add histogram functionality; recompile

scratchpad/plotting/c/attribution.md

@@ -0,0 +1,2 @@
+https://github.com/red-data-tools/YouPlot
+Design inspired by ^

scratchpad/plotting/c/example.md

@@ -0,0 +1 @@
+./samples  | hist 200 | head -n 20

scratchpad/plotting/c/histogram.c

@@ -0,0 +1,86 @@
+#include <stdio.h>
+#include <stdlib.h>
+#include <float.h>
+
+#define MAX_SAMPLES 1000000
+
+int main(int argc, char *argv[]) {
+    if (argc < 2) {
+        fprintf(stderr, "Usage: %s number_of_bins\n", argv[0]);
+        exit(EXIT_FAILURE);
+    }
+
+    int num_bins = atoi(argv[1]);
+    if (num_bins <= 0) {
+        fprintf(stderr, "Number of bins must be a positive integer.\n");
+        exit(EXIT_FAILURE);
+    }
+
+    int *bins = calloc(num_bins, sizeof(int));
+    double *samples = malloc(MAX_SAMPLES * sizeof(double));
+    if (bins == NULL || samples == NULL) {
+        fprintf(stderr, "Memory allocation failed.\n");
+        exit(EXIT_FAILURE);
+    }
+
+    double value, min_value = DBL_MAX, max_value = -DBL_MAX;
+    int sample_count = 0;
+
+    // Read numbers from stdin and store them into the samples array
+    while (sample_count < MAX_SAMPLES && scanf("%lf", &value) != EOF) {
+        samples[sample_count++] = value;
+        if (value < min_value) {
+            min_value = value;
+        }
+        if (value > max_value) {
+            max_value = value;
+        }
+    }
+
+    // Avoid division by zero for a single unique value
+    if (min_value == max_value) {
+        max_value++;
+    }
+
+    // Calculate bin width
+    double range = max_value - min_value;
+    double bin_width = range / num_bins;
+
+    // Fill the bins with sample counts
+    for (int i = 0; i < sample_count; i++) {
+        int bin_index = (int)((samples[i] - min_value) / bin_width);
+        if (bin_index == num_bins) {
+            bin_index--; // Last bin includes max_value
+        }
+        bins[bin_index]++;
+    }
+
+    // Calculate the scaling factor based on the maximum bin count
+    int max_bin_count = 0;
+    for (int i = 0; i < num_bins; i++) {
+        if (bins[i] > max_bin_count) {
+            max_bin_count = bins[i];
+        }
+    }
+    const int MAX_WIDTH = 50; // Adjust this to your terminal width
+    double scale = max_bin_count > MAX_WIDTH ? (double)MAX_WIDTH / max_bin_count : 1.0;
+
+    // Print the histogram
+    for (int i = 0; i < num_bins; i++) {
+        double bin_start = min_value + i * bin_width;
+        double bin_end = bin_start + bin_width;
+        printf("  [%4.1f, %4.1f): ", bin_start, bin_end);
+
+        int marks = (int)(bins[i] * scale);
+        for (int j = 0; j < marks; j++) {
+            printf("▇");
+        }
+        printf(" %d\n", bins[i]);
+    }
+
+    // Free the allocated memory
+    free(bins);
+    free(samples);
+
+    return 0;
+}

scratchpad/plotting/gnuplot/term4.gp

@@ -1,9 +1,11 @@
 reset
 
-stats 'data.dat' nooutput
+# stats '-' nooutput
 # show variables all
-max=STATS_max 
+# max=STATS_max 
-min=STATS_min 
+# min=STATS_min 
+min=-1
+max=25
 
 n=1000 
 width=(max-min)/n 
@@ -23,4 +25,6 @@ set tics scale 0,0,0,0
 set xlabel "x"
 
 set terminal dumb size 100, 25
-plot 'data.dat' u (hist($1,width)):(1.0) smooth freq w boxes notitle
+plot '-' u (hist($1,width)):(1.0) smooth freq w boxes notitle
+
+

scratchpad/plotting/gnuplot/term5.gp

@@ -0,0 +1,5 @@
+reset
+set terminal dumb size 100, 25
+set autoscale
+plot '<cat' using 0 with lines notitle
+

