Sample Code to use Keras Model

https://keras.io/api/applications/

Remove Sensitive Environment Variable / File That is too Big from Remote Repo

In case mistakenly pushed a large file/sensitive environment data and git rm --cached does not help (pushed to the remote repo but just untracked locally):

git filter-branch --force --index-filter \
  "git rm --cached --ignore-unmatch <path to your file>" \
  --prune-empty --tag-name-filter cat -- --all

(cd to the top level of the repo first), add -r flag if you want to remove the whole directory. REF: 

https://docs.github.com/.../removing-sensitive-data-from... 

This error may subsequently follow: fatal: refusing to merge unrelated histories Then

git pull origin the-remote-branch --allow-unrelated-histories

and resolve conflicts.

Examine Output Size in Tensorflow

When we are uncertain the output size of our tensor processed by some layer, we can go through the following:

x = tf.constant([[1, 1., 1., 2., 3.],
                 [1, 1., 4., 5., 6.],
                 [1, 1., 7., 8., 9.],
                 [1, 1., 7., 8., 9.],
                 [1, 1., 7., 8., 9.]])

x = tf.reshape(x, [1, 5, 5, 1])

print(MaxPool2D((5, 5), strides=(2, 2),  padding="same")(x))
print(math.ceil(5/2))

which yields

print(MaxPool2D((5, 5), strides=(2, 2),  padding="same")(x))
tf.Tensor(
[[[[7.]
   [9.]
   [9.]]

  [[7.]
   [9.]
   [9.]]

  [[7.]
   [9.]
   [9.]]]], shape=(1, 3, 3, 1), dtype=float32)

3

For layer that has training weight, we may try the following for testing:

model = Conv2D(3, (3, 3), strides=(2, 2), padding="same", kernel_initializer=tf.constant_initializer(1.))


x = tf.constant([[1., 2., 3., 4., 5.],
                 [1., 2., 3., 4., 5.],
                 [1., 2., 3., 4., 5.],
                 [1., 2., 3., 4., 5.],
                 [1., 2., 3., 4., 5.]])

x = tf.reshape(x, (1, 5, 5, 1))
print(model(x))

which yields

x = tf.constant([[1., 2., 3., 4., 5.],...
tf.Tensor(
[[[[ 6.  6.  6.]
   [18. 18. 18.]
   [18. 18. 18.]]

  [[ 9.  9.  9.]
   [27. 27. 27.]
   [27. 27. 27.]]

  [[ 6.  6.  6.]
   [18. 18. 18.]
   [18. 18. 18.]]]], shape=(1, 3, 3, 3), dtype=float32)

In fact it can be proved in both MaxPooling2D and Conv2D that if stride $=s$ and padding$=$same, then 

\[\text{output_width} = \left\lfloor\frac{\text{input_width}-1}{s}\right\rfloor + 1 = \left\lceil\frac{\text{input_width}}{s}\right\rceil\]

The last equality deserves a proof as it is not highly trivial:

Fact. For any positive intergers $w,s$, we have \[ \left\lfloor \frac{w-1}{s}\right\rfloor + 1 = \left\lceil \frac{w}{s}\right\rceil. \] Proof. We do case by case study. If $w=ks$ for some positive $k\in \N$, then \[\text{LHS} = \left\lfloor k - \frac{1}{s}\right\rfloor +1 = (k-1)+1=k = \lceil k\rceil = \text{RHS}. \] When $w=ks+j$, for some $k\in\N$ and $j\in \N \cap (0, s)$, then \[ \text{LHS} = \left\lfloor k+\frac{j-1}{s}\right\rfloor + 1 = k+1 = \left\lceil k+\frac{j}{s}\right\rceil = \left\lceil \frac{ks+j}{s}\right\rceil = \left\lceil\frac{w}{s}\right\rceil=\text{RHS}.\qed \]

conda virtual environment command

conda create --name tensorflow python=3.7
conda env remove --name tensorflow

conda env export --name ENVNAME > envname.yml
conda env create --file envname.yml

Record model.compile options

Suppose we have a model of two hidden layers as follows:

model = tf.keras.models.Sequential([tf.keras.layers.Flatten(), 
                                    tf.keras.layers.Dense(128, activation=tf.nn.relu), 
                                    tf.keras.layers.Dense(10, activation=tf.nn.softmax)])

Then our model.compile might have the following as arguments:

model.compile(optimizer = tf.optimizers.Adam(),
              loss = 'sparse_categorical_crossentropy',
              metrics=['accuracy'])

With a callback that stop training at desired loss:

import tensorflow as tf
print(tf.__version__)

class myCallback(tf.keras.callbacks.Callback):
  def on_epoch_end(self, epoch, logs={}):
    if(logs.get('loss')<0.4):
      print("\nReached 60% accuracy so cancelling training!")
      self.model.stop_training = True

callbacks = myCallback()
mnist = tf.keras.datasets.fashion_mnist
(training_images, training_labels), (test_images, test_labels) = mnist.load_data()
training_images=training_images/255.0
test_images=test_images/255.0
model = tf.keras.models.Sequential([
  tf.keras.layers.Flatten(),
  tf.keras.layers.Dense(512, activation=tf.nn.relu),
  tf.keras.layers.Dense(10, activation=tf.nn.softmax)
])
model.compile(optimizer='adam', loss='sparse_categorical_crossentropy')
model.fit(training_images, training_labels, epochs=5, callbacks=[callbacks])

ResNet

def identity_block(X, f, filters, stage, block):
    """
    Implementation of the identity block as defined in Figure 3
    
    Arguments:
    X -- input tensor of shape (m, n_H_prev, n_W_prev, n_C_prev)
    f -- integer, specifying the shape of the middle CONV's window for the main path
    filters -- python list of integers, defining the number of filters in the CONV layers of the main path
    stage -- integer, used to name the layers, depending on their position in the network
    block -- string/character, used to name the layers, depending on their position in the network
    
    Returns:
    X -- output of the identity block, tensor of shape (n_H, n_W, n_C)
    """
    
