diff --git a/machine-learning-ex7.zip b/machine-learning-ex7.zip new file mode 100644 index 0000000..83463a5 Binary files /dev/null and b/machine-learning-ex7.zip differ diff --git a/machine-learning-ex7/ex7.pdf b/machine-learning-ex7/ex7.pdf new file mode 100644 index 0000000..366201c Binary files /dev/null and b/machine-learning-ex7/ex7.pdf differ diff --git a/machine-learning-ex7/ex7/bird_small.mat b/machine-learning-ex7/ex7/bird_small.mat new file mode 100644 index 0000000..04c224c Binary files /dev/null and b/machine-learning-ex7/ex7/bird_small.mat differ diff --git a/machine-learning-ex7/ex7/bird_small.png b/machine-learning-ex7/ex7/bird_small.png new file mode 100644 index 0000000..a3cd00c Binary files /dev/null and b/machine-learning-ex7/ex7/bird_small.png differ diff --git a/machine-learning-ex7/ex7/computeCentroids.m b/machine-learning-ex7/ex7/computeCentroids.m new file mode 100644 index 0000000..606011e --- /dev/null +++ b/machine-learning-ex7/ex7/computeCentroids.m @@ -0,0 +1,40 @@ +function centroids = computeCentroids(X, idx, K) +%COMPUTECENTROIDS returs the new centroids by computing the means of the +%data points assigned to each centroid. +% centroids = COMPUTECENTROIDS(X, idx, K) returns the new centroids by +% computing the means of the data points assigned to each centroid. It is +% given a dataset X where each row is a single data point, a vector +% idx of centroid assignments (i.e. each entry in range [1..K]) for each +% example, and K, the number of centroids. You should return a matrix +% centroids, where each row of centroids is the mean of the data points +% assigned to it. +% + +% Useful variables +[m n] = size(X); + +% You need to return the following variables correctly. +centroids = zeros(K, n); + + +% ====================== YOUR CODE HERE ====================== +% Instructions: Go over every centroid and compute mean of all points that +% belong to it. Concretely, the row vector centroids(i, :) +% should contain the mean of the data points assigned to +% centroid i. +% +% Note: You can use a for-loop over the centroids to compute this. +% + + + + + + + + +% ============================================================= + + +end + diff --git a/machine-learning-ex7/ex7/displayData.m b/machine-learning-ex7/ex7/displayData.m new file mode 100644 index 0000000..160697e --- /dev/null +++ b/machine-learning-ex7/ex7/displayData.m @@ -0,0 +1,59 @@ +function [h, display_array] = displayData(X, example_width) +%DISPLAYDATA Display 2D data in a nice grid +% [h, display_array] = DISPLAYDATA(X, example_width) displays 2D data +% stored in X in a nice grid. It returns the figure handle h and the +% displayed array if requested. + +% Set example_width automatically if not passed in +if ~exist('example_width', 'var') || isempty(example_width) + example_width = round(sqrt(size(X, 2))); +end + +% Gray Image +colormap(gray); + +% Compute rows, cols +[m n] = size(X); +example_height = (n / example_width); + +% Compute number of items to display +display_rows = floor(sqrt(m)); +display_cols = ceil(m / display_rows); + +% Between images padding +pad = 1; + +% Setup blank display +display_array = - ones(pad + display_rows * (example_height + pad), ... + pad + display_cols * (example_width + pad)); + +% Copy each example into a patch on the display array +curr_ex = 1; +for j = 1:display_rows + for i = 1:display_cols + if curr_ex > m, + break; + end + % Copy the patch + + % Get the max value of the patch + max_val = max(abs(X(curr_ex, :))); + display_array(pad + (j - 1) * (example_height + pad) + (1:example_height), ... + pad + (i - 1) * (example_width + pad) + (1:example_width)) = ... + reshape(X(curr_ex, :), example_height, example_width) / max_val; + curr_ex = curr_ex + 1; + end + if curr_ex > m, + break; + end +end + +% Display Image +h = imagesc(display_array, [-1 1]); + +% Do not show axis +axis image off + +drawnow; + +end diff --git a/machine-learning-ex7/ex7/drawLine.m b/machine-learning-ex7/ex7/drawLine.m new file mode 100644 index 0000000..85e6c41 --- /dev/null +++ b/machine-learning-ex7/ex7/drawLine.m @@ -0,0 +1,8 @@ +function drawLine(p1, p2, varargin) +%DRAWLINE Draws a line from point p1 to point p2 +% DRAWLINE(p1, p2) Draws a line from point p1 to point p2 and holds the +% current figure + +plot([p1(1) p2(1)], [p1(2) p2(2)], varargin{:}); + +end \ No newline at end of file diff --git a/machine-learning-ex7/ex7/ex7.m b/machine-learning-ex7/ex7/ex7.m new file mode 100644 index 0000000..3a095ae --- /dev/null +++ b/machine-learning-ex7/ex7/ex7.m @@ -0,0 +1,174 @@ +%% Machine Learning Online Class +% Exercise 7 | Principle Component Analysis and K-Means Clustering +% +% Instructions +% ------------ +% +% This file contains code that helps you get started on the +% exercise. You will need to complete the following functions: +% +% pca.m +% projectData.m +% recoverData.m +% computeCentroids.m +% findClosestCentroids.m +% kMeansInitCentroids.m +% +% For this exercise, you will not need to change any code in this file, +% or any other files other than those mentioned above. +% + +%% Initialization +clear ; close all; clc + +%% ================= Part 1: Find Closest Centroids ==================== +% To help you implement K-Means, we have divided the learning algorithm +% into two functions -- findClosestCentroids and computeCentroids. In this +% part, you shoudl complete the code in the findClosestCentroids function. +% +fprintf('Finding closest centroids.\n\n'); + +% Load an example dataset that we will be using +load('ex7data2.mat'); + +% Select an initial set of centroids +K = 3; % 3 Centroids +initial_centroids = [3 3; 6 2; 8 5]; + +% Find the closest centroids for the examples using the +% initial_centroids +idx = findClosestCentroids(X, initial_centroids); + +fprintf('Closest centroids for the first 3 examples: \n') +fprintf(' %d', idx(1:3)); +fprintf('\n(the closest centroids should be 1, 3, 2 respectively)\n'); + +fprintf('Program paused. Press enter to continue.\n'); +pause; + +%% ===================== Part 2: Compute Means ========================= +% After implementing the closest centroids function, you should now +% complete the computeCentroids function. +% +fprintf('\nComputing centroids means.\n\n'); + +% Compute means based on the closest centroids found in the previous part. +centroids = computeCentroids(X, idx, K); + +fprintf('Centroids computed after initial finding of closest centroids: \n') +fprintf(' %f %f \n' , centroids'); +fprintf('\n(the centroids should be\n'); +fprintf(' [ 2.428301 3.157924 ]\n'); +fprintf(' [ 5.813503 2.633656 ]\n'); +fprintf(' [ 7.119387 3.616684 ]\n\n'); + +fprintf('Program paused. Press enter to continue.\n'); +pause; + + +%% =================== Part 3: K-Means Clustering ====================== +% After you have completed the two functions computeCentroids and +% findClosestCentroids, you have all the necessary pieces to run the +% kMeans algorithm. In this part, you will run the K-Means algorithm on +% the example dataset we have provided. +% +fprintf('\nRunning K-Means clustering on example dataset.\n\n'); + +% Load an example dataset +load('ex7data2.mat'); + +% Settings for running K-Means +K = 3; +max_iters = 10; + +% For consistency, here we set centroids to specific values +% but in practice you want to generate them automatically, such as by +% settings them to be random examples (as can be seen in +% kMeansInitCentroids). +initial_centroids = [3 3; 6 2; 8 5]; + +% Run K-Means algorithm. The 'true' at the end tells our function to plot +% the progress of K-Means +[centroids, idx] = runkMeans(X, initial_centroids, max_iters, true); +fprintf('\nK-Means Done.\n\n'); + +fprintf('Program paused. Press enter to continue.\n'); +pause; + +%% ============= Part 4: K-Means Clustering on Pixels =============== +% In this exercise, you will use K-Means to compress an image. To do this, +% you will first run K-Means on the colors of the pixels in the image and +% then you will map each pixel on to it's closest centroid. +% +% You should now complete the code in kMeansInitCentroids.m +% + +fprintf('\nRunning K-Means clustering on pixels from an image.\n\n'); + +% Load an image of a bird +A = double(imread('bird_small.png')); + +% If imread does not work for you, you can try instead +% load ('bird_small.mat'); + +A = A / 255; % Divide by 255 so that all values are in the range 0 - 1 + +% Size of the image +img_size = size(A); + +% Reshape the image into an Nx3 matrix where N = number of pixels. +% Each row will contain the Red, Green and Blue pixel values +% This gives us our dataset matrix X that we will use K-Means on. +X = reshape(A, img_size(1) * img_size(2), 3); + +% Run your K-Means algorithm on this data +% You should try different values of K and max_iters here +K = 16; +max_iters = 10; + +% When using K-Means, it is important the initialize the centroids +% randomly. +% You should complete the code in kMeansInitCentroids.m before proceeding +initial_centroids = kMeansInitCentroids(X, K); + +% Run K-Means +[centroids, idx] = runkMeans(X, initial_centroids, max_iters); + +fprintf('Program paused. Press enter to continue.\n'); +pause; + + +%% ================= Part 5: Image Compression ====================== +% In this part of the exercise, you will use the clusters of K-Means to +% compress an image. To do this, we first find the closest clusters for +% each example. After that, we + +fprintf('\nApplying K-Means to compress an image.\n\n'); + +% Find closest cluster members +idx = findClosestCentroids(X, centroids); + +% Essentially, now we have represented the image X as in terms of the +% indices in idx. + +% We can now recover the image from the indices (idx) by mapping each pixel +% (specified by it's index in idx) to the centroid value +X_recovered = centroids(idx,:); + +% Reshape the recovered image into proper dimensions +X_recovered = reshape(X_recovered, img_size(1), img_size(2), 3); + +% Display the original image +subplot(1, 2, 1); +imagesc(A); +title('Original'); + +% Display compressed image side by side +subplot(1, 2, 2); +imagesc(X_recovered) +title(sprintf('Compressed, with %d colors.', K)); + + +fprintf('Program paused. Press enter to continue.\n'); +pause; + diff --git a/machine-learning-ex7/ex7/ex7_pca.m b/machine-learning-ex7/ex7/ex7_pca.m new file mode 100644 index 0000000..de98b13 --- /dev/null +++ b/machine-learning-ex7/ex7/ex7_pca.m @@ -0,0 +1,235 @@ +%% Machine Learning Online Class +% Exercise 7 | Principle Component Analysis and K-Means Clustering +% +% Instructions +% ------------ +% +% This file contains code that helps you get started on the +% exercise. You will need to complete the following functions: +% +% pca.m +% projectData.m +% recoverData.m +% computeCentroids.m +% findClosestCentroids.m +% kMeansInitCentroids.m +% +% For this exercise, you will not need to change any code in this file, +% or any other files other than those mentioned above. +% + +%% Initialization +clear ; close all; clc + +%% ================== Part 1: Load Example Dataset =================== +% We start this exercise by using a small dataset that is easily to +% visualize +% +fprintf('Visualizing example dataset for PCA.\n\n'); + +% The following command loads the dataset. You should now have the +% variable X in your environment +load ('ex7data1.mat'); + +% Visualize the example dataset +plot(X(:, 1), X(:, 2), 'bo'); +axis([0.5 6.5 2 8]); axis square; + +fprintf('Program paused. Press enter to continue.\n'); +pause; + + +%% =============== Part 2: Principal Component Analysis =============== +% You should now implement PCA, a dimension reduction technique. You +% should complete the code in pca.m +% +fprintf('\nRunning PCA on example dataset.\n\n'); + +% Before running PCA, it is important to first normalize X +[X_norm, mu, sigma] = featureNormalize(X); + +% Run PCA +[U, S] = pca(X_norm); + +% Compute mu, the mean of the each feature + +% Draw the eigenvectors centered at mean of data. These lines show the +% directions of maximum variations in the dataset. +hold on; +drawLine(mu, mu + 1.5 * S(1,1) * U(:,1)', '-k', 'LineWidth', 2); +drawLine(mu, mu + 1.5 * S(2,2) * U(:,2)', '-k', 'LineWidth', 2); +hold off; + +fprintf('Top eigenvector: \n'); +fprintf(' U(:,1) = %f %f \n', U(1,1), U(2,1)); +fprintf('\n(you should expect to see -0.707107 -0.707107)\n'); + +fprintf('Program paused. Press enter to continue.\n'); +pause; + + +%% =================== Part 3: Dimension Reduction =================== +% You should now implement the projection step to map the data onto the +% first k eigenvectors. The code will then plot the data in this reduced +% dimensional space. This will show you what the data looks like when +% using only the corresponding eigenvectors to reconstruct it. +% +% You should complete the code in projectData.m +% +fprintf('\nDimension reduction on example dataset.\n\n'); + +% Plot the normalized dataset (returned from pca) +plot(X_norm(:, 1), X_norm(:, 2), 'bo'); +axis([-4 3 -4 3]); axis square + +% Project the data onto K = 1 dimension +K = 1; +Z = projectData(X_norm, U, K); +fprintf('Projection of the first example: %f\n', Z(1)); +fprintf('\n(this value should be about 1.481274)\n\n'); + +X_rec = recoverData(Z, U, K); +fprintf('Approximation of the first example: %f %f\n', X_rec(1, 1), X_rec(1, 2)); +fprintf('\n(this value should be about -1.047419 -1.047419)\n\n'); + +% Draw lines connecting the projected points to the original points +hold on; +plot(X_rec(:, 1), X_rec(:, 2), 'ro'); +for i = 1:size(X_norm, 1) + drawLine(X_norm(i,:), X_rec(i,:), '--k', 'LineWidth', 1); +end +hold off + +fprintf('Program paused. Press enter to continue.\n'); +pause; + +%% =============== Part 4: Loading and Visualizing Face Data ============= +% We start the exercise by first loading and visualizing the dataset. +% The following code will load the dataset into your environment +% +fprintf('\nLoading face dataset.\n\n'); + +% Load Face dataset +load ('ex7faces.mat') + +% Display the first 100 faces in the dataset +displayData(X(1:100, :)); + +fprintf('Program paused. Press enter to continue.\n'); +pause; + +%% =========== Part 5: PCA on Face Data: Eigenfaces =================== +% Run PCA and visualize the eigenvectors which are in this case eigenfaces +% We display the first 36 eigenfaces. +% +fprintf(['\nRunning PCA on face dataset.\n' ... + '(this mght take a minute or two ...)\n\n']); + +% Before running PCA, it is important to first normalize X by subtracting +% the mean value from each feature +[X_norm, mu, sigma] = featureNormalize(X); + +% Run PCA +[U, S] = pca(X_norm); + +% Visualize the top 36 eigenvectors found +displayData(U(:, 1:36)'); + +fprintf('Program paused. Press enter to continue.\n'); +pause; + + +%% ============= Part 6: Dimension Reduction for Faces ================= +% Project images to the eigen space using the top k eigenvectors +% If you are applying a machine learning algorithm +fprintf('\nDimension reduction for face dataset.\n\n'); + +K = 100; +Z = projectData(X_norm, U, K); + +fprintf('The projected data Z has a size of: ') +fprintf('%d ', size(Z)); + +fprintf('\n\nProgram paused. Press enter to continue.\n'); +pause; + +%% ==== Part 7: Visualization of Faces after PCA Dimension Reduction ==== +% Project images to the eigen space using the top K eigen vectors and +% visualize only using those K dimensions +% Compare to the original input, which is also displayed + +fprintf('\nVisualizing the projected (reduced dimension) faces.\n\n'); + +K = 100; +X_rec = recoverData(Z, U, K); + +% Display normalized data +subplot(1, 2, 1); +displayData(X_norm(1:100,:)); +title('Original faces'); +axis square; + +% Display reconstructed data from only k eigenfaces +subplot(1, 2, 2); +displayData(X_rec(1:100,:)); +title('Recovered faces'); +axis square; + +fprintf('Program paused. Press enter to continue.