py train data/FIFA.dat profiling/sample_profiling.profile`
-
-The mode should be either train or predict
-
-The train profiles should be made with the full datasets, the predict profiles should be made with the partial datasets. The `predict` method is taking more time than the `train` method, so a smaller dataset is enough to profile `predict`
-
-### Read stats
-To read stats you need to use the [pstats](https://docs.python.org/3/library/profile.html) module in python. `python -m pstats <profile_path>`
-
-## Before pushing
-Make sure you ran `make test` and `make lint` before pushing