    # defining name basis
    conv_name_base = 'res' + str(stage) + block + '_branch'    bn_name_base = 'bn' + str(stage) + block + '_branch'
    
    # Retrieve Filters
    F1, F2, F3 = filters
    
    # Save the input value. You'll need this later to add back to the main path. 
    X_shortcut = X
    
    # First component of main path
    X = Conv2D(filters = F1, kernel_size = (1, 1), strides = (1,1), padding = 'valid', name = conv_name_base + '2a', kernel_initializer = glorot_uniform(seed=0))(X)
    X = BatchNormalization(axis = 3, name = bn_name_base + '2a')(X)
    X = Activation('relu')(X)
    
    ### START CODE HERE ###
    
    # Second component of main path (≈3 lines)
    X = Conv2D(filters = F2, kernel_size = (f,f), strides = (1,1), padding = 'same', name = conv_name_base + '2b', kernel_initializer = glorot_uniform(seed=0))(X)
    X = BatchNormalization(axis=3, name=bn_name_base+'2b')(X)
    X = Activation('relu')(X)
    
    # Third component of main path (≈2 lines)
    X = Conv2D(filters = F3, kernel_size = (1,1), strides = (1,1), padding = 'valid', name = conv_name_base+'2c', kernel_initializer = glorot_uniform(seed=0))(X)
    X = BatchNormalization(axis=3, name=bn_name_base+'2c')(X)

    # Final step: Add shortcut value to main path, and pass it through a RELU activation (≈2 lines)
    X = Add()([X, X_shortcut])
    X = Activation('relu')(X)
    
    ### END CODE HERE ###

def convolutional_block(X, f, filters, stage, block, s = 2):
    """
    Implementation of the convolutional block as defined in Figure 4
    
    Arguments:
    X -- input tensor of shape (m, n_H_prev, n_W_prev, n_C_prev)
    f -- integer, specifying the shape of the middle CONV's window for the main path
    filters -- python list of integers, defining the number of filters in the CONV layers of the main path
    stage -- integer, used to name the layers, depending on their position in the network
    block -- string/character, used to name the layers, depending on their position in the network
    s -- Integer, specifying the stride to be used
    
    Returns:
    X -- output of the convolutional block, tensor of shape (n_H, n_W, n_C)
    """
    
    # defining name basis
    conv_name_base = 'res' + str(stage) + block + '_branch'
    bn_name_base = 'bn' + str(stage) + block + '_branch'
    
    # Retrieve Filters
    F1, F2, F3 = filters
    
    # Save the input value
    X_shortcut = X


    ##### MAIN PATH #####
    # First component of main path 
    X = Conv2D(F1, (1, 1), strides = (s,s), name = conv_name_base + '2a', kernel_initializer = glorot_uniform(seed=0))(X)
    X = BatchNormalization(axis = 3, name = bn_name_base + '2a')(X)
    X = Activation('relu')(X)
    
    ### START CODE HERE ###

    # Second component of main path (≈3 lines)
    X = Conv2D(F2, (f, f), strides = (1,1), padding="same", name = conv_name_base + '2b', kernel_initializer = glorot_uniform(seed=0))(X)
    X = BatchNormalization(axis = 3, name = bn_name_base + '2b')(X)
    X = Activation('relu')(X)

    # Third component of main path (≈2 lines)
    X = Conv2D(F3, (1, 1), strides = (1,1), padding="valid", name = conv_name_base + '2c', kernel_initializer = glorot_uniform(seed=0))(X)
    X = BatchNormalization(axis = 3, name = bn_name_base + '2c')(X)

    ##### SHORTCUT PATH #### (≈2 lines)
    X_shortcut = Conv2D(F3, (1, 1), strides = (s,s), padding="valid", name = conv_name_base + '1', kernel_initializer = glorot_uniform(seed=0))(X_shortcut)
    X_shortcut = BatchNormalization(axis = 3, name = bn_name_base + '1')(X_shortcut)
      
    # Final step: Add shortcut value to main path, and pass it through a RELU activation (≈2 lines)
    X = Add()([X,  X_shortcut])
    X = Activation("relu")(X)
    
    ### END CODE HERE ###
    
    return X

def ResNet50(input_shape = (64, 64, 3), classes = 6):
    """
    Implementation of the popular ResNet50 the following architecture:
    CONV2D -> BATCHNORM -> RELU -> MAXPOOL -> CONVBLOCK -> IDBLOCK*2 -> CONVBLOCK -> IDBLOCK*3
    -> CONVBLOCK -> IDBLOCK*5 -> CONVBLOCK -> IDBLOCK*2 -> AVGPOOL -> TOPLAYER

    Arguments:
    input_shape -- shape of the images of the dataset
    classes -- integer, number of classes

    Returns:
    model -- a Model() instance in Keras
    """
    
    # Define the input as a tensor with shape input_shape
    X_input = Input(input_shape)

    
    # Zero-Padding
    X = ZeroPadding2D((3, 3))(X_input)
    
    # Stage 1
    X = Conv2D(64, (7, 7), strides = (2, 2), name = 'conv1', kernel_initializer = glorot_uniform(seed=0))(X)
    X = BatchNormalization(axis = 3, name = 'bn_conv1')(X)
    X = Activation('relu')(X)
    X = MaxPooling2D((3, 3), strides=(2, 2))(X)

    # Stage 2
    X = convolutional_block(X, f = 3, filters = [64, 64, 256], stage = 2, block='a', s = 1)
    X = identity_block(X, 3, [64, 64, 256], stage=2, block='b')
    X = identity_block(X, 3, [64, 64, 256], stage=2, block='c')

    ### START CODE HERE ###

    # Stage 3 (≈4 lines)
    X = convolutional_block(X, f = 3, filters=[128, 128, 512], stage = 3, block="a", s = 2)
    X = identity_block(X, 3, filters=[128,128,512], stage=3, block="b")
    X = identity_block(X, 3, filters=[128,128,512], stage=3, block="c")
    X = identity_block(X, 3, filters=[128,128,512], stage=3, block="d")