\n'); +pause; + + +%% === Part 8(a): Optional (ungraded) Exercise: PCA for Visualization === +% One useful application of PCA is to use it to visualize high-dimensional +% data. In the last K-Means exercise you ran K-Means on 3-dimensional +% pixel colors of an image. We first visualize this output in 3D, and then +% apply PCA to obtain a visualization in 2D. + +close all; close all; clc + +% Re-load the image from the previous exercise and run K-Means on it +% For this to work, you need to complete the K-Means assignment first +A = double(imread('bird_small.png')); + +% If imread does not work for you, you can try instead +% load ('bird_small.mat'); + +A = A / 255; +img_size = size(A); +X = reshape(A, img_size(1) * img_size(2), 3); +K = 16; +max_iters = 10; +initial_centroids = kMeansInitCentroids(X, K); +[centroids, idx] = runkMeans(X, initial_centroids, max_iters); + +% Sample 1000 random indexes (since working with all the data is +% too expensive. If you have a fast computer, you may increase this. +sel = floor(rand(1000, 1) * size(X, 1)) + 1; + +% Setup Color Palette +palette = hsv(K); +colors = palette(idx(sel), :); + +% Visualize the data and centroid memberships in 3D +figure; +scatter3(X(sel, 1), X(sel, 2), X(sel, 3), 10, colors); +title('Pixel dataset plotted in 3D. Color shows centroid memberships'); +fprintf('Program paused. Press enter to continue.\n'); +pause; + +%% === Part 8(b): Optional (ungraded) Exercise: PCA for Visualization === +% Use PCA to project this cloud to 2D for visualization + +% Subtract the mean to use PCA +[X_norm, mu, sigma] = featureNormalize(X); + +% PCA and project the data to 2D +[U, S] = pca(X_norm); +Z = projectData(X_norm, U, 2); + +% Plot in 2D +figure; +plotDataPoints(Z(sel, :), idx(sel), K); +title('Pixel dataset plotted in 2D, using PCA for dimensionality reduction'); +fprintf('Program paused. Press enter to continue.\n'); +pause; diff --git a/machine-learning-ex7/ex7/ex7data1.mat b/machine-learning-ex7/ex7/ex7data1.mat new file mode 100644 index 0000000..f9c3961 Binary files /dev/null and b/machine-learning-ex7/ex7/ex7data1.mat differ diff --git a/machine-learning-ex7/ex7/ex7data2.mat b/machine-learning-ex7/ex7/ex7data2.mat new file mode 100644 index 0000000..de3f5b9 Binary files /dev/null and b/machine-learning-ex7/ex7/ex7data2.mat differ diff --git a/machine-learning-ex7/ex7/ex7faces.mat b/machine-learning-ex7/ex7/ex7faces.mat new file mode 100644 index 0000000..3965bd1 Binary files /dev/null and b/machine-learning-ex7/ex7/ex7faces.mat differ diff --git a/machine-learning-ex7/ex7/featureNormalize.m b/machine-learning-ex7/ex7/featureNormalize.m new file mode 100644 index 0000000..da03bee --- /dev/null +++ b/machine-learning-ex7/ex7/featureNormalize.m @@ -0,0 +1,17 @@ +function [X_norm, mu, sigma] = featureNormalize(X) +%FEATURENORMALIZE Normalizes the features in X +% FEATURENORMALIZE(X) returns a normalized version of X where +% the mean value of each feature is 0 and the standard deviation +% is 1. This is often a good preprocessing step to do when +% working with learning algorithms. + +mu = mean(X); +X_norm = bsxfun(@minus, X, mu); + +sigma = std(X_norm); +X_norm = bsxfun(@rdivide, X_norm, sigma); + + +% ============================================================ + +end diff --git a/machine-learning-ex7/ex7/findClosestCentroids.m b/machine-learning-ex7/ex7/findClosestCentroids.m new file mode 100644 index 0000000..52f6d8e --- /dev/null +++ b/machine-learning-ex7/ex7/findClosestCentroids.m @@ -0,0 +1,33 @@ +function idx = findClosestCentroids(X, centroids) +%FINDCLOSESTCENTROIDS computes the centroid memberships for every example +% idx = FINDCLOSESTCENTROIDS (X, centroids) returns the closest centroids +% in idx for a dataset X where each row is a single example. idx = m x 1 +% vector of centroid assignments (i.e. each entry in range [1..K]) +% + +% Set K +K = size(centroids, 1); + +% You need to return the following variables correctly. +idx = zeros(size(X,1), 1); + +% ====================== YOUR CODE HERE ====================== +% Instructions: Go over every example, find its closest centroid, and store +% the index inside idx at the appropriate location. +% Concretely, idx(i) should contain the index of the centroid +% closest to example i. Hence, it should be a value in the +% range 1..K +% +% Note: You can use a for-loop over the examples to compute this. +% + + + + + + + +% ============================================================= + +end + diff --git a/machine-learning-ex7/ex7/kMeansInitCentroids.m b/machine-learning-ex7/ex7/kMeansInitCentroids.m new file mode 100644 index 0000000..7a6d252 --- /dev/null +++ b/machine-learning-ex7/ex7/kMeansInitCentroids.m @@ -0,0 +1,26 @@ +function centroids = kMeansInitCentroids(X, K) +%KMEANSINITCENTROIDS This function initializes K centroids that are to be +%used in K-Means on the dataset X +% centroids = KMEANSINITCENTROIDS(X, K) returns K initial centroids to be +% used with the K-Means on the dataset X +% + +% You should return this values correctly +centroids = zeros(K, size(X, 2)); + +% ====================== YOUR CODE HERE ====================== +% Instructions: You should set centroids to randomly chosen examples from +% the dataset X +% + + + + + + + + +% ============================================================= + +end + diff --git a/machine-learning-ex7/ex7/lib/jsonlab/AUTHORS.txt b/machine-learning-ex7/ex7/lib/jsonlab/AUTHORS.txt new file mode 100644 index 0000000..9dd3fc7 --- /dev/null +++ b/machine-learning-ex7/ex7/lib/jsonlab/AUTHORS.txt @@ -0,0 +1,41 @@ +The author of "jsonlab" toolbox is Qianqian Fang. Qianqian +is currently an Assistant Professor at Massachusetts General Hospital, +Harvard Medical School. + +Address: Martinos Center for Biomedical Imaging, + Massachusetts General Hospital, + Harvard Medical School + Bldg 149, 13th St, Charlestown, MA 02129, USA +URL: http://nmr.mgh.harvard.edu/~fangq/ +Email: or + + +The script loadjson.m was built upon previous works by + +- Nedialko Krouchev: http://www.mathworks.com/matlabcentral/fileexchange/25713 + date: 2009/11/02 +- François Glineur: http://www.mathworks.com/matlabcentral/fileexchange/23393 + date: 2009/03/22 +- Joel Feenstra: http://www.mathworks.com/matlabcentral/fileexchange/20565 + date: 2008/07/03 + + +This toolbox contains patches submitted by the following contributors: + +- Blake Johnson + part of revision 341 + +- Niclas Borlin + various fixes in revision 394, including + - loadjson crashes for all-zero sparse matrix. + - loadjson crashes for empty sparse matrix. + - Non-zero size of 0-by-N and N-by-0 empty matrices is lost after savejson/loadjson. + - loadjson crashes for sparse real column vector. + - loadjson crashes for sparse complex column vector. + - Data is corrupted by savejson for sparse real row vector. + - savejson crashes for sparse complex row vector. + +- Yul Kang + patches for svn revision 415. + - savejson saves an empty cell array as [] instead of null + - loadjson differentiates an empty struct from an empty array diff --git a/machine-learning-ex7/ex7/lib/jsonlab/ChangeLog.txt b/machine-learning-ex7/ex7/lib/jsonlab/ChangeLog.txt new file mode 100644 index 0000000..07824f5 --- /dev/null +++ b/machine-learning-ex7/ex7/lib/jsonlab/ChangeLog.txt @@ -0,0 +1,74 @@ +============================================================================ + + JSONlab - a toolbox to encode/decode JSON/UBJSON files in MATLAB/Octave + +---------------------------------------------------------------------------- + +JSONlab ChangeLog (key features marked by *): + +== JSONlab 1.0 (codename: Optimus - Final), FangQ == + + 2015/01/02 polish help info for all major functions, update examples, finalize 1.0 + 2014/12/19 fix a bug to strictly respect NoRowBracket in savejson + +== JSONlab 1.0.0-RC2 (codename: Optimus - RC2), FangQ == + + 2014/11/22 show progress bar in loadjson ('ShowProgress') + 2014/11/17 add Compact option in savejson to output compact JSON format ('Compact') + 2014/11/17 add FastArrayParser in loadjson to specify fast parser applicable levels + 2014/09/18 start official github mirror: https://github.com/fangq/jsonlab + +== JSONlab 1.0.0-RC1 (codename: Optimus - RC1), FangQ == + + 2014/09/17 fix several compatibility issues when running on octave versions 3.2-3.8 + 2014/09/17 support 2D cell and struct arrays in both savejson and saveubjson + 2014/08/04 escape special characters in a JSON string + 2014/02/16 fix a bug when saving ubjson files + +== JSONlab 0.9.9 (codename: Optimus - beta), FangQ == + + 2014/01/22 use binary read and write in saveubjson and loadubjson + +== JSONlab 0.9.8-1 (codename: Optimus - alpha update 1), FangQ == + + 2013/10/07 better round-trip conservation for empty arrays and structs (patch submitted by Yul Kang) + +== JSONlab 0.9.8 (codename: Optimus - alpha), FangQ == + 2013/08/23 *universal Binary JSON (UBJSON) support, including both saveubjson and loadubjson + +== JSONlab 0.9.1 (codename: Rodimus, update 1), FangQ == + 2012/12/18 *handling of various empty and sparse matrices (fixes submitted by Niclas Borlin) + +== JSONlab 0.9.0 (codename: Rodimus), FangQ == + + 2012/06/17 *new format for an invalid leading char, unpacking hex code in savejson + 2012/06/01 support JSONP in savejson + 2012/05/25 fix the empty cell bug (reported by Cyril Davin) + 2012/04/05 savejson can save to a file (suggested by Patrick Rapin) + +== JSONlab 0.8.1 (codename: Sentiel, Update 1), FangQ == + + 2012/02/28 loadjson quotation mark escape bug, see http://bit.ly/yyk1nS + 2012/01/25 patch to handle root-less objects, contributed by Blake Johnson + +== JSONlab 0.8.0 (codename: Sentiel), FangQ == + + 2012/01/13 *speed up loadjson by 20 fold when parsing large data arrays in matlab + 2012/01/11 remove row bracket if an array has 1 element, suggested by Mykel Kochenderfer + 2011/12/22 *accept sequence of 'param',value input in savejson and loadjson + 2011/11/18 fix struct array bug reported by Mykel Kochenderfer + +== JSONlab 0.5.1 (codename: Nexus Update 1), FangQ == + + 2011/10/21 fix a bug in loadjson, previous code does not use any of the acceleration + 2011/10/20 loadjson supports JSON collections - concatenated JSON objects + +== JSONlab 0.5.0 (codename: Nexus), FangQ == + + 2011/10/16 package and release jsonlab 0.5.0 + 2011/10/15 *add json demo and regression test, support cpx numbers, fix double quote bug + 2011/10/11 *speed up readjson dramatically, interpret _Array* tags, show data in root level + 2011/10/10 create jsonlab project, start jsonlab website, add online documentation + 2011/10/07 *speed up savejson by 25x using sprintf instead of mat2str, add options support + 2011/10/06 *savejson works for structs, cells and arrays + 2011/09/09 derive loadjson from JSON parser from MATLAB Central, draft savejson.m diff --git a/machine-learning-ex7/ex7/lib/jsonlab/LICENSE_BSD.txt b/machine-learning-ex7/ex7/lib/jsonlab/LICENSE_BSD.txt new file mode 100644 index 0000000..32d66cb --- /dev/null +++ b/machine-learning-ex7/ex7/lib/jsonlab/LICENSE_BSD.txt @@ -0,0 +1,25 @@ +Copyright 2011-2015 Qianqian Fang . All rights reserved. + +Redistribution and use in source and binary forms, with or without modification, are +permitted provided that the following conditions are met: + + 1. Redistributions of source code must retain the above copyright notice, this list of + conditions and the following disclaimer. + + 2. Redistributions in binary form must reproduce the above copyright notice, this list + of conditions and the following disclaimer in the documentation and/or other materials + provided with the distribution. + +THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS ''AS IS'' AND ANY EXPRESS OR IMPLIED +WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND +FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDERS +OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR +CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR +SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON +ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING +NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF +ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. + +The views and conclusions contained in the software and documentation are those of the +authors and should not be interpreted as representing official policies, either expressed +or implied, of the copyright holders. diff --git a/machine-learning-ex7/ex7/lib/jsonlab/README.txt b/machine-learning-ex7/ex7/lib/jsonlab/README.txt new file mode 100644 index 0000000..7b4f732 --- /dev/null +++ b/machine-learning-ex7/ex7/lib/jsonlab/README.txt @@ -0,0 +1,394 @@ +=============================================================================== += JSONLab = += An open-source MATLAB/Octave JSON encoder and decoder = +=============================================================================== + +*Copyright (C) 2011-2015 Qianqian Fang +*License: BSD License, see License_BSD.txt for details +*Version: 1.0 (Optimus - Final) + +------------------------------------------------------------------------------- + +Table of Content: + +I. Introduction +II. Installation +III.Using JSONLab +IV. Known Issues and TODOs +V. Contribution and feedback + +------------------------------------------------------------------------------- + +I. Introduction + +JSON ([http://www.json.org/ JavaScript Object Notation]) is a highly portable, +human-readable and "[http://en.wikipedia.org/wiki/JSON fat-free]" text format +to represent complex and hierarchical data. It is as powerful as +[http://en.wikipedia.org/wiki/XML XML], but less verbose. JSON format is widely +used for data-exchange in applications, and is essential for the wild success +of [http://en.wikipedia.org/wiki/Ajax_(programming) Ajax] and +[http://en.wikipedia.org/wiki/Web_2.0 Web2.0]. + +UBJSON (Universal Binary JSON) is a binary JSON format, specifically +optimized for compact file size and better performance while keeping +the semantics as simple as the text-based JSON format. Using the UBJSON +format allows to wrap complex binary data in a flexible and extensible +structure, making it possible to process complex and large dataset +without accuracy loss due to text conversions. + +We envision that both JSON and its binary version will serve as part of +the mainstream data-exchange formats for scientific research in the future. +It will provide the flexibility and generality achieved by other popular +general-purpose file specifications, such as +[http://www.hdfgroup.org/HDF5/whatishdf5.html HDF5], with significantly +reduced complexity and enhanced performance. + +JSONLab is a free and open-source implementation of a JSON/UBJSON encoder +and a decoder in the native MATLAB language. It can be used to convert a MATLAB +data structure (array, struct, cell, struct array and cell array) into +JSON/UBJSON formatted strings, or to decode a JSON/UBJSON file into MATLAB +data structure. JSONLab supports both MATLAB and +[http://www.gnu.org/software/octave/ GNU Octave] (a free MATLAB clone). + +------------------------------------------------------------------------------- + +II. Installation + +The installation of JSONLab is no different than any other simple +MATLAB toolbox. You only need to download/unzip the JSONLab package +to a folder, and add the folder's path to MATLAB/Octave's path list +by using the following command: + + addpath('/path/to/jsonlab'); + +If you want to add this path permanently, you need to type "pathtool", +browse to the jsonlab root folder and add to the list, then click "Save". +Then, run "rehash" in MATLAB, and type "which loadjson", if you see an +output, that means JSONLab is installed for MATLAB/Octave. + +------------------------------------------------------------------------------- + +III.Using JSONLab + +JSONLab provides two functions, loadjson.m -- a MATLAB->JSON decoder, +and savejson.m -- a MATLAB->JSON encoder, for the text-based JSON, and +two equivallent functions -- loadubjson and saveubjson for the binary +JSON. The detailed help info for the four functions can be found below: + +=== loadjson.m === +
+  data=loadjson(fname,opt)
+     or
+  data=loadjson(fname,'param1',value1,'param2',value2,...)