scratchpad/plotting/gnuplot/term6.gp

@@ -0,0 +1,30 @@
+reset
+
+# stats '-' nooutput
+# show variables all
+# max=STATS_max 
+# min=STATS_min 
+min=-1
+max=25
+
+n=1000 
+width=(max-min)/n 
+hist(x,width)=width*floor(x/width)+width/2.0
+
+set xrange [min:max]
+set yrange [0:45000]
+
+unset xtics
+unset ytics
+unset border
+set xtics out nomirror norotate offset 0,0
+set ytics out nomirror norotate
+set ytics in nomirror norotate offset 0,0
+set tics scale 0,0,0,0
+
+set xlabel "x"
+
+set terminal dumb size 100, 25
+plot '<cat' u (hist(0,width)):(1.0) smooth freq w boxes notitle
+
+

scratchpad/plotting/uplot/command.sh

@@ -0,0 +1,2 @@
+cat data.dat | uplot hist --nbins 100
+cat data.dat | uplot hist --nbins 10

scratchpad/plotting/uplot/howto.md

@@ -0,0 +1,4 @@
+sudo apt install ruby-dev 
+
+sudo gem install youplot
+

squiggle_more.c

@@ -213,6 +213,81 @@ void array_print_stats(double xs[], int n){
            mean, std, ci_90.low, ci_80.low, ci_50.low, median, ci_50.high, ci_80.high, ci_90.high);
 }
 
+
+void print_histogram(double* xs, int n_samples, int n_bins) {
+    // Generated with the help of an llm; there might be subtle off-by-one errors
+    // interface inspired by <https://github.com/red-data-tools/YouPlot>
+    if (n_bins <= 0) {
+        fprintf(stderr, "Number of bins must be a positive integer.\n");
+        return;
+    } else if (n_samples <= 0) {
+        fprintf(stderr, "Number of samples must be a positive integer.\n");
+        return;
+    }
+
+    int *bins = (int*) malloc((size_t)n_bins * sizeof(int));
+    if (bins == NULL) {
+        fprintf(stderr, "Memory allocation for bins failed.\n");
+        return;
+    }
+
+    double min_value = xs[0], max_value = xs[0];
+
+    // Find the minimum and maximum values from the samples
+    for (int i = 0; i < n_samples; i++) {
+        if (xs[i] < min_value) {
+            min_value = xs[i];
+        }
+        if (xs[i] > max_value) {
+            max_value = xs[i];
+        }
+    }
+
+    // Avoid division by zero for a single unique value
+    if (min_value == max_value) {
+        max_value++;
+    }
+
+    // Calculate bin width
+    double range = max_value - min_value;
+    double bin_width = range / n_bins;
+
+    // Fill the bins with sample counts
+    for (int i = 0; i < n_samples; i++) {
+        int bin_index = (int)((xs[i] - min_value) / bin_width);
+        if (bin_index == n_bins) {
+            bin_index--; // Last bin includes max_value
+        }
+        bins[bin_index]++;
+    }
+
+    // Calculate the scaling factor based on the maximum bin count
+    int max_bin_count = 0;
+    for (int i = 0; i < n_bins; i++) {
+        if (bins[i] > max_bin_count) {
+            max_bin_count = bins[i];
+        }
+    }
+    const int MAX_WIDTH = 50; // Adjust this to your terminal width
+    double scale = max_bin_count > MAX_WIDTH ? (double)MAX_WIDTH / max_bin_count : 1.0;
+
+    // Print the histogram
+    for (int i = 0; i < n_bins; i++) {
+        double bin_start = min_value + i * bin_width;
+        double bin_end = bin_start + bin_width;
+        printf("  [%4.1f, %4.1f): ", bin_start, bin_end);
+
+        int marks = (int)(bins[i] * scale);
+        for (int j = 0; j < marks; j++) {
+            printf("█");
+        }
+        printf(" %d\n", bins[i]);
+    }
+
+    // Free the allocated memory for bins
+    free(bins);
+}
+
 // Replicate some of the above functions over samplers
 // However, in the future I'll delete this
 // There should be a clear boundary between working with samplers and working with an array of samples

squiggle_more.h

@@ -13,6 +13,7 @@ typedef struct ci_t {
 ci array_get_ci(ci interval, double* xs, int n);
 ci array_get_90_ci(double xs[], int n);
 void array_print_stats(double xs[], int n);
+void print_histogram(double* xs, int n_samples, int n_bins);
 
 // Deprecated: get confidence intervals directly from samplers
 ci sampler_get_ci(ci interval, double (*sampler)(uint64_t*), int n, uint64_t* seed);