    # Stage 4 (≈6 lines)
    X = convolutional_block(X, f = 3, filters=[256, 256, 1024], stage = 4, block="a", s = 2)
    X = identity_block(X, 3, filters=[256, 256, 1024], stage=4, block="b")
    X = identity_block(X, 3, filters=[256, 256, 1024], stage=4, block="c")
    X = identity_block(X, 3, filters=[256, 256, 1024], stage=4, block="d")
    X = identity_block(X, 3, filters=[256, 256, 1024], stage=4, block="e")
    X = identity_block(X, 3, filters=[256, 256, 1024], stage=4, block="f")

    # Stage 5 (≈3 lines)
    X = convolutional_block(X, f = 3, filters=[512, 512, 2048], stage = 5, block="a", s = 2)
    X = identity_block(X, f=3, filters=[512, 512, 2048], stage=5, block="b")
    X = identity_block(X, f=3, filters=[512, 512, 2048], stage=5, block="c")

    # AVGPOOL (≈1 line). Use "X = AveragePooling2D(...)(X)"
    X = AveragePooling2D(pool_size=(2, 2), name="avg_pool")(X)
    
    ### END CODE HERE ###

    # output layer
    X = Flatten()(X)
    X = Dense(classes, activation='softmax', name='fc' + str(classes), kernel_initializer = glorot_uniform(seed=0))(X)
    
    
    # Create model
    model = Model(inputs = X_input, outputs = X, name='ResNet50')

    return model

model.compile(optimizer='adam', loss='categorical_crossentropy', metrics=['accuracy'])

X_train_orig, Y_train_orig, X_test_orig, Y_test_orig, classes = load_dataset()

# Normalize image vectors
X_train = X_train_orig/255.
X_test = X_test_orig/255.

# Convert training and test labels to one hot matrices
Y_train = convert_to_one_hot(Y_train_orig, 6).T
Y_test = convert_to_one_hot(Y_test_orig, 6).T

model.fit(X_train, Y_train, epochs = 2, batch_size = 32)

preds = model.evaluate(X_test, Y_test)
print ("Loss = " + str(preds[0]))
print ("Test Accuracy = " + str(preds[1]))

Code Assignment

def HappyModel(input_shape):
    X_input = Input(input_shape)
    
    X = ZeroPadding2D((3,3))(X_input)
    X = Conv2D(18,(7,7),strides=(1,1),name="conv0")(X)
    X = BatchNormalization(axis=3, name="bn0")(X)
    X = Activation("relu")(X)
    
    X = MaxPooling2D((2,2), name="max_pool")(X)
    
    X = Flatten()(X)
    X = Dense(1,activation="sigmoid", name="fC")(X)
    
    model = Model(input = X_input, outputs = X, name="happy model")
	 return model

happyModel = HappyModel(X_train.shape[1:])
     
happyModel.compile(optimizer="adam", loss="binary_crossentropy", metrics=["accuracy"])
    
happyModel.fit(x=X_train, y=Y_train,epochs=10, batch_size=20)
    
preds = happyModel.evaluate(x=X_test,y=Y_test)

img_path = 'images/smile.jpg'
  
img = image.load_img(img_path, target_size=(64, 64))
imshow(img)

x = image.img_to_array(img)
x = np.expand_dims(x, axis=0)
x = preprocess_input(x)
print(happyModel.predict(x))

Derive the Formula of $\displaystyle \frac{\partial \mathcal L}{\partial W^{[\ell]}}$

Wikipedia record the following formula without proof:

I accidentally found that by the formulas in the previous post, we can already derive the following

Theorem. For every $\ell<L-1$, we let $\Phi^{[\ell]}:\R\to \R$ denote the activation function in the hidden layer, then we have\[ \frac{\partial \mathcal L }{\partial W^{[\ell]}}=\underbrace{\frac{1}{m}\Phi^{[\ell]}{}'(U^{[\ell]}) * \left[\prod_{i=\ell +1}^{L-1} (\Phi^{[i]}{}'(U^{[i]}) * W^{[i]T}\right]\cdot \frac{\partial \mathcal L}{\partial Y^{[L-1]}} }_{:=\delta_\ell} \cdot Y^{[\ell-1]T} = \delta_{\ell}\cdot Y^{[\ell-1]T}.\] Here $*$ denotes the entrywise multiplication. Since $\displaystyle \frac{\partial \mathcal L}{\partial W^{[L]}}=\frac{1}{m}\cdot \frac{\partial \mathcal L}{\partial U^{[L]}}\cdot Y^{[L-1]T}$, we also define \[ \boxed{\delta_L = \frac{1}{m}\cdot \frac{\partial \mathcal L}{\partial U^{[L]}}}    \]and since \[\frac{\partial \mathcal L}{\partial W^{[L-1]}} =\frac{1}{m}\Phi^{[L-1]}{}'(U^{[L-1]})* \left( W^{[L]T} \cdot \frac{\partial \mathcal L}{\partial U^{[L]}}\right) Y^{[L-2]T}=\delta_{L-1} Y^{[L-2]T} \] with $\delta_{L-1} :=\frac{1}{m}\cdot \left( \Phi^{[L-1]}{}'(U^{[L-1]}) * W^{[L]T}\right)\cdot \frac{\partial \mathcal L}{\partial U^{[L]}}$, by the definition of $\delta_\ell$ for $\ell<L-1$ above, we obtain for every $\ell\leq L-1$, \[\boxed{ \delta_{\ell} = \frac{1}{m}\cdot \Phi^{[\ell]}{}'(U^{[\ell]}) * \left[W^{[\ell+1]T} \cdot \delta_{\ell+1}\right]\quad \text{with}\quad \frac{\partial \mathcal L}{\partial W^{[\ell]}} = \delta_\ell Y^{[\ell-1]T}.} \] And as a side consequence of our computation, since $\displaystyle\frac{1}{m}\cdot \frac{\partial \mathcal L}{\partial U^{[\ell]}} = \delta_\ell$, \[ \boxed{\frac{\partial \mathcal L}{\partial b^{[\ell]}} = \text{np.sum}(\delta_\ell,\text{axis=1}).} \]

The last two formulars are computationally very useful. Note that in the definition of $\delta_\ell$, the multiplication in the product notation will not make sense unless they act on the rightmost matrix $\displaystyle \frac{\partial \mathcal L}{\partial Y^{[L-1]}} $ in a correct order (from the biggest index). To simplify notations we follow Andrew Ng's course to define $dW = \partial \mathcal L /\partial W$ and similarly for other matrices.

