Added new aggregate rules (first, nth, last) (#82)

This commit introduces several new aggregate rules (first, nth, last) in JSON, PHP, YML schema files, and creates corresponding classes. These additions enhance rule validation, allowing further specification of conditions based on first, nth, and last values in a column. The changes include a range of rule adjustments such as number expectations, string expectations, and "not expected" conditions, broadening the flexibility and functionality of the schema validation system.
JBZoo · Mar 24, 2024 · 409156d · 409156d
1 parent 3f0225c
commit 409156d
Show file tree

Hide file tree

Showing 26 changed files with 885 additions and 15 deletions.
diff --git a/README.md b/README.md
@@ -4,7 +4,7 @@
 [![Stable Version](https://poser.pugx.org/jbzoo/csv-blueprint/version)](https://packagist.org/packages/jbzoo/csv-blueprint/)    [![Total Downloads](https://poser.pugx.org/jbzoo/csv-blueprint/downloads)](https://packagist.org/packages/jbzoo/csv-blueprint/stats)    [![Docker Pulls](https://img.shields.io/docker/pulls/jbzoo/csv-blueprint.svg)](https://hub.docker.com/r/jbzoo/csv-blueprint/tags)    [![GitHub License](https://img.shields.io/github/license/jbzoo/csv-blueprint)](https://github.com/JBZoo/Csv-Blueprint/blob/master/LICENSE)
 
 <!-- rules-counter -->
-[![Static Badge](https://img.shields.io/badge/Rules-103-green?label=Total%20Number%20of%20Rules&labelColor=darkgreen&color=gray)](schema-examples/full.yml)    [![Static Badge](https://img.shields.io/badge/Rules-55-green?label=Cell%20Value&labelColor=blue&color=gray)](src/Rules/Cell)    [![Static Badge](https://img.shields.io/badge/Rules-45-green?label=Aggregate%20Column&labelColor=blue&color=gray)](src/Rules/Aggregate)    [![Static Badge](https://img.shields.io/badge/Rules-3-green?label=Extra%20Checks&labelColor=blue&color=gray)](#extra-checks)    [![Static Badge](https://img.shields.io/badge/Rules-329-green?label=Plan%20to%20add&labelColor=gray&color=gray)](tests/schemas/todo.yml)
+[![Static Badge](https://img.shields.io/badge/Rules-119-green?label=Total%20Number%20of%20Rules&labelColor=darkgreen&color=gray)](schema-examples/full.yml)    [![Static Badge](https://img.shields.io/badge/Rules-55-green?label=Cell%20Value&labelColor=blue&color=gray)](src/Rules/Cell)    [![Static Badge](https://img.shields.io/badge/Rules-59-green?label=Aggregate%20Column&labelColor=blue&color=gray)](src/Rules/Aggregate)    [![Static Badge](https://img.shields.io/badge/Rules-5-green?label=Extra%20Checks&labelColor=blue&color=gray)](#extra-checks)    [![Static Badge](https://img.shields.io/badge/Rules-317-green?label=Plan%20to%20add&labelColor=gray&color=gray)](tests/schemas/todo.yml)
 <!-- /rules-counter -->
 
 ## Introduction
@@ -265,6 +265,26 @@ columns:
     aggregate_rules:
       is_unique: true                   # All values in the column are unique.
 
+      # First number in the column. Expected value is float or integer.
+      first_num: 5
+      first_num_not: 4.123
+      first_num_min: -1
+      first_num_max: 2e4
+      first: Expected                   # First value in the column. Will be compared as strings.
+      first_not: 'Not Expected'         # Not allowed as the first value in the column. Will be compared as strings.
+
+      # N-th in the column.
+      nth: [ 2, Expected ]              # Nth value in the column. Will be compared as strings.
+      nth_not: [ 2, 'Not expected' ]    # Not allowed as the N-th value in the column. Will be compared as strings.
+
+      # Last number in the column. Expected value is float or integer.
+      last_num: 5
+      last_num_not: 4.123
+      last_num_min: -1
+      last_num_max: 2e4
+      last: Expected                    # Last value in the column. Will be compared as strings.
+      last_not: 'Not Expected'          # Not allowed as the last value in the column. Will be compared as strings.
+
       # Sum of the numbers in the column. Example: [1, 2, 3] => 6.
       sum: 5.123
       sum_not: 4.123
@@ -362,7 +382,9 @@ Behind the scenes to what is outlined in the yml above, there are additional che
 
