• Create test for base selectors at src/selectors/index.js between line 18 and 31.
• Refactor base selectors with ramda.

This issue can ben two PRs if you want but, do the tests first to make sure nothing breaks.

Just to be clear, you must add test and refactor this code:

export const getNetwork = state => state.network;
export const getNodes = state => state.network.nodes || state.nodes || [];
export const getPositions = state => state.network.positions || state.positions || [];
export const getBeliefs = state => state.network.beliefs;
export const getSubnetworks = state => state.network.subnetworks || [];
export const getSelectedNodes = pipe(getNetwork, prop('selectedNodes'));
export const getNetworkKind = state => state.network.kind || NETWORK_KINDS.BN;
export const getPanelVisibility = state => state.network.propertiesPanelVisible;
export const getLinkages = state => state.network.linkages;
export const getInferenceEnabled = (state) => {
  const { inferenceEnabled } = getNetwork(state);

  return inferenceEnabled === undefined ? true : inferenceEnabled;
};

Github checklist for recommended community standards.

https://github.com/bayesjs/bayesjs-editor/community

• Code of conduct
• Contributing

Open Pull requests accordingly.

Add a new section above What is Bayes' Theorem with:

## What is a [Bayesian network](https://en.wikipedia.org/wiki/Bayesian_network)

A Bayesian network is a probabilistic graphical model (a type of statistical model) that represents a set of variables and their conditional dependencies via a directed acyclic graph (DAG). Bayesian networks are ideal for taking an event that occurred and predicting the likelihood that anyone of several possible known causes was the contributing factor. For example, a Bayesian network could represent the probabilistic relationships between diseases and symptoms. Given symptoms, the network can be used to compute the probabilities of the presence of various diseases.

And fix What is Bayes' Theorem text size

- # What is [Bayes' Theorem](https://en.wikipedia.org/wiki/Bayes%27_theorem)
+ ## What is [Bayes' Theorem](https://en.wikipedia.org/wiki/Bayes%27_theorem)

Refactor and add tests for the following actions at src/actions/index.js:

• removeNode
• addParent
• removeParent

Theses three actions don't need nodes payload anymore so, you can remove it and then create the missing test for these actions.

- export const removeNode = id => (dispatch, getState) => {
+ export const removeNode = id => (dispatch) => {
  dispatch({
    type: REMOVE_NODE,
-    payload: { id, nodes: getNodes(getState()) }
+    payload: { id },
  });

  dispatch(persistState());
};

Create tests for the following actions at action/index.js:

• persistState
• newNetwork
• loadNetwork
• changeNetworkProperty
• addNode
• changeNodeId
• changeNodeDescription
• changeNodePosition
• setBelief
• addSuperNode
• removeSuperNode
• addLinkage
• removeLinkage

You don't need create a test for these actions: removeNode, addParent, removeParent, changeNodeStates. I will change them and then I will create the tests.

You can create more than one PR if you wanna so, the PR will not be too big ;)

Here is a test code example you can follow:

import { NETWORK_KINDS } from 'constants/network';
import {
  persistState,
  newNetwork,
  loadNetwork,
  PERSIST_STATE,
  NEW_NETWORK,
  LOAD_NETWORK,
} from './index';

const persistStateAction = {
  type: PERSIST_STATE,
};

describe('Actions', () => {
  let dispatch;

  beforeAll(() => {
    dispatch = jest.fn();
  });

  beforeEach(() => {
    dispatch.mockClear();
  });

  describe('persistState', () => {
    let actionResult;

    beforeEach(() => {
      actionResult = persistState();
    });

    it('calls dispatch with type PERSIST_STATE', () => {
      expect(actionResult).toEqual(persistStateAction);
    });
  });

  describe('newNetwork', () => {
    describe('When passing no kind parameter', () => {
      beforeEach(() => {
        newNetwork()(dispatch);
      });

      it('calls dispatch with type NEW_NETWORK and kind BN', () => {
        expect(dispatch).toHaveBeenNthCalledWith(1, {
          type: NEW_NETWORK,
          kind: NETWORK_KINDS.BN,
        });
      });

      it('calls dispatch persist action', () => {
        expect(dispatch).toHaveBeenNthCalledWith(2, persistStateAction);
      });
    });

    describe('When passing with kind parameter as MSBN', () => {
      beforeEach(() => {
        newNetwork(NETWORK_KINDS.MSBN)(dispatch);
      });