Proof. By repeated use of the formular $dY^{[\ell]} = [W^{[\ell+1]T}dY^{[\ell+1]}] * \Phi^{[\ell+1]}(U^{[\ell+1]})$ we have \[\begin{align*} dW^{[\ell]}& = \frac{1}{m} dU^{[\ell]} Y^{[\ell-1]T}\\ &=\frac{1}{m}\left(\left[dY^{[\ell]}\right] * \Phi^{[\ell]}{}'(U^{[\ell]})\right) Y^{[\ell-1]T}\\ &=\frac{1}{m}\left( \Phi^{[\ell]'}(U^{[\ell]})* \left[\prod_{i=\ell+1}^{L-1} \Phi^{[i]}{}'(U^{[i]}) * W^{[i]T}\right]\cdot dY^{[L-1]}\right) \cdot Y^{[\ell-1]T} \end{align*} \] And recall that $dY^{[L]} =\displaystyle \frac{\partial \mathcal L}{\partial Y^{[L]}}. \qed$

Formulas Revisit

When I study Andrew Ng course I am used to formulas like $dZ^{[\ell]}, dA$ , etc notations. Recently I revisit the topic and I am then used to using the notation \[u_i^{[\ell]} = W^{[\ell]}_{i:}\cdot y^{[\ell-1]} + b^{[\ell]}\quad\text{and}\quad y^{[\ell]} = \Phi^{[\ell]}(u),\] so I want to record the corresponding formulas for computation. Since the notation $dW$ doesn't look any cleaner than $\displaystyle\frac{\partial \mathcal L}{\partial W}$, in the sequel I write everything explicitly. \[ \boxed{\frac{\partial \mathcal L}{\partial W^{[\ell]}} = \frac{1}{m}\cdot \frac{\partial \mathcal L }{\partial U^{[\ell]}}\cdot Y^{[\ell-1]T}} \] \[ \boxed{\frac{\partial \mathcal L}{\partial U^{[\ell]}} = \brac{ W^{[\ell +1]T} \cdot \frac{\partial \mathcal L}{\partial U^{[\ell+1]}} }* \Phi^{[\ell]}{}'(U^{[\ell]})} \] where $*$ denotes entrywise product of matrices. \[ \boxed{\frac{\partial \mathcal L}{ \partial Y^{[\ell - 1]}} = W^{[\ell]T}\cdot \frac{\partial \mathcal L}{\partial U^{[\ell]}}} \] The last two yield the following for $\ell<L$ the hidden layer and for $\Phi^{[\ell]}:\R\to \R$ the activation function at $\ell$-th layer. \[ \boxed{\frac{\partial \mathcal L}{\partial U^{[\ell]}} = \frac{\partial \mathcal L}{\partial Y^{[\ell]}} * \Phi^{[\ell]}{}'(U^{[\ell]})} \] and finally \[ \boxed{\frac{\partial \mathcal L}{\partial b^{[\ell]} } =\frac{1}{m}\cdot \sum_{i=1}^m \frac{\partial \mathcal L}{\partial u^{[\ell](i)}} =\frac{1}{m}\cdot \text{np.sum}\brac{\frac{\partial \mathcal L}{\partial U^{[\ell]}},\text{axis = 1}}} \] For derivation of these formulas one can visit my another post: https://checkerlee.blogspot.com/2019/11/important-formulas-in-backward.html#more

Intutive derivation of Cross Entropy as a "loss" function

In defining "loss" function for classification problems given $p_i=\mathbb P\{\text{$i$ occurs}\}$, $i=1,2,\dots,n$,  from emperical data, we measure the accuracy of estimated data (from our output layer in neuron network) $[q_1,q_2,\dots,q_n]$ by the cross-entropy: \[L=\sum_{i=1}^n p_i\ln q_i.\] Recently I revisit this topic, and understand that this comes very naturally from solving maximum-likelihood estimation problem!

        Let's take an example, consider flipping a coin with getting a head with probability $p$ and tail with $1-p$, then the probability of getting 2 heads out of 6 flipping is \[ L(p) = \binom{6}{2} p^2 (1-p)^4 = 15 p^2(1-p)^4. \] Maximum-likelihood estimation ask the following problem:

The phenomenon of getting 2 heads is most likely to happen under what value of $p$?

In other words, the above question is the same as at what value of $p$ the proability $L(p)$ gets maximized? By simply solving $L'(p)=0$ we get the answer $p=\frac{1}{3}$.

        But in more complex problem we could not have the probability of some phenomenon to occurs based on another probablity with explicit formula. Instead of computing the probability $p$ directly, we try to estimate it such that our observation (the phenomenon from empirical data) is most likely to occur, and such an estimated value $p$ is considered as a good estimation.

        Now the derivation of cross-entropy will be very intuitive: Assume that \[ \text{mutally disjoint }E_i=\{\text{$i$ occurs}\},\quad \mathbb P(E_i) = p_i, \quad i=1,2,3,\dots,n. \] And assume further that $E_i$'s are iid events. Consider events $A_1,\dots,A_N$ are such that $A_i = \cupp_{i=1}^n E_i$ for each $i$ (for example, flipping coins $N$ times), then $p_i = N_i/N$, where $N_i$ is the number of times $i$ occures among $A_1,\dots,A_N$.