+ 
+  parse a JSON (JavaScript Object Notation) file or string
+ 
+  authors:Qianqian Fang (fangq nmr.mgh.harvard.edu)
+  created on 2011/09/09, including previous works from 
+ 
+          Nedialko Krouchev: http://www.mathworks.com/matlabcentral/fileexchange/25713
+             created on 2009/11/02
+          François Glineur: http://www.mathworks.com/matlabcentral/fileexchange/23393
+             created on  2009/03/22
+          Joel Feenstra:
+          http://www.mathworks.com/matlabcentral/fileexchange/20565
+             created on 2008/07/03
+ 
+  $Id: loadjson.m 452 2014-11-22 16:43:33Z fangq $
+ 
+  input:
+       fname: input file name, if fname contains "{}" or "[]", fname
+              will be interpreted as a JSON string
+       opt: a struct to store parsing options, opt can be replaced by 
+            a list of ('param',value) pairs - the param string is equivallent
+            to a field in opt. opt can have the following 
+            fields (first in [.|.] is the default)
+ 
+            opt.SimplifyCell [0|1]: if set to 1, loadjson will call cell2mat
+                          for each element of the JSON data, and group 
+                          arrays based on the cell2mat rules.
+            opt.FastArrayParser [1|0 or integer]: if set to 1, use a
+                          speed-optimized array parser when loading an 
+                          array object. The fast array parser may 
+                          collapse block arrays into a single large
+                          array similar to rules defined in cell2mat; 0 to 
+                          use a legacy parser; if set to a larger-than-1
+                          value, this option will specify the minimum
+                          dimension to enable the fast array parser. For
+                          example, if the input is a 3D array, setting
+                          FastArrayParser to 1 will return a 3D array;
+                          setting to 2 will return a cell array of 2D
+                          arrays; setting to 3 will return to a 2D cell
+                          array of 1D vectors; setting to 4 will return a
+                          3D cell array.
+            opt.ShowProgress [0|1]: if set to 1, loadjson displays a progress bar.
+ 
+  output:
+       dat: a cell array, where {...} blocks are converted into cell arrays,
+            and [...] are converted to arrays
+ 
+  examples:
+       dat=loadjson('{"obj":{"string":"value","array":[1,2,3]}}')
+       dat=loadjson(['examples' filesep 'example1.json'])
+       dat=loadjson(['examples' filesep 'example1.json'],'SimplifyCell',1)
+
+ +=== savejson.m === + +
+  json=savejson(rootname,obj,filename)
+     or
+  json=savejson(rootname,obj,opt)
+  json=savejson(rootname,obj,'param1',value1,'param2',value2,...)
+ 
+  convert a MATLAB object (cell, struct or array) into a JSON (JavaScript
+  Object Notation) string
+ 
+  author: Qianqian Fang (fangq nmr.mgh.harvard.edu)
+  created on 2011/09/09
+ 
+  $Id: savejson.m 458 2014-12-19 22:17:17Z fangq $
+ 
+  input:
+       rootname: the name of the root-object, when set to '', the root name
+         is ignored, however, when opt.ForceRootName is set to 1 (see below),
+         the MATLAB variable name will be used as the root name.
+       obj: a MATLAB object (array, cell, cell array, struct, struct array).
+       filename: a string for the file name to save the output JSON data.
+       opt: a struct for additional options, ignore to use default values.
+         opt can have the following fields (first in [.|.] is the default)
+ 
+         opt.FileName [''|string]: a file name to save the output JSON data
+         opt.FloatFormat ['%.10g'|string]: format to show each numeric element
+                          of a 1D/2D array;
+         opt.ArrayIndent [1|0]: if 1, output explicit data array with
+                          precedent indentation; if 0, no indentation
+         opt.ArrayToStruct[0|1]: when set to 0, savejson outputs 1D/2D
+                          array in JSON array format; if sets to 1, an
+                          array will be shown as a struct with fields
+                          "_ArrayType_", "_ArraySize_" and "_ArrayData_"; for
+                          sparse arrays, the non-zero elements will be
+                          saved to _ArrayData_ field in triplet-format i.e.
+                          (ix,iy,val) and "_ArrayIsSparse_" will be added
+                          with a value of 1; for a complex array, the 
+                          _ArrayData_ array will include two columns 
+                          (4 for sparse) to record the real and imaginary 
+                          parts, and also "_ArrayIsComplex_":1 is added. 
+         opt.ParseLogical [0|1]: if this is set to 1, logical array elem
+                          will use true/false rather than 1/0.
+         opt.NoRowBracket [1|0]: if this is set to 1, arrays with a single
+                          numerical element will be shown without a square
+                          bracket, unless it is the root object; if 0, square
+                          brackets are forced for any numerical arrays.
+         opt.ForceRootName [0|1]: when set to 1 and rootname is empty, savejson
+                          will use the name of the passed obj variable as the 
+                          root object name; if obj is an expression and 
+                          does not have a name, 'root' will be used; if this 
+                          is set to 0 and rootname is empty, the root level 
+                          will be merged down to the lower level.
+         opt.Inf ['"$1_Inf_"'|string]: a customized regular expression pattern
+                          to represent +/-Inf. The matched pattern is '([-+]*)Inf'
+                          and $1 represents the sign. For those who want to use
+                          1e999 to represent Inf, they can set opt.Inf to '$11e999'
+         opt.NaN ['"_NaN_"'|string]: a customized regular expression pattern
+                          to represent NaN
+         opt.JSONP [''|string]: to generate a JSONP output (JSON with padding),
+                          for example, if opt.JSONP='foo', the JSON data is
+                          wrapped inside a function call as 'foo(...);'
+         opt.UnpackHex [1|0]: conver the 0x[hex code] output by loadjson 
+                          back to the string form
+         opt.SaveBinary [0|1]: 1 - save the JSON file in binary mode; 0 - text mode.
+         opt.Compact [0|1]: 1- out compact JSON format (remove all newlines and tabs)
+ 
+         opt can be replaced by a list of ('param',value) pairs. The param 
+         string is equivallent to a field in opt and is case sensitive.
+  output:
+       json: a string in the JSON format (see http://json.org)
+ 
+  examples:
+       jsonmesh=struct('MeshNode',[0 0 0;1 0 0;0 1 0;1 1 0;0 0 1;1 0 1;0 1 1;1 1 1],... 
+                'MeshTetra',[1 2 4 8;1 3 4 8;1 2 6 8;1 5 6 8;1 5 7 8;1 3 7 8],...
+                'MeshTri',[1 2 4;1 2 6;1 3 4;1 3 7;1 5 6;1 5 7;...
+                           2 8 4;2 8 6;3 8 4;3 8 7;5 8 6;5 8 7],...
+                'MeshCreator','FangQ','MeshTitle','T6 Cube',...
+                'SpecialData',[nan, inf, -inf]);
+       savejson('jmesh',jsonmesh)
+       savejson('',jsonmesh,'ArrayIndent',0,'FloatFormat','\t%.5g')
+
+ +=== loadubjson.m === + +
+  data=loadubjson(fname,opt)
+     or
+  data=loadubjson(fname,'param1',value1,'param2',value2,...)
+ 
+  parse a JSON (JavaScript Object Notation) file or string
+ 
+  authors:Qianqian Fang (fangq nmr.mgh.harvard.edu)
+  created on 2013/08/01
+ 
+  $Id: loadubjson.m 436 2014-08-05 20:51:40Z fangq $
+ 
+  input:
+       fname: input file name, if fname contains "{}" or "[]", fname
+              will be interpreted as a UBJSON string
+       opt: a struct to store parsing options, opt can be replaced by 
+            a list of ('param',value) pairs - the param string is equivallent
+            to a field in opt. opt can have the following 
+            fields (first in [.|.] is the default)
+ 
+            opt.SimplifyCell [0|1]: if set to 1, loadubjson will call cell2mat
+                          for each element of the JSON data, and group 
+                          arrays based on the cell2mat rules.
+            opt.IntEndian [B|L]: specify the endianness of the integer fields
+                          in the UBJSON input data. B - Big-Endian format for 
+                          integers (as required in the UBJSON specification); 
+                          L - input integer fields are in Little-Endian order.
+ 
+  output:
+       dat: a cell array, where {...} blocks are converted into cell arrays,
+            and [...] are converted to arrays
+ 
+  examples:
+       obj=struct('string','value','array',[1 2 3]);
+       ubjdata=saveubjson('obj',obj);
+       dat=loadubjson(ubjdata)
+       dat=loadubjson(['examples' filesep 'example1.ubj'])
+       dat=loadubjson(['examples' filesep 'example1.ubj'],'SimplifyCell',1)
+
+ +=== saveubjson.m === + +
+  json=saveubjson(rootname,obj,filename)
+     or
+  json=saveubjson(rootname,obj,opt)
+  json=saveubjson(rootname,obj,'param1',value1,'param2',value2,...)
+ 
+  convert a MATLAB object (cell, struct or array) into a Universal 
+  Binary JSON (UBJSON) binary string
+ 
+  author: Qianqian Fang (fangq nmr.mgh.harvard.edu)
+  created on 2013/08/17
+ 
+  $Id: saveubjson.m 440 2014-09-17 19:59:45Z fangq $
+ 
+  input:
+       rootname: the name of the root-object, when set to '', the root name
+         is ignored, however, when opt.ForceRootName is set to 1 (see below),
+         the MATLAB variable name will be used as the root name.
+       obj: a MATLAB object (array, cell, cell array, struct, struct array)
+       filename: a string for the file name to save the output UBJSON data
+       opt: a struct for additional options, ignore to use default values.
+         opt can have the following fields (first in [.|.] is the default)
+ 
+         opt.FileName [''|string]: a file name to save the output JSON data
+         opt.ArrayToStruct[0|1]: when set to 0, saveubjson outputs 1D/2D
+                          array in JSON array format; if sets to 1, an
+                          array will be shown as a struct with fields
+                          "_ArrayType_", "_ArraySize_" and "_ArrayData_"; for
+                          sparse arrays, the non-zero elements will be
+                          saved to _ArrayData_ field in triplet-format i.e.
+                          (ix,iy,val) and "_ArrayIsSparse_" will be added
+                          with a value of 1; for a complex array, the 
+                          _ArrayData_ array will include two columns 
+                          (4 for sparse) to record the real and imaginary 
+                          parts, and also "_ArrayIsComplex_":1 is added. 
+         opt.ParseLogical [1|0]: if this is set to 1, logical array elem
+                          will use true/false rather than 1/0.
+         opt.NoRowBracket [1|0]: if this is set to 1, arrays with a single
+                          numerical element will be shown without a square
+                          bracket, unless it is the root object; if 0, square
+                          brackets are forced for any numerical arrays.
+         opt.ForceRootName [0|1]: when set to 1 and rootname is empty, saveubjson
+                          will use the name of the passed obj variable as the 
+                          root object name; if obj is an expression and 
+                          does not have a name, 'root' will be used; if this 
+                          is set to 0 and rootname is empty, the root level 
+                          will be merged down to the lower level.
+         opt.JSONP [''|string]: to generate a JSONP output (JSON with padding),
+                          for example, if opt.JSON='foo', the JSON data is
+                          wrapped inside a function call as 'foo(...);'
+         opt.UnpackHex [1|0]: conver the 0x[hex code] output by loadjson 
+                          back to the string form
+ 
+         opt can be replaced by a list of ('param',value) pairs. The param 
+         string is equivallent to a field in opt and is case sensitive.
+  output:
+       json: a binary string in the UBJSON format (see http://ubjson.org)
+ 
+  examples:
+       jsonmesh=struct('MeshNode',[0 0 0;1 0 0;0 1 0;1 1 0;0 0 1;1 0 1;0 1 1;1 1 1],... 
+                'MeshTetra',[1 2 4 8;1 3 4 8;1 2 6 8;1 5 6 8;1 5 7 8;1 3 7 8],...
+                'MeshTri',[1 2 4;1 2 6;1 3 4;1 3 7;1 5 6;1 5 7;...
+                           2 8 4;2 8 6;3 8 4;3 8 7;5 8 6;5 8 7],...
+                'MeshCreator','FangQ','MeshTitle','T6 Cube',...