      it('calls dispatch with type NEW_NETWORK and kind BN', () => {
        expect(dispatch).toHaveBeenNthCalledWith(1, {
          type: NEW_NETWORK,
          kind: NETWORK_KINDS.MSBN,
        });
      });

      it('calls dispatch persist action', () => {
        expect(dispatch).toHaveBeenNthCalledWith(2, persistStateAction);
      });
    });
  });

  describe('loadNetwork', () => {
    const state = { key: 'value' };

    beforeEach(() => {
      loadNetwork(state)(dispatch);
    });

    it('calls dispatch with type LOAD_NETWORK and payload with state', () => {
      expect(dispatch).toHaveBeenNthCalledWith(1, {
        type: LOAD_NETWORK,
        payload: { state },
      });
    });

    it('calls dispatch persist action', () => {
      expect(dispatch).toHaveBeenNthCalledWith(2, persistStateAction);
    });
  });
});

• RenderIntoBody
• storage middleware

• Create tests for src/reducers/network/subnetworks.js. PR reference: #39
• Refactor src/reducers/network/subnetworks.js.

This issue can ben two PRs but, do the tests first to make sure nothing breaks.
Also, you can use ramda to refactor:

// current
case LOAD_NETWORK: {
  const { subnetworks } = action.payload.state.network;
  return subnetworks || [];
}
// with Ramda
import { path, defaultTo } from 'ramda';

const pathPayloadStateNetwork = path(['payload', 'state', 'network'])
const defaultToEmptyArray = defaultTo([]);

case LOAD_NETWORK: {
  return defaultToEmptyArray(pathPayloadStateNetwork(action))
}
// or
const onLoadNetwork = pipe(pathPayloadStateNetwork, defaultToEmptyArray);
case LOAD_NETWORK: {
  return onLoadNetwork(action)
}

Other details about this reducer:

• There is a method called colorsNol that should be something like getRandomColor. I believe you can extract this to a new utils like /src/utils/colors or use a package for it. Just to be clear: a super-node is a sub-network that will receive a random color. Yes, I'm awere that super-node is a really bad name but I will change it in another PR.
• There is a condition that I don't know if is valid: if (nodes && positions) {. This was made because we can receive two different networks: the current format and the old.

networks.zip

In this .zip you can find the two networks. You can move theses networks to a __fixtures__ folder and then use it in your tests.

• first-version-network.json - the first/legacy version
• simple-network.json - the current version

Well, thats it 😄

Update webpack to version 4.41.0 using yarn due mixin-deep and set-value vulnerability.
https://nvd.nist.gov/vuln/detail/CVE-2019-10746
https://nvd.nist.gov/vuln/detail/CVE-2019-10747

• Create tests for src/reducers/network.js.
• Refactor src/reducers/network.js.

This reducer is not 100% pure in some cases, the current state is being changed:

if (network.kind === undefined) network.kind = NETWORK_KINDS.BN;
if (network.id === undefined) network.id = v4();

Anyway, would be great use ramda to refactor this reducer and extract all the possible logic from there.

Update README to:

- **P(A | B) = (P(B | A).P(A)) / P(B)**
+ ![Bayes' theorem formula](https://wikimedia.org/api/rest_v1/media/math/render/svg/87c061fe1c7430a5201eef3fa50f9d00eac78810)

And fix the link:

- * P(A | B) is a conditional probability(https://en.wikipedia.org/wiki/Conditional_probability): the likelihood of event A occurring given that B is true.
+ * P(A | B) is a [conditional probability](https://en.wikipedia.org/wiki/Conditional_probability): the likelihood of event A occurring given that B is true.

When there're some modifications on some node's states, parents, children, etc... the software should warn the user that the CPTs must be checked.

Currently, CHANGE_NETWORK_PROPERTY changes the following network's properties:

• name
• description
• width
• height
• inferenceEnabled
• propertiesPanelVisible

I think would be better to create a more explicit action, for example, instead of CHANGE_NETWORK_PROPERTY would be better to have a UPDATE_NETWORK_NAME.
Follow the name example, to accomplish this change you must:

Create the action

src/actions/network.js

export const onUpdateNetworkName = (name) => (dispatch) => {
  dispatch({
    type: UPDATE_NETWORK_NAME,
    payload: { name },
  });

  dispatch(persistState());
};

Remember to add this src/actions/network.js at src/actions/index.js.