        Now we get another estimation $q_i$ of the same event $E_i$ from what ever experiment we can imagine. How good is $[q_1,\dots,q_n]$ as an estimation to the past emperical data $[p_1,\dots,p_n]$? The standard distance in $\R^n$ is certainly not a good choice since an quantity $\epsilon$ from $q_i$ to $p_i$ can mean huge difference from $q_{i'}$ to $p_{i'}$. $[q_1,\dots,q_n]$ is considered as good estimation if the observed phenomenon \[ \{\text{1 appears $N_1$ times}\}, \quad \{\text{2 appears $N_2$ times}\},\quad \dots ,\quad \{\text{n appears $N_n$ times} \} \] is very likely to happen under the estimates $[q_1,\dots,q_n]$, i.e., when \[ L = \prod_{i=1}^N q_i^{N_i}\iff \frac{\ln L}{N}= \sum_{i=1}^N \frac{N_i}{N}\ln q_i = \sum_{i=1}^n p_i\ln q_i. \] is large, and we have derived the cross-entropy at this point.

Useful command in git review course:

git rm --cached -r build/

If we type git rm -h we get an explanaation that:

--cached only remove from the index

by index it means tracked files inside the staging area (files are always tracked once they get committed.) A gitignore has no effect to files that is already tracked, so we use git rm --cached.

babel-node template

npm install babel-cli babel-preset-env --save-dev

{
  "presets": [
    "env"
  ]
}

 nodemon src/index.js --exec babel-node

Completely uninstall apache2 to get fresh config

apt-get remove --purge apache2 apache2-data apache2-utils

Powershell command to debug ios in chrome

remotedebug_ios_webkit_adapter --port=9000

Open safari, browse to the page that is going to be inspected. Then in chrome go to chrome://inspect/#devices and choose the device.

Copy all lastly updated files into a single directory if the version control is horribly not done by git:

Instantiate a git repository (don't need to connect to any remote branch, do everything locally). Create a .sh file at the current directory, and paste:

#!/bin/bash

git add .
git status

updatedFiles=$(git status | awk '{print $2}' | grep -P "\..*$")

touch updates/update.txt
git status > updates/update.txt

for file in $updatedFiles
do
    cp --parents "$file" ./updates
    echo "copied $file to ./updates/$file"
done;

read -p "Press enter to exit"

mkdir updates and then run the bash script above. Files will be copied into updates directory, and we can manage it by date.

如果用了 spring-boot-devtools 的話，謹記在開啟 SpringApplication 前:

public class Main {
  public static void main(String[] args) {
    System.setProperty("spring.devtools.restart.enabled", "false");
    SpringApplication.run(Main.class, args);
  }
}

不然不知為甚麼它有 restart 機制，restart 前後的同一個 class 將不視為同一個 class，database transaction 將發生錯誤。

Hilbernate Database Configuration without XML

package com.springboot.mvc;

import java.util.Properties;
import com.springboot.mvc.models.Customer;
import org.hibernate.SessionFactory;
import org.hibernate.boot.registry.StandardServiceRegistryBuilder;
import org.hibernate.cfg.Configuration;
import org.hibernate.cfg.Environment;
import org.hibernate.service.ServiceRegistry;

public class HibernateUtil {
  private static SessionFactory sessionFactory;

  public static SessionFactory getSessionFactory() {
    if (sessionFactory == null) {
      try {
        Configuration configuration = new Configuration();
        // Hibernate settings equivalent to hibernate.cfg.xml's properties
        Properties settings = new Properties();
        settings.put(Environment.DRIVER, "com.mysql.cj.jdbc.Driver");
        settings.put(Environment.URL,
            "jdbc:mysql://192.168.99.100:3306/JDBC_spring_mvc_tutorial?useSSL=false&serverTimezone=UTC");
        settings.put(Environment.USER, "cclee");
        settings.put(Environment.PASS, "ccleedb12345");
        settings.put(Environment.DIALECT, "org.hibernate.dialect.MySQL55Dialect");
        settings.put(Environment.SHOW_SQL, "true");
        settings.put(Environment.CURRENT_SESSION_CONTEXT_CLASS, "thread");
        settings.put(Environment.HBM2DDL_AUTO, "create-drop");

        configuration.setProperties(settings);
        configuration.addAnnotatedClass(Customer.class);
       // we add more and more classes here.
       
       
        ServiceRegistry serviceRegistry = new StandardServiceRegistryBuilder()
            .applySettings(configuration.getProperties()).build();
        sessionFactory = configuration.buildSessionFactory(serviceRegistry);
      } catch (Exception e) {
        e.printStackTrace();
      }
    }
    // Then:
    // Session session = sessionFactory.openSession();
    // Transaction transaction = session.beginTransaction();

    return sessionFactory;
  }
}

basic setup for hibernate

For hiberante.cfg.xml:

    
        update
        
        com.mysql.cj.jdbc.Driver
        jdbc:mysql://192.168.99.100:3306/JDBC_test?useSSL=false
        root
        cclee12345@12345

        
        1

        
        org.hibernate.dialect.MySQL55Dialect

        
        true

  
  thread
        create-drop

        
    

and pom.xml we need:

  mysql
  mysql-connector-java
  8.0.20

in additional to the hibernate maven dependencies.

Record for my Docker Files

FROM node:10

WORKDIR /usr/src/app

COPY package*.json ./

RUN npm install && npm rebuild bcrypt --build-from-source

EXPOSE 3000

CMD ["npm", "start"]

and

version: "3.7"

services:
  db:
    container_name: postgres_screencapdic_db
    image: postgres
    restart: always
    environment:
      POSTGRES_USER: cclee11111
      POSTGRES_PASSWORD: cclee11111
      POSTGRES_DB: screencapdb
    volumes:
      - screencapdb:/var/lib/postgresql/data
    ports:
      - "5432:5432"
  # screencap_api:
  #   build:
  #     context: ./
  #     dockerfile: Dockerfile-screepcap-express
  #   container_name: screencap_api
  #   restart: always
  #   ports:
  #     - "8080:3000"
  #   volumes:
  #     - type: bind
  #       source: ./
  #       target: /usr/src/app
  #     - /usr/src/app/node_modules

volumes:
  screencapdb:

State Pattern

Trying to refactor my old project by using design pattern. It really becomes simple to implement new functionality now! What I need to do is just to create additional state. And my code get less coupled! My class mainVM know nothing about my translation service now!

https://github.com/machingclee/ScreenCapDictionaryNoteApp_refactor/tree/2020-06-23-refactor-translation-by-state-pattern/ScreenCapDictionaryNoteApp/ViewModel/Helpers/TranslationHelper

Bash Script

 for f in *\ *; do mv "$f" "${f// /_}"; done

This change all spaces in file name by a "_".