+                'SpecialData',[nan, inf, -inf]);
+       saveubjson('jsonmesh',jsonmesh)
+       saveubjson('jsonmesh',jsonmesh,'meshdata.ubj')
+
+ + +=== examples === + +Under the "examples" folder, you can find several scripts to demonstrate the +basic utilities of JSONLab. Running the "demo_jsonlab_basic.m" script, you +will see the conversions from MATLAB data structure to JSON text and backward. +In "jsonlab_selftest.m", we load complex JSON files downloaded from the Internet +and validate the loadjson/savejson functions for regression testing purposes. +Similarly, a "demo_ubjson_basic.m" script is provided to test the saveubjson +and loadubjson pairs for various matlab data structures. + +Please run these examples and understand how JSONLab works before you use +it to process your data. + +------------------------------------------------------------------------------- + +IV. Known Issues and TODOs + +JSONLab has several known limitations. We are striving to make it more general +and robust. Hopefully in a few future releases, the limitations become less. + +Here are the known issues: + +# 3D or higher dimensional cell/struct-arrays will be converted to 2D arrays; +# When processing names containing multi-byte characters, Octave and MATLAB \ +can give different field-names; you can use feature('DefaultCharacterSet','latin1') \ +in MATLAB to get consistant results +# savejson can not handle class and dataset. +# saveubjson converts a logical array into a uint8 ([U]) array +# an unofficial N-D array count syntax is implemented in saveubjson. We are \ +actively communicating with the UBJSON spec maintainer to investigate the \ +possibility of making it upstream +# loadubjson can not parse all UBJSON Specification (Draft 9) compliant \ +files, however, it can parse all UBJSON files produced by saveubjson. + +------------------------------------------------------------------------------- + +V. Contribution and feedback + +JSONLab is an open-source project. This means you can not only use it and modify +it as you wish, but also you can contribute your changes back to JSONLab so +that everyone else can enjoy the improvement. For anyone who want to contribute, +please download JSONLab source code from it's subversion repository by using the +following command: + + svn checkout svn://svn.code.sf.net/p/iso2mesh/code/trunk/jsonlab jsonlab + +You can make changes to the files as needed. Once you are satisfied with your +changes, and ready to share it with others, please cd the root directory of +JSONLab, and type + + svn diff > yourname_featurename.patch + +You then email the .patch file to JSONLab's maintainer, Qianqian Fang, at +the email address shown in the beginning of this file. Qianqian will review +the changes and commit it to the subversion if they are satisfactory. + +We appreciate any suggestions and feedbacks from you. Please use iso2mesh's +mailing list to report any questions you may have with JSONLab: + +http://groups.google.com/group/iso2mesh-users?hl=en&pli=1 + +(Subscription to the mailing list is needed in order to post messages). diff --git a/machine-learning-ex7/ex7/lib/jsonlab/jsonopt.m b/machine-learning-ex7/ex7/lib/jsonlab/jsonopt.m new file mode 100644 index 0000000..0bebd8d --- /dev/null +++ b/machine-learning-ex7/ex7/lib/jsonlab/jsonopt.m @@ -0,0 +1,32 @@ +function val=jsonopt(key,default,varargin) +% +% val=jsonopt(key,default,optstruct) +% +% setting options based on a struct. The struct can be produced +% by varargin2struct from a list of 'param','value' pairs +% +% authors:Qianqian Fang (fangq nmr.mgh.harvard.edu) +% +% $Id: loadjson.m 371 2012-06-20 12:43:06Z fangq $ +% +% input: +% key: a string with which one look up a value from a struct +% default: if the key does not exist, return default +% optstruct: a struct where each sub-field is a key +% +% output: +% val: if key exists, val=optstruct.key; otherwise val=default +% +% license: +% BSD, see LICENSE_BSD.txt files for details +% +% -- this function is part of jsonlab toolbox (http://iso2mesh.sf.net/cgi-bin/index.cgi?jsonlab) +% + +val=default; +if(nargin<=2) return; end +opt=varargin{1}; +if(isstruct(opt) && isfield(opt,key)) + val=getfield(opt,key); +end + diff --git a/machine-learning-ex7/ex7/lib/jsonlab/loadjson.m b/machine-learning-ex7/ex7/lib/jsonlab/loadjson.m new file mode 100644 index 0000000..42798c0 --- /dev/null +++ b/machine-learning-ex7/ex7/lib/jsonlab/loadjson.m @@ -0,0 +1,566 @@ +function data = loadjson(fname,varargin) +% +% data=loadjson(fname,opt) +% or +% data=loadjson(fname,'param1',value1,'param2',value2,...) +% +% parse a JSON (JavaScript Object Notation) file or string +% +% authors:Qianqian Fang (fangq nmr.mgh.harvard.edu) +% created on 2011/09/09, including previous works from +% +% Nedialko Krouchev: http://www.mathworks.com/matlabcentral/fileexchange/25713 +% created on 2009/11/02 +% François Glineur: http://www.mathworks.com/matlabcentral/fileexchange/23393 +% created on 2009/03/22 +% Joel Feenstra: +% http://www.mathworks.com/matlabcentral/fileexchange/20565 +% created on 2008/07/03 +% +% $Id: loadjson.m 460 2015-01-03 00:30:45Z fangq $ +% +% input: +% fname: input file name, if fname contains "{}" or "[]", fname +% will be interpreted as a JSON string +% opt: a struct to store parsing options, opt can be replaced by +% a list of ('param',value) pairs - the param string is equivallent +% to a field in opt. opt can have the following +% fields (first in [.|.] is the default) +% +% opt.SimplifyCell [0|1]: if set to 1, loadjson will call cell2mat +% for each element of the JSON data, and group +% arrays based on the cell2mat rules. +% opt.FastArrayParser [1|0 or integer]: if set to 1, use a +% speed-optimized array parser when loading an +% array object. The fast array parser may +% collapse block arrays into a single large +% array similar to rules defined in cell2mat; 0 to +% use a legacy parser; if set to a larger-than-1 +% value, this option will specify the minimum +% dimension to enable the fast array parser. For +% example, if the input is a 3D array, setting +% FastArrayParser to 1 will return a 3D array; +% setting to 2 will return a cell array of 2D +% arrays; setting to 3 will return to a 2D cell +% array of 1D vectors; setting to 4 will return a +% 3D cell array. +% opt.ShowProgress [0|1]: if set to 1, loadjson displays a progress bar. +% +% output: +% dat: a cell array, where {...} blocks are converted into cell arrays, +% and [...] are converted to arrays +% +% examples: +% dat=loadjson('{"obj":{"string":"value","array":[1,2,3]}}') +% dat=loadjson(['examples' filesep 'example1.json']) +% dat=loadjson(['examples' filesep 'example1.json'],'SimplifyCell',1) +% +% license: +% BSD, see LICENSE_BSD.txt files for details +% +% -- this function is part of JSONLab toolbox (http://iso2mesh.sf.net/cgi-bin/index.cgi?jsonlab) +% + +global pos inStr len esc index_esc len_esc isoct arraytoken + +if(regexp(fname,'[\{\}\]\[]','once')) + string=fname; +elseif(exist(fname,'file')) + fid = fopen(fname,'rb'); + string = fread(fid,inf,'uint8=>char')'; + fclose(fid); +else + error('input file does not exist'); +end + +pos = 1; len = length(string); inStr = string; +isoct=exist('OCTAVE_VERSION','builtin'); +arraytoken=find(inStr=='[' | inStr==']' | inStr=='"'); +jstr=regexprep(inStr,'\\\\',' '); +escquote=regexp(jstr,'\\"'); +arraytoken=sort([arraytoken escquote]); + +% String delimiters and escape chars identified to improve speed: +esc = find(inStr=='"' | inStr=='\' ); % comparable to: regexp(inStr, '["\\]'); +index_esc = 1; len_esc = length(esc); + +opt=varargin2struct(varargin{:}); + +if(jsonopt('ShowProgress',0,opt)==1) + opt.progressbar_=waitbar(0,'loading ...'); +end +jsoncount=1; +while pos <= len + switch(next_char) + case '{' + data{jsoncount} = parse_object(opt); + case '[' + data{jsoncount} = parse_array(opt); + otherwise + error_pos('Outer level structure must be an object or an array'); + end + jsoncount=jsoncount+1; +end % while + +jsoncount=length(data); +if(jsoncount==1 && iscell(data)) + data=data{1}; +end + +if(~isempty(data)) + if(isstruct(data)) % data can be a struct array + data=jstruct2array(data); + elseif(iscell(data)) + data=jcell2array(data); + end +end +if(isfield(opt,'progressbar_')) + close(opt.progressbar_); +end + +%% +function newdata=jcell2array(data) +len=length(data); +newdata=data; +for i=1:len + if(isstruct(data{i})) + newdata{i}=jstruct2array(data{i}); + elseif(iscell(data{i})) + newdata{i}=jcell2array(data{i}); + end +end + +%%------------------------------------------------------------------------- +function newdata=jstruct2array(data) +fn=fieldnames(data); +newdata=data; +len=length(data); +for i=1:length(fn) % depth-first + for j=1:len + if(isstruct(getfield(data(j),fn{i}))) + newdata(j)=setfield(newdata(j),fn{i},jstruct2array(getfield(data(j),fn{i}))); + end + end +end +if(~isempty(strmatch('x0x5F_ArrayType_',fn)) && ~isempty(strmatch('x0x5F_ArrayData_',fn))) + newdata=cell(len,1); + for j=1:len + ndata=cast(data(j).x0x5F_ArrayData_,data(j).x0x5F_ArrayType_); + iscpx=0; + if(~isempty(strmatch('x0x5F_ArrayIsComplex_',fn))) + if(data(j).x0x5F_ArrayIsComplex_) + iscpx=1; + end + end + if(~isempty(strmatch('x0x5F_ArrayIsSparse_',fn))) + if(data(j).x0x5F_ArrayIsSparse_) + if(~isempty(strmatch('x0x5F_ArraySize_',fn))) + dim=data(j).x0x5F_ArraySize_; + if(iscpx && size(ndata,2)==4-any(dim==1)) + ndata(:,end-1)=complex(ndata(:,end-1),ndata(:,end)); + end + if isempty(ndata) + % All-zeros sparse + ndata=sparse(dim(1),prod(dim(2:end))); + elseif dim(1)==1 + % Sparse row vector + ndata=sparse(1,ndata(:,1),ndata(:,2),dim(1),prod(dim(2:end))); + elseif dim(2)==1 + % Sparse column vector + ndata=sparse(ndata(:,1),1,ndata(:,2),dim(1),prod(dim(2:end))); + else + % Generic sparse array. + ndata=sparse(ndata(:,1),ndata(:,2),ndata(:,3),dim(1),prod(dim(2:end))); + end + else + if(iscpx && size(ndata,2)==4) + ndata(:,3)=complex(ndata(:,3),ndata(:,4)); + end + ndata=sparse(ndata(:,1),ndata(:,2),ndata(:,3)); + end + end + elseif(~isempty(strmatch('x0x5F_ArraySize_',fn))) + if(iscpx && size(ndata,2)==2) + ndata=complex(ndata(:,1),ndata(:,2)); + end + ndata=reshape(ndata(:),data(j).x0x5F_ArraySize_); + end + newdata{j}=ndata; + end + if(len==1) + newdata=newdata{1}; + end +end + +%%------------------------------------------------------------------------- +function object = parse_object(varargin) + parse_char('{'); + object = []; + if next_char ~= '}' + while 1 + str = parseStr(varargin{:}); + if isempty(str) + error_pos('Name of value at position %d cannot be empty'); + end + parse_char(':'); + val = parse_value(varargin{:}); + eval( sprintf( 'object.%s = val;', valid_field(str) ) ); + if next_char == '}' + break; + end + parse_char(','); + end + end + parse_char('}'); + +%%------------------------------------------------------------------------- + +function object = parse_array(varargin) % JSON array is written in row-major order +global pos inStr isoct + parse_char('['); + object = cell(0, 1); + dim2=[]; + arraydepth=jsonopt('JSONLAB_ArrayDepth_',1,varargin{:}); + pbar=jsonopt('progressbar_',-1,varargin{:}); + + if next_char ~= ']' + if(jsonopt('FastArrayParser',1,varargin{:})>=1 && arraydepth>=jsonopt('FastArrayParser',1,varargin{:})) + [endpos, e1l, e1r, maxlevel]=matching_bracket(inStr,pos); + arraystr=['[' inStr(pos:endpos)]; + arraystr=regexprep(arraystr,'"_NaN_"','NaN'); + arraystr=regexprep(arraystr,'"([-+]*)_Inf_"','$1Inf'); + arraystr(arraystr==sprintf('\n'))=[]; + arraystr(arraystr==sprintf('\r'))=[]; + %arraystr=regexprep(arraystr,'\s*,',','); % this is slow,sometimes needed + if(~isempty(e1l) && ~isempty(e1r)) % the array is in 2D or higher D + astr=inStr((e1l+1):(e1r-1)); + astr=regexprep(astr,'"_NaN_"','NaN'); + astr=regexprep(astr,'"([-+]*)_Inf_"','$1Inf'); + astr(astr==sprintf('\n'))=[]; + astr(astr==sprintf('\r'))=[]; + astr(astr==' ')=''; + if(isempty(find(astr=='[', 1))) % array is 2D + dim2=length(sscanf(astr,'%f,',[1 inf])); + end + else % array is 1D + astr=arraystr(2:end-1); + astr(astr==' ')=''; + [obj, count, errmsg, nextidx]=sscanf(astr,'%f,',[1,inf]); + if(nextidx>=length(astr)-1) + object=obj; + pos=endpos; + parse_char(']'); + return; + end + end + if(~isempty(dim2)) + astr=arraystr; + astr(astr=='[')=''; + astr(astr==']')=''; + astr(astr==' ')=''; + [obj, count, errmsg, nextidx]=sscanf(astr,'%f,',inf); + if(nextidx>=length(astr)-1) + object=reshape(obj,dim2,numel(obj)/dim2)'; + pos=endpos; + parse_char(']'); + if(pbar>0) + waitbar(pos/length(inStr),pbar,'loading ...'); + end + return; + end + end + arraystr=regexprep(arraystr,'\]\s*,','];'); + else + arraystr='['; + end + try + if(isoct && regexp(arraystr,'"','once')) + error('Octave eval can produce empty cells for JSON-like input'); + end + object=eval(arraystr); + pos=endpos; + catch + while 1 + newopt=varargin2struct(varargin{:},'JSONLAB_ArrayDepth_',arraydepth+1); + val = parse_value(newopt); + object{end+1} = val; + if next_char == ']' + break; + end + parse_char(','); + end + end + end + if(jsonopt('SimplifyCell',0,varargin{:})==1) + try + oldobj=object; + object=cell2mat(object')'; + if(iscell(oldobj) && isstruct(object) && numel(object)>1 && jsonopt('SimplifyCellArray',1,varargin{:})==0) + object=oldobj; + elseif(size(object,1)>1 && ndims(object)==2) + object=object'; + end + catch + end + end + parse_char(']'); + + if(pbar>0) + waitbar(pos/length(inStr),pbar,'loading ...'); + end +%%------------------------------------------------------------------------- + +function parse_char(c) + global pos inStr len + skip_whitespace; + if pos > len || inStr(pos) ~= c + error_pos(sprintf('Expected %c at position %%d', c)); + else + pos = pos + 1; + skip_whitespace; + end + +%%------------------------------------------------------------------------- + +function c = next_char + global pos inStr len + skip_whitespace; + if pos > len + c = []; + else + c = inStr(pos); + end + +%%------------------------------------------------------------------------- + +function skip_whitespace + global pos inStr len + while pos <= len && isspace(inStr(pos)) + pos = pos + 1; + end + +%%------------------------------------------------------------------------- +function str = parseStr(varargin) + global pos inStr len esc index_esc len_esc + % len, ns = length(inStr), keyboard + if inStr(pos) ~= '"' + error_pos('String starting with " expected at position %d'); + else + pos = pos + 1; + end + str = ''; + while pos <= len + while index_esc <= len_esc && esc(index_esc) < pos + index_esc = index_esc + 1; + end + if index_esc > len_esc + str = [str inStr(pos:len)]; + pos = len + 1; + break; + else + str = [str inStr(pos:esc(index_esc)-1)]; + pos = esc(index_esc); + end + nstr = length(str); switch inStr(pos) + case '"' + pos = pos + 1; + if(~isempty(str)) + if(strcmp(str,'_Inf_')) + str=Inf; + elseif(strcmp(str,'-_Inf_')) + str=-Inf; + elseif(strcmp(str,'_NaN_')) + str=NaN; + end + end + return; + case '\' + if pos+1 > len + error_pos('End of file reached right after escape character'); + end + pos = pos + 1; + switch inStr(pos) + case {'"' '\' '/'} + str(nstr+1) = inStr(pos); + pos = pos + 1; + case {'b' 'f' 'n' 'r' 't'} + str(nstr+1) = sprintf(['\' inStr(pos)]); + pos = pos + 1; + case 'u' + if pos+4 > len + error_pos('End of file reached in escaped unicode character'); + end + str(nstr+(1:6)) = inStr(pos-1:pos+4); + pos = pos + 5; + end + otherwise % should never happen + str(nstr+1) = inStr(pos), keyboard + pos = pos + 1; + end + end + error_pos('End of file while expecting end of inStr'); + +%%------------------------------------------------------------------------- + +function num = parse_number(varargin) + global pos inStr len isoct + currstr=inStr(pos:end); + numstr=0; + if(isoct~=0) + numstr=regexp(currstr,'^\s*-?(?:0|[1-9]\d*)(?:\.\d+)?(?:[eE][+\-]?\d+)?','end'); + [num, one] = sscanf(currstr, '%f', 1); + delta=numstr+1; + else + [num, one, err, delta] = sscanf(currstr, '%f', 1); + if ~isempty(err) + error_pos('Error reading number at position %d'); + end + end + pos = pos + delta-1; + +%%------------------------------------------------------------------------- + +function val = parse_value(varargin) + global pos inStr len + true = 1; false = 0; + + pbar=jsonopt('progressbar_',-1,varargin{:}); + if(pbar>0) + waitbar(pos/len,pbar,'loading ...'); + end + + switch(inStr(pos)) + case '"' + val = parseStr(varargin{:}); + return; + case '[' + val = parse_array(varargin{:}); + return; + case '{' + val = parse_object(varargin{:}); + if isstruct(val) + if(~isempty(strmatch('x0x5F_ArrayType_',fieldnames(val), 'exact'))) + val=jstruct2array(val); + end + elseif isempty(val) + val = struct; + end + return; + case {'-','0','1','2','3','4','5','6','7','8','9'} + val = parse_number(varargin{:}); + return; + case 't' + if pos+3 <= len && strcmpi(inStr(pos:pos+3), 'true') + val = true; + pos = pos + 4; + return; + end + case 'f' + if pos+4 <= len && strcmpi(inStr(pos:pos+4), 'false') + val = false; + pos = pos + 5; + return; + end + case 'n' + if pos+3 <= len && strcmpi(inStr(pos:pos+3), 'null') + val = []; + pos = pos + 4; + return; + end + end + error_pos('Value expected at position %d'); +%%------------------------------------------------------------------------- + +function error_pos(msg) + global pos inStr len + poShow = max(min([pos-15 pos-1 pos pos+20],len),1); + if poShow(3) == poShow(2) + poShow(3:4) = poShow(2)+[0 -1]; % display nothing after + end + msg = [sprintf(msg, pos) ': ' ... + inStr(poShow(1):poShow(2)) '' inStr(poShow(3):poShow(4)) ]; + error( ['JSONparser:invalidFormat: ' msg] ); + +%%------------------------------------------------------------------------- + +function str = valid_field(str) +global isoct +% From MATLAB doc: field names must begin with a letter, which may be +% followed by any combination of letters, digits, and underscores. +% Invalid characters will be converted to underscores, and the prefix +% "x0x[Hex code]_" will be added if the first character is not a letter. + pos=regexp(str,'^[^A-Za-z]','once'); + if(~isempty(pos)) + if(~isoct) + str=regexprep(str,'^([^A-Za-z])','x0x${sprintf(''%X'',unicode2native($1))}_','once'); + else + str=sprintf('x0x%X_%s',char(str(1)),str(2:end)); + end + end + if(isempty(regexp(str,'[^0-9A-Za-z_]', 'once' ))) return; end + if(~isoct) + str=regexprep(str,'([^0-9A-Za-z_])','_0x${sprintf(''%X'',unicode2native($1))}_'); + else + pos=regexp(str,'[^0-9A-Za-z_]'); + if(isempty(pos)) return; end + str0=str; + pos0=[0 pos(:)' length(str)]; + str=''; + for i=1:length(pos) + str=[str str0(pos0(i)+1:pos(i)-1) sprintf('_0x%X_',str0(pos(i)))]; + end + if(pos(end)~=length(str)) + str=[str str0(pos0(end-1)+1:pos0(end))]; + end + end + %str(~isletter(str) & ~('0' <= str & str <= '9')) = '_'; + +%%------------------------------------------------------------------------- +function endpos = matching_quote(str,pos) +len=length(str); +while(pos1 && str(pos-1)=='\')) + endpos=pos; + return; + end + end + pos=pos+1; +end +error('unmatched quotation mark'); +%%------------------------------------------------------------------------- +function [endpos, e1l, e1r, maxlevel] = matching_bracket(str,pos) +global arraytoken +level=1; +maxlevel=level; +endpos=0; +bpos=arraytoken(arraytoken>=pos); +tokens=str(bpos); +len=length(tokens); +pos=1; +e1l=[]; +e1r=[]; +while(pos<=len) + c=tokens(pos); + if(c==']') + level=level-1; + if(isempty(e1r)) e1r=bpos(pos); end + if(level==0) + endpos=bpos(pos); + return + end + end + if(c=='[') + if(isempty(e1l)) e1l=bpos(pos); end + level=level+1; + maxlevel=max(maxlevel,level); + end + if(c=='"') + pos=matching_quote(tokens,pos+1); + end + pos=pos+1; +end +if(endpos==0) + error('unmatched "]"'); +end + diff --git a/machine-learning-ex7/ex7/lib/jsonlab/loadubjson.m b/machine-learning-ex7/ex7/lib/jsonlab/loadubjson.m new file mode 100644 index 0000000..0155115 --- /dev/null +++ b/machine-learning-ex7/ex7/lib/jsonlab/loadubjson.m @@ -0,0 +1,528 @@ +function data = loadubjson(fname,varargin) +% +% data=loadubjson(fname,opt) +% or +% data=loadubjson(fname,'param1',value1,'param2',value2,...) +% +% parse a JSON (JavaScript Object Notation) file or string +% +% authors:Qianqian Fang (fangq nmr.mgh.harvard.edu) +% created on 2013/08/01 +% +% $Id: loadubjson.m 460 2015-01-03 00:30:45Z fangq $ +% +% input: +% fname: input file name, if fname contains "{}" or "[]", fname +% will be interpreted as a UBJSON string +% opt: a struct to store parsing options, opt can be replaced by +% a list of ('param',value) pairs - the param string is equivallent +% to a field in opt. opt can have the following +% fields (first in [.|.] is the default) +% +% opt.SimplifyCell [0|1]: if set to 1, loadubjson will call cell2mat +% for each element of the JSON data, and group +% arrays based on the cell2mat rules. +% opt.IntEndian [B|L]: specify the endianness of the integer fields +% in the UBJSON input data. B - Big-Endian format for +% integers (as required in the UBJSON specification); +% L - input integer fields are in Little-Endian order. +% +% output: +% dat: a cell array, where {...} blocks are converted into cell arrays, +% and [...] are converted to arrays +% +% examples: +% obj=struct('string','value','array',[1 2 3]); +% ubjdata=saveubjson('obj',obj); +% dat=loadubjson(ubjdata) +% dat=loadubjson(['examples' filesep 'example1.ubj']) +% dat=loadubjson(['examples' filesep 'example1.ubj'],'SimplifyCell',1) +% +% license: +% BSD, see LICENSE_BSD.txt files for details +% +% -- this function is part of JSONLab toolbox (http://iso2mesh.sf.net/cgi-bin/index.cgi?jsonlab) +% + +global pos inStr len esc index_esc len_esc isoct arraytoken fileendian systemendian + +if(regexp(fname,'[\{\}\]\[]','once')) + string=fname; +elseif(exist(fname,'file')) + fid = fopen(fname,'rb'); + string = fread(fid,inf,'uint8=>char')'; + fclose(fid); +else + error('input file does not exist'); +end + +pos = 1; len = length(string); inStr = string; +isoct=exist('OCTAVE_VERSION','builtin'); +arraytoken=find(inStr=='[' | inStr==']' | inStr=='"'); +jstr=regexprep(inStr,'\\\\',' '); +escquote=regexp(jstr,'\\"'); +arraytoken=sort([arraytoken escquote]); + +% String delimiters and escape chars identified to improve speed: +esc = find(inStr=='"' | inStr=='\' ); % comparable to: regexp(inStr, '["\\]'); +index_esc = 1; len_esc = length(esc); + +opt=varargin2struct(varargin{:}); +fileendian=upper(jsonopt('IntEndian','B',opt)); +[os,maxelem,systemendian]=computer; + +jsoncount=1; +while pos <= len + switch(next_char) + case '{' + data{jsoncount} = parse_object(opt); + case '[' + data{jsoncount} = parse_array(opt); + otherwise + error_pos('Outer level structure must be an object or an array'); + end + jsoncount=jsoncount+1; +end % while + +jsoncount=length(data); +if(jsoncount==1 && iscell(data)) + data=data{1}; +end + +if(~isempty(data)) + if(isstruct(data)) % data can be a struct array + data=jstruct2array(data); + elseif(iscell(data)) + data=jcell2array(data); + end +end + + +%% +function newdata=parse_collection(id,data,obj) + +if(jsoncount>0 && exist('data','var')) + if(~iscell(data)) + newdata=cell(1); + newdata{1}=data; + data=newdata; + end +end + +%% +function newdata=jcell2array(data) +len=length(data); +newdata=data; +for i=1:len + if(isstruct(data{i})) + newdata{i}=jstruct2array(data{i}); + elseif(iscell(data{i})) + newdata{i}=jcell2array(data{i}); + end +end + +%%------------------------------------------------------------------------- +function newdata=jstruct2array(data) +fn=fieldnames(data); +newdata=data; +len=length(data); +for i=1:length(fn) % depth-first + for j=1:len + if(isstruct(getfield(data(j),fn{i}))) + newdata(j)=setfield(newdata(j),fn{i},jstruct2array(getfield(data(j),fn{i}))); + end + end +end +if(~isempty(strmatch('x0x5F_ArrayType_',fn)) && ~isempty(strmatch('x0x5F_ArrayData_',fn))) + newdata=cell(len,1); + for j=1:len + ndata=cast(data(j).x0x5F_ArrayData_,data(j).x0x5F_ArrayType_); + iscpx=0; + if(~isempty(strmatch('x0x5F_ArrayIsComplex_',fn))) + if(data(j).x0x5F_ArrayIsComplex_) + iscpx=1; + end + end + if(~isempty(strmatch('x0x5F_ArrayIsSparse_',fn))) + if(data(j).x0x5F_ArrayIsSparse_) + if(~isempty(strmatch('x0x5F_ArraySize_',fn))) + dim=double(data(j).x0x5F_ArraySize_); + if(iscpx && size(ndata,2)==4-any(dim==1)) + ndata(:,end-1)=complex(ndata(:,end-1),ndata(:,end)); + end + if isempty(ndata) + % All-zeros sparse + ndata=sparse(dim(1),prod(dim(2:end))); + elseif dim(1)==1 + % Sparse row vector + ndata=sparse(1,ndata(:,1),ndata(:,2),dim(1),prod(dim(2:end))); + elseif dim(2)==1 + % Sparse column vector + ndata=sparse(ndata(:,1),1,ndata(:,2),dim(1),prod(dim(2:end))); + else + % Generic sparse array. + ndata=sparse(ndata(:,1),ndata(:,2),ndata(:,3),dim(1),prod(dim(2:end))); + end + else + if(iscpx && size(ndata,2)==4) + ndata(:,3)=complex(ndata(:,3),ndata(:,4)); + end + ndata=sparse(ndata(:,1),ndata(:,2),ndata(:,3)); + end + end + elseif(~isempty(strmatch('x0x5F_ArraySize_',fn))) + if(iscpx && size(ndata,2)==2) + ndata=complex(ndata(:,1),ndata(:,2)); + end + ndata=reshape(ndata(:),data(j).x0x5F_ArraySize_); + end + newdata{j}=ndata; + end + if(len==1) + newdata=newdata{1}; + end +end + +%%------------------------------------------------------------------------- +function object = parse_object(varargin) + parse_char('{'); + object = []; + type=''; + count=-1; + if(next_char == '$') + type=inStr(pos+1); % TODO + pos=pos+2; + end + if(next_char == '#') + pos=pos+1; + count=double(parse_number()); + end + if next_char ~= '}' + num=0; + while 1 + str = parseStr(varargin{:}); + if isempty(str) + error_pos('Name of value at position %d cannot be empty'); + end + %parse_char(':'); + val = parse_value(varargin{:}); + num=num+1; + eval( sprintf( 'object.