 * With `filename_pattern` rule, you can check if the file name matches the pattern.
 * Property `name` is not defined in a column. If `csv.header: true`.
-* Schema contains an unknown column `name` that is not found in the CSV file. If `csv.header: true`
+* Check that each row matches the number of columns.
+* If `csv.header: true`. Schema contains an unknown column `name` that is not found in the CSV file.
+* If `csv.header: false`. Compare the number of columns in the schema and the CSV file.
 
 <!-- /extra-rules -->
 
@@ -381,11 +403,11 @@ You can find launch examples in the [workflow demo](https://github.com/JBZoo/Csv
   with:
     # Path(s) to validate. You can specify path in which CSV files will be searched. Feel free to use glob pattrens. Usage examples: /full/path/file.csv, p/file.csv, p/*.csv, p/**/*.csv, p/**/name-*.csv, **/*.csv, etc.
     # Required: true
-    csv: ./tests/**/*.csv
+    csv: './tests/**/*.csv'
 
     # Schema filepath. It can be a YAML, JSON or PHP. See examples on GitHub.
     # Required: true
-    schema: ./tests/schema.yml
+    schema: './tests/**/*.yml'
 
     # Report format. Available options: text, table, github, gitlab, teamcity, junit.
     # Default value: table

diff --git a/schema-examples/full.json b/schema-examples/full.json
@@ -87,6 +87,23 @@
             "aggregate_rules" : {
                 "is_unique"               : true,
 
+                "first_num"               : 5,
+                "first_num_not"           : 4.123,
+                "first_num_min"           : -1,
+                "first_num_max"           : 20000,
+                "first"                   : "Expected",
+                "first_not"               : "Not Expected",
+
+                "nth"                     : [2, "Expected"],
+                "nth_not"                 : [2, "Not expected"],
+
+                "last_num"                : 5,
+                "last_num_not"            : 4.123,
+                "last_num_min"            : -1,
+                "last_num_max"            : 20000,
+                "last"                    : "Expected",
+                "last_not"                : "Not Expected",
+
                 "sum"                     : 5.123,
                 "sum_not"                 : 4.123,
                 "sum_min"                 : 1.123,

diff --git a/schema-examples/full.php b/schema-examples/full.php
@@ -108,6 +108,23 @@
             'aggregate_rules' => [
                 'is_unique' => true,
 
+                'first_num'     => 5,
+                'first_num_not' => 4.123,
+                'first_num_min' => -1,
+                'first_num_max' => 20000.0,
+                'first'         => 'Expected',
+                'first_not'     => 'Not Expected',
+
+                'nth'     => [2, 'Expected'],
+                'nth_not' => [2, 'Not expected'],
+
+                'last_num'     => 5,
+                'last_num_not' => 4.123,
+                'last_num_min' => -1,
+                'last_num_max' => 20000.0,
+                'last'         => 'Expected',
+                'last_not'     => 'Not Expected',
+
                 'sum'     => 5.123,
                 'sum_not' => 4.123,
                 'sum_min' => 1.123,

diff --git a/schema-examples/full.yml b/schema-examples/full.yml
@@ -179,6 +179,26 @@ columns:
     aggregate_rules:
       is_unique: true                   # All values in the column are unique.
 