Use Sequelize Migration with ES6 Syntax

After yarn add sequelize-cli it is clear from the --help command how to generate a migration folder and migration file. The only trouble is to use them with ES6 syntax. From the official document:

https://sequelize.org/master/manual/migrations.html#using-babel

we add

yarn add babel-register

and add a .sequelizerc runtime config with

// .sequelizerc
require("babel-register");

const path = require('path');

module.exports = {
  'config': path.resolve('config', 'config.json'),
  'models-path': path.resolve('models'),
  'seeders-path': path.resolve('seeders'),
  'migrations-path': path.resolve('migrations')
}

We can copy the implementation of altering, creating, deleting table from official documentation:

https://sequelize.org/master/manual/query-interface.html

Official document also says that in migration file we can export async function up and async function down instead of returning a chain of promises (i.e., a promise)! For example it happens that I want to add a column for users to implement mobile push notification, then I need to add a column called push_notification_token, I can do the following in our migration file:

"use strict";
import { modelNames } from "../src/enums/modelNames";
import { Sequelize, DataTypes } from "sequelize";

module.exports = {
  async up(queryInterface, Sequelize) {
    await queryInterface.addColumn(modelNames.USER + "s", "push_notification_token", {
      type: DataTypes.STRING,
      allowNull: true
    });
  },

  async down(queryInterface, Sequelize) {
    await queryInterface.removeColumn(
      modelNames.USER + "s",
      "push_notification_token",
      {}
    );
  }
};

Now if you run the code, we encounter the following error

Loaded configuration file "config\config.js".
Using environment "development".
== 20200615141047-add-push-notification-token-to-users-table: migrating =======

ERROR: regeneratorRuntime is not defined

so we need the transform runtime plugin by babel,

yarn add babel-plugin-transform-runtime

and in our .babelrc add:

{
  "presets": ["env"],
  "plugins": [
    ["transform-runtime", {
      "regenerator": true
    }]
  ]
}

and we are done!

SQL injection 及 prepared statement already exists

雖然知道 sql injection 成功後果很嚴重，但因為我很少打 raw sql，而我知道很多 library (sequelize, knex 等) 都會避免 sql injection 所以沒有特別去深究。

今天突然想起這個問題，而我也在為自己的手機 app 寫一個 backend 及想有一個自己 customize 的 query 結果。翻查 sequelize 的 doc 這件事沒有比寫 raw query 簡單，所以就開始自己寫 raw sequel。嘗試 sql inject 自己一下。發現如果沒有做任何預防操施真的很危險，甚至把我整個 database 毀掉:

所以開始學習寫 prepare statement:

PREPARE get_notes (int) AS
  SELECT v."id", v."word", v."pronounciation", v."explanation", p."dateTime", p."croppedScreenshot"
  FROM vocabs v
  INNER JOIN pages p 
  ON v."sqlitePageId"=p."sqliteId"
  WHERE p."sqliteNoteId"=$1;
EXECUTE get_notes(${sqliteNoteId});

在 postman get request 了一次，一切都很美好，再 get request 多一次，誒?

error: prepared statement "get_notes" already exists

搜尋了一下解決方法，最後只要每一次完成 EXECUTE 後把儲存好的 prepared statement 移除就好，整句變成:

PREPARE get_notes (int) AS
  SELECT v."id", v."word", v."pronounciation", v."explanation", p."dateTime", p."croppedScreenshot"
  FROM vocabs v
  INNER JOIN pages p 
  ON v."sqlitePageId"=p."sqliteId"
  WHERE p."sqliteNoteId"=$1;
EXECUTE get_notes(${sqliteNoteId});
DEALLOCATE get_notes;

Sequelize API for CRUD operation

更多: https://dwatow.github.io/2018/09-24-sequelize/sequelize-R-of-CRUD/

module.exports = (app, db) => {
  app.get( "/posts", (req, res) =>
    db.post.findAll().then( (result) => res.json(result) )
  );

  app.get( "/post/:id", (req, res) =>
    db.post.findByPk(req.params.id).then( (result) => res.json(result))
  );

  app.post("/post", (req, res) => 
    db.post.create({
      title: req.body.title,
      content: req.body.content
    }).then( (result) => res.json(result) )
  );

  app.put( "/post/:id", (req, res) =>
    db.post.update({
      title: req.body.title,
      content: req.body.content
    },
    {
      where: {
        id: req.params.id
      }
    }).then( (result) => res.json(result) )
  );

  app.delete( "/post/:id", (req, res) =>
    db.post.destroy({
      where: {
        id: req.params.id
      }
    }).then( (result) => res.json(result) )
  );
}

Use docker-compose up instead of docker container run -v blablabla for local developement

version: "3.7"

services:
  app:
    container_name: docker-node-mongo
    restart: always
    build: .
    ports:
      - "80:3000"
    volumes:
      - type: bind
        source: ./
        target: /usr/src/app

  mongo:
    container_name: mongo
    image: mongo
    ports:
      - "27017:27017"