%s = val;', valid_field(str) ) ); + if next_char == '}' || (count>=0 && num>=count) + break; + end + %parse_char(','); + end + end + if(count==-1) + parse_char('}'); + end + +%%------------------------------------------------------------------------- +function [cid,len]=elem_info(type) +id=strfind('iUIlLdD',type); +dataclass={'int8','uint8','int16','int32','int64','single','double'}; +bytelen=[1,1,2,4,8,4,8]; +if(id>0) + cid=dataclass{id}; + len=bytelen(id); +else + error_pos('unsupported type at position %d'); +end +%%------------------------------------------------------------------------- + + +function [data adv]=parse_block(type,count,varargin) +global pos inStr isoct fileendian systemendian +[cid,len]=elem_info(type); +datastr=inStr(pos:pos+len*count-1); +if(isoct) + newdata=int8(datastr); +else + newdata=uint8(datastr); +end +id=strfind('iUIlLdD',type); +if(id<=5 && fileendian~=systemendian) + newdata=swapbytes(typecast(newdata,cid)); +end +data=typecast(newdata,cid); +adv=double(len*count); + +%%------------------------------------------------------------------------- + + +function object = parse_array(varargin) % JSON array is written in row-major order +global pos inStr isoct + parse_char('['); + object = cell(0, 1); + dim=[]; + type=''; + count=-1; + if(next_char == '$') + type=inStr(pos+1); + pos=pos+2; + end + if(next_char == '#') + pos=pos+1; + if(next_char=='[') + dim=parse_array(varargin{:}); + count=prod(double(dim)); + else + count=double(parse_number()); + end + end + if(~isempty(type)) + if(count>=0) + [object adv]=parse_block(type,count,varargin{:}); + if(~isempty(dim)) + object=reshape(object,dim); + end + pos=pos+adv; + return; + else + endpos=matching_bracket(inStr,pos); + [cid,len]=elem_info(type); + count=(endpos-pos)/len; + [object adv]=parse_block(type,count,varargin{:}); + pos=pos+adv; + parse_char(']'); + return; + end + end + if next_char ~= ']' + while 1 + val = parse_value(varargin{:}); + object{end+1} = val; + if next_char == ']' + break; + end + %parse_char(','); + end + end + if(jsonopt('SimplifyCell',0,varargin{:})==1) + try + oldobj=object; + object=cell2mat(object')'; + if(iscell(oldobj) && isstruct(object) && numel(object)>1 && jsonopt('SimplifyCellArray',1,varargin{:})==0) + object=oldobj; + elseif(size(object,1)>1 && ndims(object)==2) + object=object'; + end + catch + end + end + if(count==-1) + parse_char(']'); + end + +%%------------------------------------------------------------------------- + +function parse_char(c) + global pos inStr len + skip_whitespace; + if pos > len || inStr(pos) ~= c + error_pos(sprintf('Expected %c at position %%d', c)); + else + pos = pos + 1; + skip_whitespace; + end + +%%------------------------------------------------------------------------- + +function c = next_char + global pos inStr len + skip_whitespace; + if pos > len + c = []; + else + c = inStr(pos); + end + +%%------------------------------------------------------------------------- + +function skip_whitespace + global pos inStr len + while pos <= len && isspace(inStr(pos)) + pos = pos + 1; + end + +%%------------------------------------------------------------------------- +function str = parseStr(varargin) + global pos inStr esc index_esc len_esc + % len, ns = length(inStr), keyboard + type=inStr(pos); + if type ~= 'S' && type ~= 'C' && type ~= 'H' + error_pos('String starting with S expected at position %d'); + else + pos = pos + 1; + end + if(type == 'C') + str=inStr(pos); + pos=pos+1; + return; + end + bytelen=double(parse_number()); + if(length(inStr)>=pos+bytelen-1) + str=inStr(pos:pos+bytelen-1); + pos=pos+bytelen; + else + error_pos('End of file while expecting end of inStr'); + end + +%%------------------------------------------------------------------------- + +function num = parse_number(varargin) + global pos inStr len isoct fileendian systemendian + id=strfind('iUIlLdD',inStr(pos)); + if(isempty(id)) + error_pos('expecting a number at position %d'); + end + type={'int8','uint8','int16','int32','int64','single','double'}; + bytelen=[1,1,2,4,8,4,8]; + datastr=inStr(pos+1:pos+bytelen(id)); + if(isoct) + newdata=int8(datastr); + else + newdata=uint8(datastr); + end + if(id<=5 && fileendian~=systemendian) + newdata=swapbytes(typecast(newdata,type{id})); + end + num=typecast(newdata,type{id}); + pos = pos + bytelen(id)+1; + +%%------------------------------------------------------------------------- + +function val = parse_value(varargin) + global pos inStr len + true = 1; false = 0; + + switch(inStr(pos)) + case {'S','C','H'} + val = parseStr(varargin{:}); + return; + case '[' + val = parse_array(varargin{:}); + return; + case '{' + val = parse_object(varargin{:}); + if isstruct(val) + if(~isempty(strmatch('x0x5F_ArrayType_',fieldnames(val), 'exact'))) + val=jstruct2array(val); + end + elseif isempty(val) + val = struct; + end + return; + case {'i','U','I','l','L','d','D'} + val = parse_number(varargin{:}); + return; + case 'T' + val = true; + pos = pos + 1; + return; + case 'F' + val = false; + pos = pos + 1; + return; + case {'Z','N'} + val = []; + pos = pos + 1; + return; + end + error_pos('Value expected at position %d'); +%%------------------------------------------------------------------------- + +function error_pos(msg) + global pos inStr len + poShow = max(min([pos-15 pos-1 pos pos+20],len),1); + if poShow(3) == poShow(2) + poShow(3:4) = poShow(2)+[0 -1]; % display nothing after + end + msg = [sprintf(msg, pos) ': ' ... + inStr(poShow(1):poShow(2)) '' inStr(poShow(3):poShow(4)) ]; + error( ['JSONparser:invalidFormat: ' msg] ); + +%%------------------------------------------------------------------------- + +function str = valid_field(str) +global isoct +% From MATLAB doc: field names must begin with a letter, which may be +% followed by any combination of letters, digits, and underscores. +% Invalid characters will be converted to underscores, and the prefix +% "x0x[Hex code]_" will be added if the first character is not a letter. + pos=regexp(str,'^[^A-Za-z]','once'); + if(~isempty(pos)) + if(~isoct) + str=regexprep(str,'^([^A-Za-z])','x0x${sprintf(''%X'',unicode2native($1))}_','once'); + else + str=sprintf('x0x%X_%s',char(str(1)),str(2:end)); + end + end + if(isempty(regexp(str,'[^0-9A-Za-z_]', 'once' ))) return; end + if(~isoct) + str=regexprep(str,'([^0-9A-Za-z_])','_0x${sprintf(''%X'',unicode2native($1))}_'); + else + pos=regexp(str,'[^0-9A-Za-z_]'); + if(isempty(pos)) return; end + str0=str; + pos0=[0 pos(:)' length(str)]; + str=''; + for i=1:length(pos) + str=[str str0(pos0(i)+1:pos(i)-1) sprintf('_0x%X_',str0(pos(i)))]; + end + if(pos(end)~=length(str)) + str=[str str0(pos0(end-1)+1:pos0(end))]; + end + end + %str(~isletter(str) & ~('0' <= str & str <= '9')) = '_'; + +%%------------------------------------------------------------------------- +function endpos = matching_quote(str,pos) +len=length(str); +while(pos1 && str(pos-1)=='\')) + endpos=pos; + return; + end + end + pos=pos+1; +end +error('unmatched quotation mark'); +%%------------------------------------------------------------------------- +function [endpos e1l e1r maxlevel] = matching_bracket(str,pos) +global arraytoken +level=1; +maxlevel=level; +endpos=0; +bpos=arraytoken(arraytoken>=pos); +tokens=str(bpos); +len=length(tokens); +pos=1; +e1l=[]; +e1r=[]; +while(pos<=len) + c=tokens(pos); + if(c==']') + level=level-1; + if(isempty(e1r)) e1r=bpos(pos); end + if(level==0) + endpos=bpos(pos); + return + end + end + if(c=='[') + if(isempty(e1l)) e1l=bpos(pos); end + level=level+1; + maxlevel=max(maxlevel,level); + end + if(c=='"') + pos=matching_quote(tokens,pos+1); + end + pos=pos+1; +end +if(endpos==0) + error('unmatched "]"'); +end + diff --git a/machine-learning-ex7/ex7/lib/jsonlab/mergestruct.m b/machine-learning-ex7/ex7/lib/jsonlab/mergestruct.m new file mode 100644 index 0000000..6de6100 --- /dev/null +++ b/machine-learning-ex7/ex7/lib/jsonlab/mergestruct.m @@ -0,0 +1,33 @@ +function s=mergestruct(s1,s2) +% +% s=mergestruct(s1,s2) +% +% merge two struct objects into one +% +% authors:Qianqian Fang (fangq nmr.mgh.harvard.edu) +% date: 2012/12/22 +% +% input: +% s1,s2: a struct object, s1 and s2 can not be arrays +% +% output: +% s: the merged struct object. fields in s1 and s2 will be combined in s. +% +% license: +% BSD, see LICENSE_BSD.txt files for details +% +% -- this function is part of jsonlab toolbox (http://iso2mesh.sf.net/cgi-bin/index.cgi?jsonlab) +% + +if(~isstruct(s1) || ~isstruct(s2)) + error('input parameters contain non-struct'); +end +if(length(s1)>1 || length(s2)>1) + error('can not merge struct arrays'); +end +fn=fieldnames(s2); +s=s1; +for i=1:length(fn) + s=setfield(s,fn{i},getfield(s2,fn{i})); +end + diff --git a/machine-learning-ex7/ex7/lib/jsonlab/savejson.m b/machine-learning-ex7/ex7/lib/jsonlab/savejson.m new file mode 100644 index 0000000..7e84076 --- /dev/null +++ b/machine-learning-ex7/ex7/lib/jsonlab/savejson.m @@ -0,0 +1,475 @@ +function json=savejson(rootname,obj,varargin) +% +% json=savejson(rootname,obj,filename) +% or +% json=savejson(rootname,obj,opt) +% json=savejson(rootname,obj,'param1',value1,'param2',value2,...) +% +% convert a MATLAB object (cell, struct or array) into a JSON (JavaScript +% Object Notation) string +% +% author: Qianqian Fang (fangq nmr.mgh.harvard.edu) +% created on 2011/09/09 +% +% $Id: savejson.m 460 2015-01-03 00:30:45Z fangq $ +% +% input: +% rootname: the name of the root-object, when set to '', the root name +% is ignored, however, when opt.ForceRootName is set to 1 (see below), +% the MATLAB variable name will be used as the root name. +% obj: a MATLAB object (array, cell, cell array, struct, struct array). +% filename: a string for the file name to save the output JSON data. +% opt: a struct for additional options, ignore to use default values. +% opt can have the following fields (first in [.|.] is the default) +% +% opt.FileName [''|string]: a file name to save the output JSON data +% opt.FloatFormat ['%.10g'|string]: format to show each numeric element +% of a 1D/2D array; +% opt.ArrayIndent [1|0]: if 1, output explicit data array with +% precedent indentation; if 0, no indentation +% opt.ArrayToStruct[0|1]: when set to 0, savejson outputs 1D/2D +% array in JSON array format; if sets to 1, an +% array will be shown as a struct with fields +% "_ArrayType_", "_ArraySize_" and "_ArrayData_"; for +% sparse arrays, the non-zero elements will be +% saved to _ArrayData_ field in triplet-format i.e. +% (ix,iy,val) and "_ArrayIsSparse_" will be added +% with a value of 1; for a complex array, the +% _ArrayData_ array will include two columns +% (4 for sparse) to record the real and imaginary +% parts, and also "_ArrayIsComplex_":1 is added. +% opt.ParseLogical [0|1]: if this is set to 1, logical array elem +% will use true/false rather than 1/0. +% opt.NoRowBracket [1|0]: if this is set to 1, arrays with a single +% numerical element will be shown without a square +% bracket, unless it is the root object; if 0, square +% brackets are forced for any numerical arrays. +% opt.ForceRootName [0|1]: when set to 1 and rootname is empty, savejson +% will use the name of the passed obj variable as the +% root object name; if obj is an expression and +% does not have a name, 'root' will be used; if this +% is set to 0 and rootname is empty, the root level +% will be merged down to the lower level. +% opt.Inf ['"$1_Inf_"'|string]: a customized regular expression pattern +% to represent +/-Inf. The matched pattern is '([-+]*)Inf' +% and $1 represents the sign. For those who want to use +% 1e999 to represent Inf, they can set opt.Inf to '$11e999' +% opt.NaN ['"_NaN_"'|string]: a customized regular expression pattern +% to represent NaN +% opt.JSONP [''|string]: to generate a JSONP output (JSON with padding), +% for example, if opt.JSONP='foo', the JSON data is +% wrapped inside a function call as 'foo(...);' +% opt.UnpackHex [1|0]: conver the 0x[hex code] output by loadjson +% back to the string form +% opt.SaveBinary [0|1]: 1 - save the JSON file in binary mode; 0 - text mode. +% opt.Compact [0|1]: 1- out compact JSON format (remove all newlines and tabs) +% +% opt can be replaced by a list of ('param',value) pairs. The param +% string is equivallent to a field in opt and is case sensitive. +% output: +% json: a string in the JSON format (see http://json.org) +% +% examples: +% jsonmesh=struct('MeshNode',[0 0 0;1 0 0;0 1 0;1 1 0;0 0 1;1 0 1;0 1 1;1 1 1],... +% 'MeshTetra',[1 2 4 8;1 3 4 8;1 2 6 8;1 5 6 8;1 5 7 8;1 3 7 8],... +% 'MeshTri',[1 2 4;1 2 6;1 3 4;1 3 7;1 5 6;1 5 7;... +% 2 8 4;2 8 6;3 8 4;3 8 7;5 8 6;5 8 7],... +% 'MeshCreator','FangQ','MeshTitle','T6 Cube',... +% 'SpecialData',[nan, inf, -inf]); +% savejson('jmesh',jsonmesh) +% savejson('',jsonmesh,'ArrayIndent',0,'FloatFormat','\t%.5g') +% +% license: +% BSD, see LICENSE_BSD.txt files for details +% +% -- this function is part of JSONLab toolbox (http://iso2mesh.sf.net/cgi-bin/index.cgi?jsonlab) +% + +if(nargin==1) + varname=inputname(1); + obj=rootname; + if(isempty(varname)) + varname='root'; + end + rootname=varname; +else + varname=inputname(2); +end +if(length(varargin)==1 && ischar(varargin{1})) + opt=struct('FileName',varargin{1}); +else + opt=varargin2struct(varargin{:}); +end +opt.IsOctave=exist('OCTAVE_VERSION','builtin'); +rootisarray=0; +rootlevel=1; +forceroot=jsonopt('ForceRootName',0,opt); +if((isnumeric(obj) || islogical(obj) || ischar(obj) || isstruct(obj) || iscell(obj)) && isempty(rootname) && forceroot==0) + rootisarray=1; + rootlevel=0; +else + if(isempty(rootname)) + rootname=varname; + end +end +if((isstruct(obj) || iscell(obj))&& isempty(rootname) && forceroot) + rootname='root'; +end + +whitespaces=struct('tab',sprintf('\t'),'newline',sprintf('\n'),'sep',sprintf(',\n')); +if(jsonopt('Compact',0,opt)==1) + whitespaces=struct('tab','','newline','','sep',','); +end +if(~isfield(opt,'whitespaces_')) + opt.whitespaces_=whitespaces; +end + +nl=whitespaces.newline; + +json=obj2json(rootname,obj,rootlevel,opt); +if(rootisarray) + json=sprintf('%s%s',json,nl); +else + json=sprintf('{%s%s%s}\n',nl,json,nl); +end + +jsonp=jsonopt('JSONP','',opt); +if(~isempty(jsonp)) + json=sprintf('%s(%s);%s',jsonp,json,nl); +end + +% save to a file if FileName is set, suggested by Patrick Rapin +if(~isempty(jsonopt('FileName','',opt))) + if(jsonopt('SaveBinary',0,opt)==1) + fid = fopen(opt.FileName, 'wb'); + fwrite(fid,json); + else + fid = fopen(opt.FileName, 'wt'); + fwrite(fid,json,'char'); + end + fclose(fid); +end + +%%------------------------------------------------------------------------- +function txt=obj2json(name,item,level,varargin) + +if(iscell(item)) + txt=cell2json(name,item,level,varargin{:}); +elseif(isstruct(item)) + txt=struct2json(name,item,level,varargin{:}); +elseif(ischar(item)) + txt=str2json(name,item,level,varargin{:}); +else + txt=mat2json(name,item,level,varargin{:}); +end + +%%------------------------------------------------------------------------- +function txt=cell2json(name,item,level,varargin) +txt=''; +if(~iscell(item)) + error('input is not a cell'); +end + +dim=size(item); +if(ndims(squeeze(item))>2) % for 3D or higher dimensions, flatten to 2D for now + item=reshape(item,dim(1),numel(item)/dim(1)); + dim=size(item); +end +len=numel(item); +ws=jsonopt('whitespaces_',struct('tab',sprintf('\t'),'newline',sprintf('\n'),'sep',sprintf(',\n')),varargin{:}); +padding0=repmat(ws.tab,1,level); +padding2=repmat(ws.tab,1,level+1); +nl=ws.newline; +if(len>1) + if(~isempty(name)) + txt=sprintf('%s"%s": [%s',padding0, checkname(name,varargin{:}),nl); name=''; + else + txt=sprintf('%s[%s',padding0,nl); + end +elseif(len==0) + if(~isempty(name)) + txt=sprintf('%s"%s": []',padding0, checkname(name,varargin{:})); name=''; + else + txt=sprintf('%s[]',padding0); + end +end +for j=1:dim(2) + if(dim(1)>1) txt=sprintf('%s%s[%s',txt,padding2,nl); end + for i=1:dim(1) + txt=sprintf('%s%s',txt,obj2json(name,item{i,j},level+(dim(1)>1)+1,varargin{:})); + if(i1) txt=sprintf('%s%s%s]',txt,nl,padding2); end + if(j1) txt=sprintf('%s%s%s]',txt,nl,padding0); end + +%%------------------------------------------------------------------------- +function txt=struct2json(name,item,level,varargin) +txt=''; +if(~isstruct(item)) + error('input is not a struct'); +end +dim=size(item); +if(ndims(squeeze(item))>2) % for 3D or higher dimensions, flatten to 2D for now + item=reshape(item,dim(1),numel(item)/dim(1)); + dim=size(item); +end +len=numel(item); +ws=struct('tab',sprintf('\t'),'newline',sprintf('\n')); +ws=jsonopt('whitespaces_',ws,varargin{:}); +padding0=repmat(ws.tab,1,level); +padding2=repmat(ws.tab,1,level+1); +padding1=repmat(ws.tab,1,level+(dim(1)>1)+(len>1)); +nl=ws.newline; + +if(~isempty(name)) + if(len>1) txt=sprintf('%s"%s": [%s',padding0,checkname(name,varargin{:}),nl); end +else + if(len>1) txt=sprintf('%s[%s',padding0,nl); end +end +for j=1:dim(2) + if(dim(1)>1) txt=sprintf('%s%s[%s',txt,padding2,nl); end + for i=1:dim(1) + names = fieldnames(item(i,j)); + if(~isempty(name) && len==1) + txt=sprintf('%s%s"%s": {%s',txt,padding1, checkname(name,varargin{:}),nl); + else + txt=sprintf('%s%s{%s',txt,padding1,nl); + end + if(~isempty(names)) + for e=1:length(names) + txt=sprintf('%s%s',txt,obj2json(names{e},getfield(item(i,j),... + names{e}),level+(dim(1)>1)+1+(len>1),varargin{:})); + if(e1) txt=sprintf('%s%s%s]',txt,nl,padding2); end + if(j1) txt=sprintf('%s%s%s]',txt,nl,padding0); end + +%%------------------------------------------------------------------------- +function txt=str2json(name,item,level,varargin) +txt=''; +if(~ischar(item)) + error('input is not a string'); +end +item=reshape(item, max(size(item),[1 0])); +len=size(item,1); +ws=struct('tab',sprintf('\t'),'newline',sprintf('\n'),'sep',sprintf(',\n')); +ws=jsonopt('whitespaces_',ws,varargin{:}); +padding1=repmat(ws.tab,1,level); +padding0=repmat(ws.tab,1,level+1); +nl=ws.newline; +sep=ws.sep; + +if(~isempty(name)) + if(len>1) txt=sprintf('%s"%s": [%s',padding1,checkname(name,varargin{:}),nl); end +else + if(len>1) txt=sprintf('%s[%s',padding1,nl); end +end +isoct=jsonopt('IsOctave',0,varargin{:}); +for e=1:len + if(isoct) + val=regexprep(item(e,:),'\\','\\'); + val=regexprep(val,'"','\"'); + val=regexprep(val,'^"','\"'); + else + val=regexprep(item(e,:),'\\','\\\\'); + val=regexprep(val,'"','\\"'); + val=regexprep(val,'^"','\\"'); + end + val=escapejsonstring(val); + if(len==1) + obj=['"' checkname(name,varargin{:}) '": ' '"',val,'"']; + if(isempty(name)) obj=['"',val,'"']; end + txt=sprintf('%s%s%s%s',txt,padding1,obj); + else + txt=sprintf('%s%s%s%s',txt,padding0,['"',val,'"']); + end + if(e==len) sep=''; end + txt=sprintf('%s%s',txt,sep); +end +if(len>1) txt=sprintf('%s%s%s%s',txt,nl,padding1,']'); end + +%%------------------------------------------------------------------------- +function txt=mat2json(name,item,level,varargin) +if(~isnumeric(item) && ~islogical(item)) + error('input is not an array'); +end +ws=struct('tab',sprintf('\t'),'newline',sprintf('\n'),'sep',sprintf(',\n')); +ws=jsonopt('whitespaces_',ws,varargin{:}); +padding1=repmat(ws.tab,1,level); +padding0=repmat(ws.tab,1,level+1); +nl=ws.newline; +sep=ws.sep; + +if(length(size(item))>2 || issparse(item) || ~isreal(item) || ... + isempty(item) ||jsonopt('ArrayToStruct',0,varargin{:})) + if(isempty(name)) + txt=sprintf('%s{%s%s"_ArrayType_": "%s",%s%s"_ArraySize_": %s,%s',... + padding1,nl,padding0,class(item),nl,padding0,regexprep(mat2str(size(item)),'\s+',','),nl); + else + txt=sprintf('%s"%s": {%s%s"_ArrayType_": "%s",%s%s"_ArraySize_": %s,%s',... + padding1,checkname(name,varargin{:}),nl,padding0,class(item),nl,padding0,regexprep(mat2str(size(item)),'\s+',','),nl); + end +else + if(numel(item)==1 && jsonopt('NoRowBracket',1,varargin{:})==1 && level>0) + numtxt=regexprep(regexprep(matdata2json(item,level+1,varargin{:}),'^\[',''),']',''); + else + numtxt=matdata2json(item,level+1,varargin{:}); + end + if(isempty(name)) + txt=sprintf('%s%s',padding1,numtxt); + else + if(numel(item)==1 && jsonopt('NoRowBracket',1,varargin{:})==1) + txt=sprintf('%s"%s": %s',padding1,checkname(name,varargin{:}),numtxt); + else + txt=sprintf('%s"%s": %s',padding1,checkname(name,varargin{:}),numtxt); + end + end + return; +end +dataformat='%s%s%s%s%s'; + +if(issparse(item)) + [ix,iy]=find(item); + data=full(item(find(item))); + if(~isreal(item)) + data=[real(data(:)),imag(data(:))]; + if(size(item,1)==1) + % Kludge to have data's 'transposedness' match item's. + % (Necessary for complex row vector handling below.) + data=data'; + end + txt=sprintf(dataformat,txt,padding0,'"_ArrayIsComplex_": ','1', sep); + end + txt=sprintf(dataformat,txt,padding0,'"_ArrayIsSparse_": ','1', sep); + if(size(item,1)==1) + % Row vector, store only column indices. + txt=sprintf(dataformat,txt,padding0,'"_ArrayData_": ',... + matdata2json([iy(:),data'],level+2,varargin{:}), nl); + elseif(size(item,2)==1) + % Column vector, store only row indices. + txt=sprintf(dataformat,txt,padding0,'"_ArrayData_": ',... + matdata2json([ix,data],level+2,varargin{:}), nl); + else + % General case, store row and column indices. + txt=sprintf(dataformat,txt,padding0,'"_ArrayData_": ',... + matdata2json([ix,iy,data],level+2,varargin{:}), nl); + end +else + if(isreal(item)) + txt=sprintf(dataformat,txt,padding0,'"_ArrayData_": ',... + matdata2json(item(:)',level+2,varargin{:}), nl); + else + txt=sprintf(dataformat,txt,padding0,'"_ArrayIsComplex_": ','1', sep); + txt=sprintf(dataformat,txt,padding0,'"_ArrayData_": ',... + matdata2json([real(item(:)) imag(item(:))],level+2,varargin{:}), nl); + end +end +txt=sprintf('%s%s%s',txt,padding1,'}'); + +%%------------------------------------------------------------------------- +function txt=matdata2json(mat,level,varargin) + +ws=struct('tab',sprintf('\t'),'newline',sprintf('\n'),'sep',sprintf(',\n')); +ws=jsonopt('whitespaces_',ws,varargin{:}); +tab=ws.tab; +nl=ws.newline; + +if(size(mat,1)==1) + pre=''; + post=''; + level=level-1; +else + pre=sprintf('[%s',nl); + post=sprintf('%s%s]',nl,repmat(tab,1,level-1)); +end + +if(isempty(mat)) + txt='null'; + return; +end +floatformat=jsonopt('FloatFormat','%.10g',varargin{:}); +%if(numel(mat)>1) + formatstr=['[' repmat([floatformat ','],1,size(mat,2)-1) [floatformat sprintf('],%s',nl)]]; +%else +% formatstr=[repmat([floatformat ','],1,size(mat,2)-1) [floatformat sprintf(',\n')]]; +%end + +if(nargin>=2 && size(mat,1)>1 && jsonopt('ArrayIndent',1,varargin{:})==1) + formatstr=[repmat(tab,1,level) formatstr]; +end + +txt=sprintf(formatstr,mat'); +txt(end-length(nl):end)=[]; +if(islogical(mat) && jsonopt('ParseLogical',0,varargin{:})==1) + txt=regexprep(txt,'1','true'); + txt=regexprep(txt,'0','false'); +end +%txt=regexprep(mat2str(mat),'\s+',','); +%txt=regexprep(txt,';',sprintf('],\n[')); +% if(nargin>=2 && size(mat,1)>1) +% txt=regexprep(txt,'\[',[repmat(sprintf('\t'),1,level) '[']); +% end +txt=[pre txt post]; +if(any(isinf(mat(:)))) + txt=regexprep(txt,'([-+]*)Inf',jsonopt('Inf','"$1_Inf_"',varargin{:})); +end +if(any(isnan(mat(:)))) + txt=regexprep(txt,'NaN',jsonopt('NaN','"_NaN_"',varargin{:})); +end + +%%------------------------------------------------------------------------- +function newname=checkname(name,varargin) +isunpack=jsonopt('UnpackHex',1,varargin{:}); +newname=name; +if(isempty(regexp(name,'0x([0-9a-fA-F]+)_','once'))) + return +end +if(isunpack) + isoct=jsonopt('IsOctave',0,varargin{:}); + if(~isoct) + newname=regexprep(name,'(^x|_){1}0x([0-9a-fA-F]+)_','${native2unicode(hex2dec($2))}'); + else + pos=regexp(name,'(^x|_){1}0x([0-9a-fA-F]+)_','start'); + pend=regexp(name,'(^x|_){1}0x([0-9a-fA-F]+)_','end'); + if(isempty(pos)) return; end + str0=name; + pos0=[0 pend(:)' length(name)]; + newname=''; + for i=1:length(pos) + newname=[newname str0(pos0(i)+1:pos(i)-1) char(hex2dec(str0(pos(i)+3:pend(i)-1)))]; + end + if(pos(end)~=length(name)) + newname=[newname str0(pos0(end-1)+1:pos0(end))]; + end + end +end + +%%------------------------------------------------------------------------- +function newstr=escapejsonstring(str) +newstr=str; +isoct=exist('OCTAVE_VERSION','builtin'); +if(isoct) + vv=sscanf(OCTAVE_VERSION,'%f'); + if(vv(1)>=3.8) isoct=0; end +end +if(isoct) + escapechars={'\a','\f','\n','\r','\t','\v'}; + for i=1:length(escapechars); + newstr=regexprep(newstr,escapechars{i},escapechars{i}); + end +else + escapechars={'\a','\b','\f','\n','\r','\t','\v'}; + for i=1:length(escapechars); + newstr=regexprep(newstr,escapechars{i},regexprep(escapechars{i},'\\','\\\\')); + end +end diff --git a/machine-learning-ex7/ex7/lib/jsonlab/saveubjson.m b/machine-learning-ex7/ex7/lib/jsonlab/saveubjson.m new file mode 100644 index 0000000..eaec433 --- /dev/null +++ b/machine-learning-ex7/ex7/lib/jsonlab/saveubjson.m @@ -0,0 +1,504 @@ +function json=saveubjson(rootname,obj,varargin) +% +% json=saveubjson(rootname,obj,filename) +% or +% json=saveubjson(rootname,obj,opt) +% json=saveubjson(rootname,obj,'param1',value1,'param2',value2,...) +% +% convert a MATLAB object (cell, struct or array) into a Universal +% Binary JSON (UBJSON) binary string +% +% author: Qianqian Fang (fangq nmr.mgh.harvard.edu) +% created on 2013/08/17 +% +% $Id: saveubjson.m 460 2015-01-03 00:30:45Z fangq $ +% +% input: +% rootname: the name of the root-object, when set to '', the root name +% is ignored, however, when opt.ForceRootName is set to 1 (see below), +% the MATLAB variable name will be used as the root name. +% obj: a MATLAB object (array, cell, cell array, struct, struct array) +% filename: a string for the file name to save the output UBJSON data +% opt: a struct for additional options, ignore to use default values. +% opt can have the following fields (first in [.