+      # First number in the column. Expected value is float or integer.
+      first_num: 5
+      first_num_not: 4.123
+      first_num_min: -1
+      first_num_max: 2e4
+      first: Expected                   # First value in the column. Will be compared as strings.
+      first_not: 'Not Expected'         # Not allowed as the first value in the column. Will be compared as strings.
+
+      # N-th in the column.
+      nth: [ 2, Expected ]              # Nth value in the column. Will be compared as strings.
+      nth_not: [ 2, 'Not expected' ]    # Not allowed as the N-th value in the column. Will be compared as strings.
+
+      # Last number in the column. Expected value is float or integer.
+      last_num: 5
+      last_num_not: 4.123
+      last_num_min: -1
+      last_num_max: 2e4
+      last: Expected                    # Last value in the column. Will be compared as strings.
+      last_not: 'Not Expected'          # Not allowed as the last value in the column. Will be compared as strings.
+
       # Sum of the numbers in the column. Example: [1, 2, 3] => 6.
       sum: 5.123
       sum_not: 4.123

diff --git a/schema-examples/full_clean.yml b/schema-examples/full_clean.yml
@@ -94,46 +94,74 @@ columns:
 
     aggregate_rules:
       is_unique: true
+
+      first_num: 5
+      first_num_not: 4.123
+      first_num_min: -1
+      first_num_max: 20000.0
+      first: Expected
+      first_not: 'Not Expected'
+
+      nth: [ 2, Expected ]
+      nth_not: [ 2, 'Not expected' ]
+
+      last_num: 5
+      last_num_not: 4.123
+      last_num_min: -1
+      last_num_max: 20000.0
+      last: Expected
+      last_not: 'Not Expected'
+
       sum: 5.123
       sum_not: 4.123
       sum_min: 1.123
       sum_max: 10.123
+
       average: 5.123
       average_not: 4.123
       average_min: 1.123
       average_max: 10.123
+
       count: 5
       count_not: 4
       count_min: 1
       count_max: 10
+
       count_empty: 5
       count_empty_not: 4
       count_empty_min: 1
       count_empty_max: 10
+
       count_not_empty: 5
       count_not_empty_not: 4
       count_not_empty_min: 1
       count_not_empty_max: 10
+
       median: 5.123
       median_not: 4.123
       median_min: 1.123
       median_max: 10.123
+
       population_variance: 5.123
       population_variance_not: 4.123
       population_variance_min: 1.123
       population_variance_max: 10.123
+
       sample_variance: 5.123
       sample_variance_not: 4.123
       sample_variance_min: 1.123
       sample_variance_max: 10.123
+
       stddev: 5.123
       stddev_not: 4.123
       stddev_min: 1.123
       stddev_max: 10.123
+
       stddev_pop: 5.123
       stddev_pop_not: 4.123
       stddev_pop_min: 1.123
       stddev_pop_max: 10.123
+
       coef_of_var: 5.123
       coef_of_var_not: 4.123
       coef_of_var_min: 1.123

diff --git a/src/Rules/AbstarctRule.php b/src/Rules/AbstarctRule.php
@@ -26,11 +26,11 @@ abstract class AbstarctRule
 {
     public const INPUT_TYPE = self::INPUT_TYPE_UNDEF;
 
-    public const INPUT_TYPE_UNDEF   = self::INPUT_TYPE_STRINGS;
     public const INPUT_TYPE_BOOL    = 0;
     public const INPUT_TYPE_INTS    = 1;
     public const INPUT_TYPE_FLOATS  = 2;
     public const INPUT_TYPE_STRINGS = 3;
+    public const INPUT_TYPE_UNDEF   = 4;
 
     // Modes
     public const DEFAULT = 'default';
@@ -213,9 +213,6 @@ protected function getOptionAsFloat(): float
         return (float)$this->options;
     }
 
-    /**
-     * @return string[]
-     */
     protected function getOptionAsArray(): array
     {
         // TODO: Replace to warning message

diff --git a/src/Rules/Aggregate/ComboFirstNum.php b/src/Rules/Aggregate/ComboFirstNum.php
@@ -0,0 +1,44 @@
+<?php
+
+/**
+ * JBZoo Toolbox - Csv-Blueprint.
+ *
+ * This file is part of the JBZoo Toolbox project.
+ * For the full copyright and license information, please view the LICENSE
+ * file that was distributed with this source code.
+ *
+ * @license    MIT
+ * @copyright  Copyright (C) JBZoo.com, All rights reserved.
+ * @see        https://github.com/JBZoo/Csv-Blueprint
+ */
+
+declare(strict_types=1);
+
+namespace JBZoo\CsvBlueprint\Rules\Aggregate;
+
+use JBZoo\CsvBlueprint\Rules\AbstarctRule;
+
+final class ComboFirstNum extends AbstarctAggregateRuleCombo
+{
+    public const INPUT_TYPE = AbstarctRule::INPUT_TYPE_FLOATS;
+
+    protected const NAME = 'first value';
+
+    protected const HELP_TOP = ['First number in the column. Expected value is float or integer.'];
+
+    protected const HELP_OPTIONS = [
+        self::EQ  => ['5', ''],
+        self::NOT => ['4.123', ''],
+        self::MIN => ['-1', ''],
+        self::MAX => ['2e4', ''],
+    ];
+
+    protected function getActualAggregate(array $colValues): ?float
+    {
+        if (!isset($colValues[0])) {
+            return null;
+        }
+
+        return (float)$colValues[0];
+    }
+}
diff --git a/src/Rules/Aggregate/ComboLastNum.php b/src/Rules/Aggregate/ComboLastNum.php
@@ -0,0 +1,44 @@
+<?php
+
+/**
+ * JBZoo Toolbox - Csv-Blueprint.
+ *
+ * This file is part of the JBZoo Toolbox project.
+ * For the full copyright and license information, please view the LICENSE
+ * file that was distributed with this source code.
+ *
+ * @license    MIT
+ * @copyright  Copyright (C) JBZoo.com, All rights reserved.
+ * @see        https://github.com/JBZoo/Csv-Blueprint
+ */
+
+declare(strict_types=1);
+
+namespace JBZoo\CsvBlueprint\Rules\Aggregate;
+
+use JBZoo\CsvBlueprint\Rules\AbstarctRule;
+
+final class ComboLastNum extends AbstarctAggregateRuleCombo
+{
+    public const INPUT_TYPE = AbstarctRule::INPUT_TYPE_FLOATS;
+
+    protected const NAME = 'last value';
+
+    protected const HELP_TOP = ['Last number in the column. Expected value is float or integer.'];
+
+    protected const HELP_OPTIONS = [
+        self::EQ  => ['5', ''],
+        self::NOT => ['4.123', ''],
+        self::MIN => ['-1', ''],
+        self::MAX => ['2e4', ''],
+    ];
+
+    protected function getActualAggregate(array $colValues): ?float
+    {
+        if (\count($colValues) === 0) {
+            return null;
+        }
+
+        return (float)\end($colValues);
+    }
+}
diff --git a/src/Rules/Aggregate/First.php b/src/Rules/Aggregate/First.php
@@ -0,0 +1,43 @@
+<?php
+
+/**
+ * JBZoo Toolbox - Csv-Blueprint.
+ *
+ * This file is part of the JBZoo Toolbox project.
+ * For the full copyright and license information, please view the LICENSE
+ * file that was distributed with this source code.
+ *
+ * @license    MIT
+ * @copyright  Copyright (C) JBZoo.com, All rights reserved.
+ * @see        https://github.com/JBZoo/Csv-Blueprint
+ */
+
+declare(strict_types=1);
+
+namespace JBZoo\CsvBlueprint\Rules\Aggregate;
+
+use JBZoo\CsvBlueprint\Rules\AbstarctRule;
+
+final class First extends AbstarctAggregateRule
+{
+    public const INPUT_TYPE = AbstarctRule::INPUT_TYPE_STRINGS;
+
+    protected const HELP_OPTIONS = [
+        self::DEFAULT => ['Expected', 'First value in the column. Will be compared as strings.'],
+    ];
+
+    public function validateRule(array &$columnValues): ?string
+    {
+        if (\count($columnValues) === 0) {
+            return null;
+        }
+
+        $first = \reset($columnValues);
+        if ($first !== $this->getOptionAsString()) {
+            return "The first value in the column is \"<c>{$first}</c>\", " .
+                "which is not equal than the expected \"<green>{$this->getOptionAsString()}</green>\"";
+        }
+
+        return null;
+    }
+}