Redux Setup that also applies to React-Native

import { createStore, combineReducers } from "redux";
import trackFormReducer from "./reducers/trackFormReducer";
const rootReducer = combineReducers({
  trackFormReducer: trackFormReducer
});

const store = createStore(
  rootReducer /* preloadedState, */,
  window.__REDUX_DEVTOOLS_EXTENSION__ && window.__REDUX_DEVTOOLS_EXTENSION__()
);

export default store;

and

import { Provider as StoreProvider } from "react-redux";

and wrap the app component by StoreProvider

Make sure swarm manager role not to run any container

docker node update --availability drain <manager host name>

Docker Exercise: docker-compose

The following create drupal service with persistent named datas: (docker-compose.yml)

version: "2"

services:
  drupal:
    image: drupal
    ports:
      - "8080:80"
    volumes:
      - drupal-modules:/var/www/html/modules
      - drupal-profiles:/var/www/html/profiles
      - drupal-sites:/var/www/html/sites
      - drupal-themes:/var/www/html/themes
  postgres:
    image: postgres
    environment:
      - POSTGRES_PASSWORD=1234

volumes:
  drupal-modules:
  drupal-profiles:
  drupal-sites:
  drupal-themes:

The options can be found in the official docker page from hub.docker.com. cd into the directory that contains the above docker-compose.yml and run docker-compose up.

When we are done, we use docker-compose down -v to remove everything.

Note that the service name is implicitly also the hostname of the service. For instance, when we try to connect to postgres database inside the network from one of the container (for example, our drupal service), the hostname is can be put as postgres.

Another example of docker-compose:

version: "2"

services:
  proxy:
    build:
      context: .
      dockerfile: nginx.Dockerfile
    image: nginx-custom
    ports:
      - "80:80"
  web:
    image: httpd
    volumes:
      - ./html:/usr/local/apache2/htdocs/

If image is not found, it will run the build command and tag it with the image name. There are at least two ways to customize what to build.

1. use COPY in Dockerfile
2. use volumes in docker-compose.yml.

Docker Exercise

FROM node:14.2.0-alpine3.10

EXPOSE 3000

RUN apk add --update tini

RUN mkdir -p /usr/src/app

WORKDIR /usr/src/app

COPY package.json package.json

RUN npm install

COPY . .

CMD ["tini", "--", "node", "./bin/www"]

cd into the directory that contains the above as Dockerfile, and run

docker build -t testnode

The above will be built as an image and tagged with testnode. After we make sure testnode can be run properly (by running docker container run --rm -p 80:3000 testnode), we change the tag name as follows:

docker tag testnode machingclee/testnode

and then

docker push machingclee/testnode

if we want.

Standard docker command

The following create a new container that redirect traffic to our host 192.168.99.100 at port 3306 into port 3306 of the running container.

docker container run -d -p 3306:3306 --name db -e MYSQL_RANDOM_ROOT_PASSWORD=yes mysql

To test linux, we can run another container that run an image of minimal version of ubuntu as follow:

docker container run -it --name ubuntu ubuntu 

Here -it (i.e., -i and -t) allows us to kind of SSH into the container and get the shell ready. By typing

exit

inside the shell we return to our docker prompt. To go back to our shell inside ubuntu container, we run

docker container start -ai ubuntu

Bash Script to Batch Renaming

The following replaces "abcd_mixdown.mp3" to "abcd_vo.mp3" for every mp3 file of the current directory.

for mp3 in *.mp3; do renamedmp3=$(echo "$mp3" | sed 's/_mixdown.mp3$/.mp3/'); mv ./"$mp3" ./"$renamedmp3; done"

The next one goes even further, it does the same thing with all current and subdirectory:

#!/bin/bash

list=$(find -name '*.mp3' | grep '_mixdown\.mp3');

for mp3FilePath in $list
do
    newMp3FilePath=$(echo "$mp3FilePath" | sed 's/_mixdown\.mp3/_vo\.mp3/')
    mv $mp3FilePath $newMp3FilePath
done

webpack config

https://github.com/Microsoft/TypeScript-Babel-Starter

const path = require("path");

module.exports = {
  entry: "./src/index.js",
  output: {
    path: __dirname,
    filename: "app.js"
  },
  module: {
    rules: [
      {
        test: /\.js$/,
        exclude: /node_modules/,
        use: {
          loader: "babel-loader",
          options: {
            presets: ["@babel/preset-env"],
            plugins: [
              [
                "@babel/plugin-transform-runtime",
                {
                  regenerator: true
                }
              ]
            ]
          }
        }
      }
    ]
  },
  target: "node",
  node: {
    __dirname: false,
    __filename: false
  },
  externals: {
    fs: "commonjs fs"
  },
  mode: "development"
};

Data Validation using Decorator in Typescript

In index.html we have

<form>
  <input type="text" placeholder="Course Title" id="title" />
  <input type="text" placeholder="Price" id="price" />
  <button type="submit">Submit</button>
</form>

and inside our ts (transpiled into js file and link it into our index.html), we write:

enum Validation {
  required = "required",
  positive = "positive"
}

interface ValidatorConfig {
  [property: string]: {
    [validatableProp: string]: Validation[]; //e.g., [Validation.required, Validation.positive]
  };
}

const registeredValidators: ValidatorConfig = {};

function Required(target: any, propName: string) {
  //we don't have propertyDescriptor for property, it exists only for methods
  registeredValidators[target.constructor.name] = {
    ...registeredValidators[target.constructor.name],
    [propName]: [Validation.required]
  };
}

function Positive(target: any, propName: string) {
  registeredValidators[target.constructor.name] = {
    ...registeredValidators[target.constructor.name],
    [propName]: [Validation.positive]
  };
}

function validate(obj: any) {
  const objValidatorConfig = registeredValidators[obj.constructor.name];
  if (!objValidatorConfig) {
    return true;
  } else {
    let validated = true;
    for (const prop in objValidatorConfig) {
      for (const validator of objValidatorConfig[prop]) {
        if (validator === (Validation.required as string)) {
          console.log("run?");
          validated = validated && obj[prop].trim().length > 0;
        }
        if (validator === (Validation.positive as string)) {
          validated = validated && obj[prop] > 0;
        }
      }
    }
    return validated;
  }
}

class Course {
  @Required
  public title: string;

  @Positive
  public price: number;

  constructor(t: string, p: number) {
    this.title = t;
    this.price = p;
  }
}

const courseForm = document.querySelector("form")!;
courseForm.addEventListener("submit", e => {
  e.preventDefault();
  const titleEl = document.getElementById("title") as HTMLInputElement;
  const priceEl = document.getElementById("price") as HTMLInputElement;
  const title = titleEl.value;
  const price = +priceEl.value;
  const newCourse = new Course(title, price);

  if (!validate(newCourse)) {
    alert("Invalid input, please try again!");
    return;
  }

  console.log(newCourse);
});

Wordpress Study Notes

Extension in VSCode

• I install beautify for cleaning up the indentations of both php and html code at the same time.
• I also install PHP Intelephense to give auto completion on html tag.