Create the action test

Check src/actions/index.test.js example ;)

Update network's name reducer

src/reducers/network/name.js

- const updateProperty = updateNetworkProperty('name');
+ const getName = path(['payload', 'name']);

export default (state = defaultValue, action) => {
  switch (action.type) {
    case LOAD_NETWORK:
      return getNetworkName(action);
    case NEW_NETWORK:
      return defaultValue;
-    case CHANGE_NETWORK_PROPERTY:
-      return updateProperty(state, action);
+    case UPDATE_NETWORK_NAME:
+      return getName(action);
    default:
      return state;
  }
};

Update network's name reducer

src/reducers/network/name.test.js

describe('Network Name Reducer', () => {
  ...
-  describe('CHANGE_NETWORK_PROPERTY', () => {
-    describe('When property name is "name"', () => {
-      it('returns new value', () => {
-        expect(
-          reducer(
-            'old name',
-            {
-              type: CHANGE_NETWORK_PROPERTY,
-              payload: { name: 'name', value: 'new name' },
-            },
-          ),
-        ).toBe('new name');
-      });
-    });
-
-    describe('When property name is not "name"', () => {
-      it('returns current state value', () => {
-        expect(
-          reducer(
-            'old name',
-            {
-              type: CHANGE_NETWORK_PROPERTY,
-              payload: { name: 'other', value: 'new name' },
-            },
-          ),
-        ).toBe('old name');
-      });
-    });
-  });

+  describe('UPDATE_NETWORK_NAME', () => {
+    it('returns new value', () => {
+      expect(
+        reducer(
+          'old name',
+          {
+            type: CHANGE_NETWORK_PROPERTY,
+            payload: { name: 'name', value: 'new name' },
+          },
+        ),
+      ).toBe('new name');
+    });
+  });
});

Hey @fhelwanger ,

I think would be great if you set a Description and Website for this repository, something like:
Description - An simple editor for Bayesian Networks
Website - http://fhelwanger.github.io/bayesjs-editor/

• Create tests for src/reducers/nodes.js. PR reference: #39
• Refactor src/reducers/nodes.js.

To test it properly you will need a network so, you can use these networks: networks.zip
You can create a folder __fixtures__ and paste these JSONs there.

You can extract hasCycles to a utils file.
There is a better hasCycles method in src/components/NetworkMSBN/helpers.js. You can use it but extract it first to a utils file and test it with a network.

Until a MVP is released, I'm not giving too much attention to validations (duplicate ids, max length, required, etc.).

There are a lot of errors on lint:

I will create a PR to fix all this errors to keep this project up to date.

Instead of showing numbers, show bars that have a width relative to its probability.

• It should be optional (an option to toggle it)
• It should have a tooltip to show the value

Add more details about bayes' theorem.

Update:

# BayesJS Editor

An editor for Bayesian Networks built in React the use bayes' theorem.

Add

# What is bayes' theorem

In probability theory and statistics, Bayes’ theorem describes the probability of an event, based on prior knowledge of conditions that might be related to the event. For example, if cancer is related to age, then, using Bayes’ theorem, a person's age can be used to more accurately assess the probability that they have cancer than can be done without knowledge of the person’s age.

Also, you can add links from Wikipedia like:

• https://en.wikipedia.org/wiki/Bayes%27_theorem
• https://en.wikipedia.org/wiki/Probability_theory
• https://en.wikipedia.org/wiki/Statistics

Use Cypress to test essential things like:

• Create a node
• Edit CPTs
• Add/Remove states
• Open a network
• Change node name
  ...

I'm already working on it.

IE 11 doesn't support Array#find.

Add polyfill for it and explore if something else is broken.

I think would be great to have why did you update running in DEV.

In this project, we already had the error of render unnecessary thing like, render all network again just because of one node change it's position. So, to make sure we don't have this kind of problem again this package can help us :)

• Create tests for src/reducers/linkages.js. PR reference: #39
• Refactor src/reducers/linkages.js.

This issue can ben two PRs but, do the tests first to make sure nothing breaks.
Also, you can use ramda to refactor:

// current
case REMOVE_LINKAGE: {
  const { id } = action.payload;
  const newState = { ...state };
  delete newState[id];

  return newState;
}
// with Ramda
import { dissoc } from 'ramda';
... 
case REMOVE_LINKAGE: {
  const { id } = action.payload;

  return dissoc(id, state);
}

When there are multiple arrows coming from and to the same point, it's very confusing.

The arrows on the SPRINKLER node illustrates this.