Merge pull request #162 from bagustris/master

Add balancing for finetune and update data README

.github/workflows/format_code.yml

@@ -0,0 +1,12 @@
+name: Check code formatting
+
+on: [push, pull_request]
+
+jobs:
+  lint:
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@v4
+      - uses: psf/black@stable
+      # - uses: psf/black@552baf822992936134cbd31a38f69c8cfe7c0f05
+

.github/workflows/isort.yaml

@@ -0,0 +1,16 @@
+name: Run isort
+
+on: [push, pull_request]
+
+jobs:
+  build:
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@v3
+      - uses: isort/isort-action@v1
+      # - uses: isort/isort-action@master
+        with:
+            # isortVersion: 5.13.2
+            sortPaths: 'nkululeko'
+            configuration: '--profile black'
+

CONTRIBUTING.md

@@ -21,7 +21,7 @@ The preferred way to contribute to nkululeko is to fork the [main repository](ht
 
 	```bash
 	git clone https://github.com/YourLogin/nkululeko.git
-	cd spafe
+	cd nkululeko
 	```
 
 3.	Remove any previously installed nkululeko versions, then install your local copy with testing dependencies:
@@ -43,9 +43,14 @@ The preferred way to contribute to nkululeko is to fork the [main repository](ht
 	-> Please never work directly on the `master` branch!
 	```
 
-6.	Once you are done, make sure to format the code using black to fit spafe's codestyle.
+6.	Once you are done, make sure to format the code using black to fit Nkululeko's codestyle.
 
-	```black nkululeko/```
+	```bash
+	black nkululeko/
+	isort --profile black nkululeko/
+	# Alternatively and additionaly, use ruff:
+	ruff check --fix --output-format=full nkululeko
+	```
 
 7.	Make sure that the tests succeed and have enough coverage.
 

data/README.md

@@ -5,7 +5,7 @@ Nkululeko database repository
 # Data
 
 
-This is the default top directory for Nkululeko data import.Each database should be in its own subfolder (you can also use `ln -sf`` to soft link original database path to these subfolders) and contain a README how to import the data to Nkululeko CSV or audformat.
+This is the default top directory for Nkululeko data import. Each database should be in its own subfolder (you can also use `ln -sf` to soft link original database path to these subfolders) and contain a README how to import the data to Nkululeko CSV or audformat.
 ## Accessibility
 
 
@@ -15,12 +15,10 @@ The column `access` in the table below indicates the database's accessability. T
 - `private`: the database is not publicly available on the internet and requires the private information of the owner of the dataset.
 
 
-To support open science and reproducible research, we only accept PR and recipes for public dataset for now on.
-## Databases
-
+To support open science and reproducible research, we encourage to submit PR and recipes for public dataset for now on.
 |Name|Target|Description|Access|
 | :--- | :--- | :--- | :--- |
-|emorynlp|emotion|English, From Friends TV|public|
+|emorynlp|emotion|English Emotion Dataset from Friends TV Show|public|
 |emns|emotion,intensity|British, singles peaker, UAR=.479|public|
 |test|none|Test data for nkululeko|public|
 |catsvsdogs|cats_dogs|kaggle test set|public|
@@ -72,11 +70,13 @@ To support open science and reproducible research, we only accept PR and recipes
 |urdu|emotion|Urdu|public|
 |polish|emotion|Polish|public|
 |cmu-mosei|sentiment,emotion|English, original link dead|public|
-|SVD|pathologicalspeech|German|public|
+|svd|pahtological speech|German speech data for detecting various pathological voices|public|
 |msp-improv|emotion,VAD,naturalness|English|restricted|
 |shemo|emotion|Persian|public|
 |esd|emotion|English,Chinese|public|
 
+
+This recipe contains information about 56 datasets.
 ## Performance
 
 ![Nkululeko performance](../meta/images/nkululeko_ser_20240719.png)  

data/androids/process_database.py

@@ -15,9 +15,10 @@
 
 """
 
-import pandas as pd
 import os
+
 import audeer
+import pandas as pd
 
 dataset_name = 'androids'
 data_root = './Androids-Corpus/'

data/banglaser/process_database.py

@@ -14,12 +14,14 @@
 GG = Actor ID, 01-34 (odd: male, even: female)
 """
 
-import pandas as pd
-from nkululeko.utils.files import find_files
 import argparse
-from sklearn.model_selection import train_test_split
 from pathlib import Path
 
+import pandas as pd
+from sklearn.model_selection import train_test_split
+
+from nkululeko.utils.files import find_files
+
 
 def process_database(data_dir, output_dir):
     # check if data_dir exists

data/crema-d/load_db.py

@@ -2,7 +2,6 @@
 
 import audb
 
-
 # set download directory to current
 cwd = os.getcwd()
 audb.config.CACHE_ROOT = cwd