Change validatetext API to return array of characters (#1380)

Update the validatetext API to return a simple array of invalid characters with
their name rather than a complex object with their position. This shifts the
burden of finding the characters to the caller but is seen as more flexible.

SETUP/tests/unittests/ApiTest.php

@@ -674,21 +674,19 @@ public function test_info_for_previous_proofreaders()
         $this->assertEquals($this->TEST_OLDUSERNAME, $info[1]->username);
     }
 
-    public function test_validate_text()
+    public function test_validate_bad_text()
     {
         $project = $this->_create_available_project();
         $response = $this->validate_text($project->projectid, "This is an invĀlid test file");
-        $expected = [
-            'valid' => false,
-            'mark_array' => [["This is an inv", "", 0], ["Ā", "LATIN CAPITAL LETTER A WITH MACRON", 1], ["lid test file", "", 0]],
-        ];
+        $expected = ["invalid_chars" => ["Ā" => "LATIN CAPITAL LETTER A WITH MACRON"]];
         $this->assertEquals($expected, $response);
+    }
 
+    public function test_validate_good_text()
+    {
+        $project = $this->_create_available_project();
         $response = $this->validate_text($project->projectid, "This is a valid test file");
-        $expected = [
-            'valid' => true,
-            'mark_array' => [["This is a valid test file", "", 0]],
-        ];
+        $expected = ["invalid_chars" => []];
         $this->assertEquals($expected, $response);
     }
 }

api/dp-openapi.yaml

@@ -956,18 +956,25 @@ paths:
     put:
       tags:
       - project
-      description: Checks the given text for invalid characters, returns data indicating them
+      description: Checks the given text for invalid characters and returns any found with their names
       parameters:
       - $ref: '#/components/parameters/projectid'
       requestBody:
         $ref: '#/components/requestBodies/page_text'
       responses:
         200:
-          description: analysis data
+          description: invalid characters
           content:
             application/json:
               schema:
-                $ref: '#/components/schemas/text_analysis'
+                type: object
+                properties:
+                  invalid_chars:
+                    type: array
+                    items:
+                      type: object
+                      additionalProperties:
+                        type: string
         '400':
           $ref: '#/components/responses/InvalidValue'
         '401':
@@ -1604,21 +1611,6 @@ components:
           description: true if the page has been saved.
           type: boolean
 
-    text_analysis:
-      type: object
-      properties:
-        valid:
-          type: boolean
-        mark_array:
-          type: array
-          items:
-            type: array
-            items:
-            # grapheme string, utf8_character_name of grapheme, code integer: 0=valid 1=invalid
-              oneOf:
-                - type: string
-                - type: integer
-
     Error:
       required:
       - message

api/v1_projects.inc

@@ -793,10 +793,11 @@ function api_v1_project_checkout($method, array $data, array $query_params)
     }
 }
 
-function api_v1_project_validatetext($method, $data, array $query_params)
+function api_v1_project_validatetext(string $method, array $data, array $query_params)
 {
-    $text_checker = new TextChecker();
-    return $text_checker->analyse(receive_project_text_from_request_body('text'), $data[":projectid"]);
+    $project = $data[":projectid"];
+    $invalid_characters = $project->find_invalid_characters(receive_project_text_from_request_body('text'));
+    return ["invalid_chars" => $invalid_characters];
 }
 
 function api_v1_project_page(string $method, array $data, array $query_params)

pinc/Project.inc

@@ -1813,6 +1813,18 @@ class Project
             }
         }
     }
+
+    public function find_invalid_characters(string $text): array
+    {
+        $pattern_string = build_character_regex_filter($this->get_valid_codepoints());
+        $invalid_characters = [];
+        foreach (split_graphemes($text) as $grapheme) {
+            if (1 !== preg_match("/$pattern_string/u", $grapheme)) {
+                $invalid_characters[$grapheme] = utf8_character_name($grapheme);
+            }
+        }
+        return $invalid_characters;
+    }
 }
 
 // XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX

pinc/ProofProject.inc

@@ -118,43 +118,3 @@ class ProofProject
         return $proof_project_page->pp_checkout($pguser);
     }
 }
-
-class TextChecker
-{
-    private const VALID_TEXT = 0;
-    private const INVALID_TEXT = 1;
-
-    private $text_array;
-    private $valid_text_block;
-
-    private function append_valid(): void
-    {
-        if ($this->valid_text_block != "") {
-            $this->text_array[] = [$this->valid_text_block, "", self::VALID_TEXT];
-            $this->valid_text_block = "";
-        }
-    }
-
-    public function analyse(string $page_text, object $project): array
-    {
-        $pattern_string = build_character_regex_filter($project->get_valid_codepoints());
-        $this->text_array = [];
-        $this->valid_text_block = "";
-        $all_valid = true;
-        foreach (split_graphemes($page_text) as $grapheme) {
-            if (1 === preg_match("/$pattern_string/u", $grapheme)) {
-                $this->valid_text_block .= $grapheme;
-            } else {
-                $all_valid = false;
-                $this->append_valid();
-                $this->text_array[] = [$grapheme, utf8_character_name($grapheme), self::INVALID_TEXT];
-            }
-        }
-        // append any remaining valid text
-        $this->append_valid();
-        return [
-            "mark_array" => $this->text_array,
-            "valid" => $all_valid,
-        ];
-    }
-}