updated readmes and changed gitignore

.github/workflows/pipeline.yml

@@ -28,6 +28,7 @@ jobs:
       - name: flake8 Lint
         uses: py-actions/flake8@v2
         with:
+          ignore: "E402"
           max-line-length: "120"
           path: .
       - uses: actions/checkout@v2
@@ -83,9 +84,10 @@ jobs:
           npm install
           npm run build
           npm run deploy
+      - uses: actions/checkout@v2
       - name: Bump version and push tag
         uses: anothrNick/[email protected]
-        if: ${{ github.event_name == 'push' && !contains(join(github.event.commits.*.message, ' '), '#skip') }}
+        if: ${{ !contains(join(github.event.commits.*.message, ' '), '#skip') }}
         env:
           GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
           DEFAULT_BUMP: patch

.gitignore

@@ -0,0 +1,24 @@
+#common
+.vscode/
+__pycache__/
+venv/
+.env
+
+#train_and_predict
+labeled*/
+unlabeled*/
+*.h5
+
+#preprocess
+ts/
+temp/
+unlabeled*/
+
+#webapp
+node_modules/
+dist/
+.DS_Store
+
+#api
+app.db
+test.db

README.md

@@ -21,7 +21,7 @@ The tool has been divided into different modules:
 -   **Preprocessing of unlabeled data:** it contains a Python script that generates mp3 files and spectrograms, given a directory containing audio data.
 -   **A Machine Learning Endpoint:** which is a small flask app that -- given an h5 file, a labeled dataset and an unlabeled dataset -- trains an ML model on the labeled data and predicts on the unlabeled data.
 
-Go into each directory to learn how to get started with it.
+Go into each directory to learn how to get started with it. You will want to visit the directories in this order: preprocess --> train_and_predict --> api --> webapp.
 
 **Flowchart of how the tool works:**  
 ![flowchart](api/assets/flowchart.jpg)

api/README.md

@@ -215,7 +215,7 @@ headers: {
 
 # Getting Started
 
-This API requires a database and a ML endpoint to run, the easiest way to do that would be run the ML endpoint's docker container by following the [train_and_predict quick method](../train_and_predict/README.md#quick-method), and to start a docker postgres database with the following command: `docker run --name postgres -p 5432:5432 -e POSTGRES_DB=orcagsoc -e POSTGRES_PASSWORD=<database-password> -d postgres`.
+This API requires a database and a ML endpoint to run, the easiest way to do that would be run the ML endpoint's docker container by following the [train_and_predict quick method](../train_and_predict/README.md#quick-method), and to start a docker postgres database with the following command: `docker run --name postgres -p 5432:5432 -e POSTGRES_DB=orcagsoc -e POSTGRES_PASSWORD=<database-password> -d postgres`. After these steps, choose either the quick method or flexible method to proceed.
 
 ### Quick Method
 
@@ -236,7 +236,7 @@ This API requires a database and a ML endpoint to run, the easiest way to do tha
     ML_ENDPOINT_URL=http://127.0.0.1:5001
     S3_LABELED_PATH=s3://orcagsoc/labeled_test/
     S3_UNLABELED_PATH=s3://orcagsoc/unlabeled_test/
-    S3_MODEL_PATH=s3://orcagsoc/models/srkw_cnn_0.h5
+    S3_MODEL_PATH=s3://orcagsoc/models/srkw_0.h5
     RETRAIN_TARGET=20
     IMG_WIDTH=607
     IMG_HEIGHT=617
@@ -246,6 +246,7 @@ This API requires a database and a ML endpoint to run, the easiest way to do tha
 -   Run `flask db upgrade` to update the tables of the database
 -   `flask run` starts a development server in http://localhost:5000
 -   If you plan to contribute, please configure your text editor / IDE to use Flake8 to lint and YAPF to format Python code
+-   N.B. If pipenv does not work for you, a requirements.txt file has been provided which you can use with pip to install packages instead
 
 ### Testing
 

api/app/__init__.py

@@ -1,8 +1,3 @@
-# Handle circular imports
-from .active_learning import train_and_predict
-# from app import routes, models
-from app.models import LabeledFile, Model, Prediction, ConfusionMatrix, Accuracy
-
 from flask import Flask
 from flask_cors import CORS
 from flask_sqlalchemy import SQLAlchemy
@@ -35,6 +30,11 @@
 db = SQLAlchemy(app)
 migrate = Migrate(app, db)
 
+# Handle circular imports
+from .active_learning import train_and_predict
+# from app import routes, models
+from app.models import LabeledFile, Model, Prediction, ConfusionMatrix, Accuracy
+
 if not app.debug or os.environ.get("WERKZEUG_RUN_MAIN") == "true":
     # The app is not in debug mode or we are in the reloaded process
     # Start training if the tables generated after each training round are empty

webapp/README.md

@@ -6,7 +6,7 @@
 
 ## Getting Started
 
--   Make sure [Node.js](http://nodejs.org) is installed
+-   Make sure [Node.js v14.x](http://nodejs.org) is installed
 -   Clone the repo and `cd` into the project directory
 -   Run `npm install` to install all the dependencies
 -   Run `npm start` to start a development server in [http://localhost:8080](http://localhost:8080) (For the app to work correctly, you need to start the API server as well)