|.] is the default) +% +% opt.FileName [''|string]: a file name to save the output JSON data +% opt.ArrayToStruct[0|1]: when set to 0, saveubjson outputs 1D/2D +% array in JSON array format; if sets to 1, an +% array will be shown as a struct with fields +% "_ArrayType_", "_ArraySize_" and "_ArrayData_"; for +% sparse arrays, the non-zero elements will be +% saved to _ArrayData_ field in triplet-format i.e. +% (ix,iy,val) and "_ArrayIsSparse_" will be added +% with a value of 1; for a complex array, the +% _ArrayData_ array will include two columns +% (4 for sparse) to record the real and imaginary +% parts, and also "_ArrayIsComplex_":1 is added. +% opt.ParseLogical [1|0]: if this is set to 1, logical array elem +% will use true/false rather than 1/0. +% opt.NoRowBracket [1|0]: if this is set to 1, arrays with a single +% numerical element will be shown without a square +% bracket, unless it is the root object; if 0, square +% brackets are forced for any numerical arrays. +% opt.ForceRootName [0|1]: when set to 1 and rootname is empty, saveubjson +% will use the name of the passed obj variable as the +% root object name; if obj is an expression and +% does not have a name, 'root' will be used; if this +% is set to 0 and rootname is empty, the root level +% will be merged down to the lower level. +% opt.JSONP [''|string]: to generate a JSONP output (JSON with padding), +% for example, if opt.JSON='foo', the JSON data is +% wrapped inside a function call as 'foo(...);' +% opt.UnpackHex [1|0]: conver the 0x[hex code] output by loadjson +% back to the string form +% +% opt can be replaced by a list of ('param',value) pairs. The param +% string is equivallent to a field in opt and is case sensitive. +% output: +% json: a binary string in the UBJSON format (see http://ubjson.org) +% +% examples: +% jsonmesh=struct('MeshNode',[0 0 0;1 0 0;0 1 0;1 1 0;0 0 1;1 0 1;0 1 1;1 1 1],... +% 'MeshTetra',[1 2 4 8;1 3 4 8;1 2 6 8;1 5 6 8;1 5 7 8;1 3 7 8],... +% 'MeshTri',[1 2 4;1 2 6;1 3 4;1 3 7;1 5 6;1 5 7;... +% 2 8 4;2 8 6;3 8 4;3 8 7;5 8 6;5 8 7],... +% 'MeshCreator','FangQ','MeshTitle','T6 Cube',... +% 'SpecialData',[nan, inf, -inf]); +% saveubjson('jsonmesh',jsonmesh) +% saveubjson('jsonmesh',jsonmesh,'meshdata.ubj') +% +% license: +% BSD, see LICENSE_BSD.txt files for details +% +% -- this function is part of JSONLab toolbox (http://iso2mesh.sf.net/cgi-bin/index.cgi?jsonlab) +% + +if(nargin==1) + varname=inputname(1); + obj=rootname; + if(isempty(varname)) + varname='root'; + end + rootname=varname; +else + varname=inputname(2); +end +if(length(varargin)==1 && ischar(varargin{1})) + opt=struct('FileName',varargin{1}); +else + opt=varargin2struct(varargin{:}); +end +opt.IsOctave=exist('OCTAVE_VERSION','builtin'); +rootisarray=0; +rootlevel=1; +forceroot=jsonopt('ForceRootName',0,opt); +if((isnumeric(obj) || islogical(obj) || ischar(obj) || isstruct(obj) || iscell(obj)) && isempty(rootname) && forceroot==0) + rootisarray=1; + rootlevel=0; +else + if(isempty(rootname)) + rootname=varname; + end +end +if((isstruct(obj) || iscell(obj))&& isempty(rootname) && forceroot) + rootname='root'; +end +json=obj2ubjson(rootname,obj,rootlevel,opt); +if(~rootisarray) + json=['{' json '}']; +end + +jsonp=jsonopt('JSONP','',opt); +if(~isempty(jsonp)) + json=[jsonp '(' json ')']; +end + +% save to a file if FileName is set, suggested by Patrick Rapin +if(~isempty(jsonopt('FileName','',opt))) + fid = fopen(opt.FileName, 'wb'); + fwrite(fid,json); + fclose(fid); +end + +%%------------------------------------------------------------------------- +function txt=obj2ubjson(name,item,level,varargin) + +if(iscell(item)) + txt=cell2ubjson(name,item,level,varargin{:}); +elseif(isstruct(item)) + txt=struct2ubjson(name,item,level,varargin{:}); +elseif(ischar(item)) + txt=str2ubjson(name,item,level,varargin{:}); +else + txt=mat2ubjson(name,item,level,varargin{:}); +end + +%%------------------------------------------------------------------------- +function txt=cell2ubjson(name,item,level,varargin) +txt=''; +if(~iscell(item)) + error('input is not a cell'); +end + +dim=size(item); +if(ndims(squeeze(item))>2) % for 3D or higher dimensions, flatten to 2D for now + item=reshape(item,dim(1),numel(item)/dim(1)); + dim=size(item); +end +len=numel(item); % let's handle 1D cell first +if(len>1) + if(~isempty(name)) + txt=[S_(checkname(name,varargin{:})) '[']; name=''; + else + txt='['; + end +elseif(len==0) + if(~isempty(name)) + txt=[S_(checkname(name,varargin{:})) 'Z']; name=''; + else + txt='Z'; + end +end +for j=1:dim(2) + if(dim(1)>1) txt=[txt '[']; end + for i=1:dim(1) + txt=[txt obj2ubjson(name,item{i,j},level+(len>1),varargin{:})]; + end + if(dim(1)>1) txt=[txt ']']; end +end +if(len>1) txt=[txt ']']; end + +%%------------------------------------------------------------------------- +function txt=struct2ubjson(name,item,level,varargin) +txt=''; +if(~isstruct(item)) + error('input is not a struct'); +end +dim=size(item); +if(ndims(squeeze(item))>2) % for 3D or higher dimensions, flatten to 2D for now + item=reshape(item,dim(1),numel(item)/dim(1)); + dim=size(item); +end +len=numel(item); + +if(~isempty(name)) + if(len>1) txt=[S_(checkname(name,varargin{:})) '[']; end +else + if(len>1) txt='['; end +end +for j=1:dim(2) + if(dim(1)>1) txt=[txt '[']; end + for i=1:dim(1) + names = fieldnames(item(i,j)); + if(~isempty(name) && len==1) + txt=[txt S_(checkname(name,varargin{:})) '{']; + else + txt=[txt '{']; + end + if(~isempty(names)) + for e=1:length(names) + txt=[txt obj2ubjson(names{e},getfield(item(i,j),... + names{e}),level+(dim(1)>1)+1+(len>1),varargin{:})]; + end + end + txt=[txt '}']; + end + if(dim(1)>1) txt=[txt ']']; end +end +if(len>1) txt=[txt ']']; end + +%%------------------------------------------------------------------------- +function txt=str2ubjson(name,item,level,varargin) +txt=''; +if(~ischar(item)) + error('input is not a string'); +end +item=reshape(item, max(size(item),[1 0])); +len=size(item,1); + +if(~isempty(name)) + if(len>1) txt=[S_(checkname(name,varargin{:})) '[']; end +else + if(len>1) txt='['; end +end +isoct=jsonopt('IsOctave',0,varargin{:}); +for e=1:len + val=item(e,:); + if(len==1) + obj=['' S_(checkname(name,varargin{:})) '' '',S_(val),'']; + if(isempty(name)) obj=['',S_(val),'']; end + txt=[txt,'',obj]; + else + txt=[txt,'',['',S_(val),'']]; + end +end +if(len>1) txt=[txt ']']; end + +%%------------------------------------------------------------------------- +function txt=mat2ubjson(name,item,level,varargin) +if(~isnumeric(item) && ~islogical(item)) + error('input is not an array'); +end + +if(length(size(item))>2 || issparse(item) || ~isreal(item) || ... + isempty(item) || jsonopt('ArrayToStruct',0,varargin{:})) + cid=I_(uint32(max(size(item)))); + if(isempty(name)) + txt=['{' S_('_ArrayType_'),S_(class(item)),S_('_ArraySize_'),I_a(size(item),cid(1)) ]; + else + if(isempty(item)) + txt=[S_(checkname(name,varargin{:})),'Z']; + return; + else + txt=[S_(checkname(name,varargin{:})),'{',S_('_ArrayType_'),S_(class(item)),S_('_ArraySize_'),I_a(size(item),cid(1))]; + end + end +else + if(isempty(name)) + txt=matdata2ubjson(item,level+1,varargin{:}); + else + if(numel(item)==1 && jsonopt('NoRowBracket',1,varargin{:})==1) + numtxt=regexprep(regexprep(matdata2ubjson(item,level+1,varargin{:}),'^\[',''),']',''); + txt=[S_(checkname(name,varargin{:})) numtxt]; + else + txt=[S_(checkname(name,varargin{:})),matdata2ubjson(item,level+1,varargin{:})]; + end + end + return; +end +if(issparse(item)) + [ix,iy]=find(item); + data=full(item(find(item))); + if(~isreal(item)) + data=[real(data(:)),imag(data(:))]; + if(size(item,1)==1) + % Kludge to have data's 'transposedness' match item's. + % (Necessary for complex row vector handling below.) + data=data'; + end + txt=[txt,S_('_ArrayIsComplex_'),'T']; + end + txt=[txt,S_('_ArrayIsSparse_'),'T']; + if(size(item,1)==1) + % Row vector, store only column indices. + txt=[txt,S_('_ArrayData_'),... + matdata2ubjson([iy(:),data'],level+2,varargin{:})]; + elseif(size(item,2)==1) + % Column vector, store only row indices. + txt=[txt,S_('_ArrayData_'),... + matdata2ubjson([ix,data],level+2,varargin{:})]; + else + % General case, store row and column indices. + txt=[txt,S_('_ArrayData_'),... + matdata2ubjson([ix,iy,data],level+2,varargin{:})]; + end +else + if(isreal(item)) + txt=[txt,S_('_ArrayData_'),... + matdata2ubjson(item(:)',level+2,varargin{:})]; + else + txt=[txt,S_('_ArrayIsComplex_'),'T']; + txt=[txt,S_('_ArrayData_'),... + matdata2ubjson([real(item(:)) imag(item(:))],level+2,varargin{:})]; + end +end +txt=[txt,'}']; + +%%------------------------------------------------------------------------- +function txt=matdata2ubjson(mat,level,varargin) +if(isempty(mat)) + txt='Z'; + return; +end +if(size(mat,1)==1) + level=level-1; +end +type=''; +hasnegtive=(mat<0); +if(isa(mat,'integer') || isinteger(mat) || (isfloat(mat) && all(mod(mat(:),1) == 0))) + if(isempty(hasnegtive)) + if(max(mat(:))<=2^8) + type='U'; + end + end + if(isempty(type)) + % todo - need to consider negative ones separately + id= histc(abs(max(mat(:))),[0 2^7 2^15 2^31 2^63]); + if(isempty(find(id))) + error('high-precision data is not yet supported'); + end + key='iIlL'; + type=key(find(id)); + end + txt=[I_a(mat(:),type,size(mat))]; +elseif(islogical(mat)) + logicalval='FT'; + if(numel(mat)==1) + txt=logicalval(mat+1); + else + txt=['[$U#' I_a(size(mat),'l') typecast(swapbytes(uint8(mat(:)')),'uint8')]; + end +else + if(numel(mat)==1) + txt=['[' D_(mat) ']']; + else + txt=D_a(mat(:),'D',size(mat)); + end +end + +%txt=regexprep(mat2str(mat),'\s+',','); +%txt=regexprep(txt,';',sprintf('],[')); +% if(nargin>=2 && size(mat,1)>1) +% txt=regexprep(txt,'\[',[repmat(sprintf('\t'),1,level) '[']); +% end +if(any(isinf(mat(:)))) + txt=regexprep(txt,'([-+]*)Inf',jsonopt('Inf','"$1_Inf_"',varargin{:})); +end +if(any(isnan(mat(:)))) + txt=regexprep(txt,'NaN',jsonopt('NaN','"_NaN_"',varargin{:})); +end + +%%------------------------------------------------------------------------- +function newname=checkname(name,varargin) +isunpack=jsonopt('UnpackHex',1,varargin{:}); +newname=name; +if(isempty(regexp(name,'0x([0-9a-fA-F]+)_','once'))) + return +end +if(isunpack) + isoct=jsonopt('IsOctave',0,varargin{:}); + if(~isoct) + newname=regexprep(name,'(^x|_){1}0x([0-9a-fA-F]+)_','${native2unicode(hex2dec($2))}'); + else + pos=regexp(name,'(^x|_){1}0x([0-9a-fA-F]+)_','start'); + pend=regexp(name,'(^x|_){1}0x([0-9a-fA-F]+)_','end'); + if(isempty(pos)) return; end + str0=name; + pos0=[0 pend(:)' length(name)]; + newname=''; + for i=1:length(pos) + newname=[newname str0(pos0(i)+1:pos(i)-1) char(hex2dec(str0(pos(i)+3:pend(i)-1)))]; + end + if(pos(end)~=length(name)) + newname=[newname str0(pos0(end-1)+1:pos0(end))]; + end + end +end +%%------------------------------------------------------------------------- +function val=S_(str) +if(length(str)==1) + val=['C' str]; +else + val=['S' I_(int32(length(str))) str]; +end +%%------------------------------------------------------------------------- +function val=I_(num) +if(~isinteger(num)) + error('input is not an integer'); +end +if(num>=0 && num<255) + val=['U' data2byte(swapbytes(cast(num,'uint8')),'uint8')]; + return; +end +key='iIlL'; +cid={'int8','int16','int32','int64'}; +for i=1:4 + if((num>0 && num<2^(i*8-1)) || (num<0 && num>=-2^(i*8-1))) + val=[key(i) data2byte(swapbytes(cast(num,cid{i})),'uint8')]; + return; + end +end +error('unsupported integer'); + +%%------------------------------------------------------------------------- +function val=D_(num) +if(~isfloat(num)) + error('input is not a float'); +end + +if(isa(num,'single')) + val=['d' data2byte(num,'uint8')]; +else + val=['D' data2byte(num,'uint8')]; +end +%%------------------------------------------------------------------------- +function data=I_a(num,type,dim,format) +id=find(ismember('iUIlL',type)); + +if(id==0) + error('unsupported integer array'); +end + +% based on UBJSON specs, all integer types are stored in big endian format + +if(id==1) + data=data2byte(swapbytes(int8(num)),'uint8'); + blen=1; +elseif(id==2) + data=data2byte(swapbytes(uint8(num)),'uint8'); + blen=1; +elseif(id==3) + data=data2byte(swapbytes(int16(num)),'uint8'); + blen=2; +elseif(id==4) + data=data2byte(swapbytes(int32(num)),'uint8'); + blen=4; +elseif(id==5) + data=data2byte(swapbytes(int64(num)),'uint8'); + blen=8; +end + +if(nargin>=3 && length(dim)>=2 && prod(dim)~=dim(2)) + format='opt'; +end +if((nargin<4 || strcmp(format,'opt')) && numel(num)>1) + if(nargin>=3 && (length(dim)==1 || (length(dim)>=2 && prod(dim)~=dim(2)))) + cid=I_(uint32(max(dim))); + data=['$' type '#' I_a(dim,cid(1)) data(:)']; + else + data=['$' type '#' I_(int32(numel(data)/blen)) data(:)']; + end + data=['[' data(:)']; +else + data=reshape(data,blen,numel(data)/blen); + data(2:blen+1,:)=data; + data(1,:)=type; + data=data(:)'; + data=['[' data(:)' ']']; +end +%%------------------------------------------------------------------------- +function data=D_a(num,type,dim,format) +id=find(ismember('dD',type)); + +if(id==0) + error('unsupported float array'); +end + +if(id==1) + data=data2byte(single(num),'uint8'); +elseif(id==2) + data=data2byte(double(num),'uint8'); +end + +if(nargin>=3 && length(dim)>=2 && prod(dim)~=dim(2)) + format='opt'; +end +if((nargin<4 || strcmp(format,'opt')) && numel(num)>1) + if(nargin>=3 && (length(dim)==1 || (length(dim)>=2 && prod(dim)~=dim(2)))) + cid=I_(uint32(max(dim))); + data=['$' type '#' I_a(dim,cid(1)) data(:)']; + else + data=['$' type '#' I_(int32(numel(data)/(id*4))) data(:)']; + end + data=['[' data]; +else + data=reshape(data,(id*4),length(data)/(id*4)); + data(2:(id*4+1),:)=data; + data(1,:)=type; + data=data(:)'; + data=['[' data(:)' ']']; +end +%%------------------------------------------------------------------------- +function bytes=data2byte(varargin) +bytes=typecast(varargin{:}); +bytes=bytes(:)'; diff --git a/machine-learning-ex7/ex7/lib/jsonlab/varargin2struct.m b/machine-learning-ex7/ex7/lib/jsonlab/varargin2struct.m new file mode 100644 index 0000000..9a5c2b6 --- /dev/null +++ b/machine-learning-ex7/ex7/lib/jsonlab/varargin2struct.m @@ -0,0 +1,40 @@ +function opt=varargin2struct(varargin) +% +% opt=varargin2struct('param1',value1,'param2',value2,...) +% or +% opt=varargin2struct(...,optstruct,...) +% +% convert a series of input parameters into a structure +% +% authors:Qianqian Fang (fangq nmr.mgh.harvard.edu) +% date: 2012/12/22 +% +% input: +% 'param', value: the input parameters should be pairs of a string and a value +% optstruct: if a parameter is a struct, the fields will be merged to the output struct +% +% output: +% opt: a struct where opt.param1=value1, opt.param2=value2 ... +% +% license: +% BSD, see LICENSE_BSD.txt files for details +% +% -- this function is part of jsonlab toolbox (http://iso2mesh.sf.net/cgi-bin/index.cgi?jsonlab) +% + +len=length(varargin); +opt=struct; +if(len==0) return; end +i=1; +while(i<=len) + if(isstruct(varargin{i})) + opt=mergestruct(opt,varargin{i}); + elseif(ischar(varargin{i}) && i