Prelude

Hierarchy of wordpress php files: Assume that we have a post type called "program", as registered by using

 
<?php 
function university_post_types()
{
  //Program Post Type
  register_post_type("program", array(
  'supports' => array('title', 'editor'),
  'rewrite'=>array('slug'=>'programs'),
  'has_archive' => true,
  'public' => true,
  'labels' => array(
  'name' => 'Programs',
  'add_new_item'=>'Add New Program',
  'edit_item'=>'Edit Program',
  'all_items'=> 'All Programs',
  'singular_name' => 'Program'
  ),
  'menu_icon' => 'dashicons-awards'
  ));
}

add_action("init", "university_post_types");
?>

inside \wp-content\mu-plugins\*.php  (mu stands for "must-use") , then we get:

in our dashboard. This is a kind of customized post type, therefore we will create a php file that is dedicated to customizing post of type "program".

php installation

1. Follow the link to install php
   https://www.youtube.com/watch?v=4_-12QSaaFg
2. Following this to install xampp:
   https://www.youtube.com/watch?v=TjFRTkw6GDQ
3. visual studio plug-in that I have used:
   1. phpfmt - PHP formatter
   2. Format HTML in PHP
   refer to my config file sent in gmail

Factory Design Pattern in Typescript

Javascript is not enough to implement ordinary factory design patternt that I want to mimic from C#, therefore I try to do it on Typescript, though nested class declaration is still not that straight forward in typescript:

class Point {
  x: number;
  y: number;

  private constructor(x: number, y: number) {
    this.x = x;
    this.y = y;
  }

  static Factory = class {
    static pointXY(x: number, y: number) {
      return new Point(x, y);
    }
    static pointPolar(r: number, theta: number) {
      return new Point(
        r * Math.cos((theta * Math.PI) / 180),
        r * Math.sin((theta * Math.PI) / 180)
      );
    }
  };
}

now we can create our point by calling

const pt = Point.Factory.pointXY(10, 20);

Certbot

SSH into our ubuntu device and do the following. The commands are subject to changes, we should use all latest relevant commands from certbot's website.

$ sudo apt update
$ clear
$ sudo apt install apache2
$ cd /etc/apache2/sites-available/
$ clear
$ ls
$ sudo vi ridiculous-inc.com.conf
$ cd /var/www
$ sudo git clone https://github.com/ridiculous-ijquery-todo.git ridic
$ sudo a2ensite ridiculous-inc.com.conf 
$ sudo service apache2 restart
$ sudo apt-get update
$ clear
$ sudo apt-get install software-properties-common
$ sudo add-apt-repository ppa:certbot/certbot
$ clear
$ sudo apt-get update
$ sudo apt-get install python-certbot-apache 
$ clear
$ sudo certbot --apache
$ history

and

<VirtualHost *:80>
    DocumentRoot /var/www/ridic
    ServerName ridiculous-inc.com
    <Directory "/var/www/ridic">
        allow from all
        AllowOverride All
        Order allow,deny
        Options +Indexes
    </Directory>
</VirtualHost>

To redirect request to port 80 to other internal port we use the following setting of virtualhost:

<VirtualHost *:80>
    ServerName api.screencapdictionary.com
    <Location "/">
        ProxyPreserveHost On
        ProxyPass http://localhost:5000/
        ProxyPassReverse http://localhost:5000/
    </Location>
</VirtualHost>

Git command that I should know

Configuration

git config --global --edit

This will open the .gitconfig file by our default editor. Make changes, save, and close the editor, changes will take place. To view the changes:

git config --global --list

Branches

List out all branches:

git branch -a

Create new branch:

git branch mynewbranch

switch to new branch:

git checkout mynewbranch

Rename a branch (-m for move, as in mv in bash command):

 git branch -m mynewbranch newbranch

Delete a branch:

 git branch -d newbranch

Specific branch

git clone a specific branch instead of all the branches and checkout specific one. For example, in my private repo I want to clone the "forth-branch" only, then write:

git clone --single-branch --branch forth-branch https://github.com/machingclee/2020-English-Learning-Website.git

without --single-branch the above will fetch all branches and checkout the forth-branch.

P4Merge Configuration

git config --global merge.tool p4merge
git config --global mergetool.p4merge.path "C:/Program Files/Perforce/p4merge.exe"
git config --global mergetool.prompt false
git config --global diff.tool p4merge
git config --global difftool.p4merge.path "C:/Program Files/Perforce/p4merge.exe"
git config --global difftool.prompt false

and git config --global --list to double check the configuration:

core.editor="C:\Users\Ching-Cheong Lee\AppData\Local\Programs\Microsoft VS Code\Code.exe" --wait
user.name=James Lee
user.email=machingclee@gmail.com
color.ui=true
merge.tool=p4merge
mergetool.p4merge.path=C:/Program Files/Perforce/p4merge.exe
mergetool.prompt=false
diff.tool=p4merge
difftool.p4merge.path=C:/Program Files/Perforce/p4merge.exe
difftool.prompt=false

moving WPF from .net framework to .net core.

It is just a simple function that takes users to a route by using his/her default browser. The migration is fine except ......

the adjustment: