Support PICT-compatible inline schema definitions (#187)

Author: eviltester

README.md

@@ -88,7 +88,7 @@ The spec is a paragraph of text where each line is either a 'name' or a 'rule':
 
 - **Comments**: lines starting with `#` (optionally prefixed by whitespace) are treated as comments.
 - **Blank lines**: blank lines are allowed and ignored, so you can separate column groups for readability.
-- **Column definitions**: each column is defined as `name` followed by `rule` on the next logical content line.
+- **Column definitions**: each column can be defined either as `name` followed by `rule` on the next logical content line, or inline as `name: rule`.
 - **Constraints**: optional `IF ... THEN ...` statements may appear in text mode after the field definitions, terminated by either `;` or `ENDIF`.
 
 ```
@@ -101,6 +101,9 @@ rule
 name
 rule
 
+# compact pict-style alternative
+status: enum(active,inactive)
+
 IF [name] = "Bob" THEN [status] = "active" ENDIF
 ```
 

docs-src/docs/040-test-data/018-Schema-Definition.md

@@ -16,7 +16,7 @@ This page explains:
 
 ## Basic schema format
 
-A schema is written as repeating two-line field definitions:
+A schema is usually written as repeating two-line field definitions:
 
 ```text
 Column Name
@@ -32,6 +32,14 @@ enum("Open","In Progress","Closed")
 
 This creates one output column called `Status`.
 
+You can also use a compact inline form when you prefer a PICT-style layout:
+
+```text
+Status: enum("Open","In Progress","Closed")
+```
+
+Both formats are supported, and you can mix them in the same schema.
+
 ## Field rule examples
 
 ### Literal values

packages/core/js/data_generation/rulesParser.js

@@ -6,6 +6,57 @@ function startsConstraint(trimmedLine) {
   return /^IF\s+(?:\[|\(|NOT\b)/i.test(trimmedLine);
 }
 
+function looksLikeInlineRuleSpec(ruleText) {
+  const trimmed = String(ruleText ?? '').trim();
+  if (trimmed.length === 0 || startsConstraint(trimmed)) {
+    return false;
+  }
+
+  if (
+    /^(?:enum|literal|regex|datatype\.(?:enum|literal|regex)|awd\.datatype\.(?:enum|literal|regex))\s*\(/i.test(trimmed)
+  ) {
+    return true;
+  }
+
+  if (/^(?:faker\.)?[A-Za-z][A-Za-z0-9_]*(?:\.[A-Za-z][A-Za-z0-9_]*)+(?:\s*\(.*\)\s*|\s*)$/i.test(trimmed)) {
+    return true;
+  }
+
+  if (!trimmed.includes(',')) {
+    return false;
+  }
+
+  const values = trimmed.split(',').map((value) => value.trim());
+  if (values.length < 2 || values.some((value) => value.length === 0 || value.length > 50)) {
+    return false;
+  }
+
+  return !values.some((value) => /[[\]{}()^$*+?|\\]/.test(value) || (value.includes('.') && /[A-Z]/.test(value)));
+}
+
+function parseInlineRuleDefinition(line) {
+  const source = String(line ?? '');
+  for (let index = 0; index < source.length; index += 1) {
+    if (source[index] !== ':') {
+      continue;
+    }
+
+    const name = source.slice(0, index).trim();
+    const rule = source.slice(index + 1).trim();
+    if (name.length === 0 || !looksLikeInlineRuleSpec(rule)) {
+      continue;
+    }
+
+    return {
+      name,
+      rule,
+      separator: ': ',
+    };
+  }
+
+  return null;
+}
+
 function isEscapedQuote(text, index) {
   let backslashCount = 0;
   let back = index - 1;
@@ -144,6 +195,22 @@ export class RulesParser {
           pendingLeadingTextLines = [];
           continue;
         }
+        const inlineRule = parseInlineRuleDefinition(line);
+        if (inlineRule) {
+          this.testDataRules.addRule(inlineRule.name, inlineRule.rule, {
+            comments: pendingLeadingTextLines.join('\n'),
+          });
+          this.schemaTokens.push({
+            kind: 'rule',
+            name: inlineRule.name,
+            rule: inlineRule.rule,
+            line: index + 1,
+            inline: true,
+            separator: inlineRule.separator,
+          });
+          pendingLeadingTextLines = [];
+          continue;
+        }
         pendingName = trimmed;
         pendingNameLine = index + 1;
         continue;
@@ -216,8 +283,12 @@ export class RulesParser {
       }
       if (token.kind === 'rule') {
         if (rowIndex < rows.length) {
-          outputLines.push(rows[rowIndex].name);
-          outputLines.push(rows[rowIndex].rule);
+          if (token.inline) {
+            outputLines.push(`${rows[rowIndex].name}${token.separator || ': '}${rows[rowIndex].rule}`);
+          } else {
+            outputLines.push(rows[rowIndex].name);
+            outputLines.push(rows[rowIndex].rule);
+          }
           rowIndex += 1;
         }
       }

packages/core/js/data_generation/schema-conversion.js

@@ -100,8 +100,12 @@ function renderSpecFromRulesWithTokens(rules, constraints, schemaTokens) {
       return;
     }
     if (token?.kind === 'rule' && rowIndex < rows.length) {
-      outputLines.push(rows[rowIndex].name);
-      outputLines.push(rows[rowIndex].rule);
+      if (token.inline) {
+        outputLines.push(`${rows[rowIndex].name}${token.separator || ': '}${rows[rowIndex].rule}`);
+      } else {
+        outputLines.push(rows[rowIndex].name);
+        outputLines.push(rows[rowIndex].rule);
+      }
       rowIndex += 1;
     }
   });

packages/core/src/index.js

@@ -78,14 +78,73 @@ const SUPPORTED_FORMATS = [
   'asciitable',
 ];
 
+function looksLikeInlineSchemaRule(ruleText) {
+  const trimmed = String(ruleText ?? '').trim();
+  if (trimmed.length === 0 || /^IF\s+(?:\[|\(|NOT\b)/i.test(trimmed)) {
+    return false;
+  }
+
+  if (
+    /^(?:enum|literal|regex|datatype\.(?:enum|literal|regex)|awd\.datatype\.(?:enum|literal|regex))\s*\(/i.test(trimmed)
+  ) {
+    return true;
+  }
+
+  if (/^(?:faker\.)?[A-Za-z][A-Za-z0-9_]*(?:\.[A-Za-z][A-Za-z0-9_]*)+(?:\s*\(.*\)\s*|\s*)$/i.test(trimmed)) {
+    return true;
+  }
+
+  if (!trimmed.includes(',')) {
+    return false;
+  }
+
+  const values = trimmed.split(',').map((value) => value.trim());
+  if (values.length < 2 || values.some((value) => value.length === 0 || value.length > 50)) {
+    return false;
+  }
+
+  return !values.some((value) => /[[\]{}()^$*+?|\\]/.test(value) || (value.includes('.') && /[A-Z]/.test(value)));
+}
+
 function extractRuleLines(textSpec) {
   if (typeof textSpec !== 'string') {
     return [];
   }
   const lines = textSpec.split(/\r?\n/);
   const ruleLines = [];
-  for (let i = 1; i < lines.length; i += 2) {
-    ruleLines.push(lines[i].trim());
+  let pendingName = null;
+  for (const line of lines) {
+    const trimmed = line.trim();
+    if (trimmed.length === 0 || /^\s*#/.test(line) || /^IF\s+(?:\[|\(|NOT\b)/i.test(trimmed)) {
+      pendingName = null;
+      continue;
+    }
+
+    let matchedInlineRule = false;
+    for (let separatorIndex = 0; separatorIndex < line.length; separatorIndex += 1) {
+      if (line[separatorIndex] !== ':') {
+        continue;
+      }
+      const rule = line.slice(separatorIndex + 1).trim();
+      if (looksLikeInlineSchemaRule(rule)) {
+        ruleLines.push(rule);
+        pendingName = null;
+        matchedInlineRule = true;
+        break;
+      }
+    }
+
+    if (matchedInlineRule) {
+      continue;
+    }
+
+    if (pendingName === null) {
+      pendingName = trimmed;
+      continue;
+    }
+
+    ruleLines.push(trimmed);
+    pendingName = null;
   }
   return ruleLines;
 }

packages/core/src/tests/core-api/amendFromTextSpecAndData.test.js

@@ -57,6 +57,24 @@ test('defaults rowCount to imported row count', () => {
   expect(result.diagnostics.importedRowCount).toBe(2);
 });
 
+test('supports pict-style inline schema definitions for amend flows', () => {
+  const result = amendFromTextSpecAndData({
+    textSpec: 'Status: literal(Active)\nRole: enum(Admin,User)',
+    inputData: '"Name"\n"Alice"\n"Eve"',
+    inputFormat: 'csv',
+    outputFormat: 'json',
+  });
+
+  expect(result.ok).toBe(true);
+  expect(result.headers).toEqual(['Name', 'Status', 'Role']);
+  expect(result.rows).toHaveLength(2);
+  result.rows.forEach((row) => {
+    expect(['Alice', 'Eve']).toContain(row[0]);
+    expect(row[1]).toBe('Active');
+    expect(['Admin', 'User']).toContain(row[2]);
+  });
+});
+
 test('amends only first N rows when rowCount is smaller', () => {
   const result = amendFromTextSpecAndData({
     textSpec: 'Name\nBob',

packages/core/src/tests/core-api/generateFromTextSpec.test.js

@@ -40,6 +40,24 @@ test('generateFromTextSpec generates rows for valid spec', () => {
   assertNoCommonErrorPatternsInRows(result.rows);
 });
 
+test('generateFromTextSpec supports pict-style inline schema definitions', () => {
+  const result = generateFromTextSpec({
+    textSpec: 'Browser: Chrome,Firefox,Safari\nStatus: enum("Open","Closed")\nName: person.fullName',
+    rowCount: 3,
+    outputFormat: 'json',
+  });
+
+  expect(result.ok).toBe(true);
+  expect(result.headers).toEqual(['Browser', 'Status', 'Name']);
+  expect(result.rows).toHaveLength(3);
+  result.rows.forEach((row) => {
+    expect(['Chrome', 'Firefox', 'Safari']).toContain(row[0]);
+    expect(['Open', 'Closed']).toContain(row[1]);
+    expect(String(row[2]).length).toBeGreaterThan(0);
+  });
+  assertNoCommonErrorPatternsInRows(result.rows);
+});
+
 test('generateFromTextSpec serializes object return values as JSON strings', () => {
   const result = generateFromTextSpec({
     textSpec: 'Currency\nfinance.currency',
@@ -255,6 +273,11 @@ test('validateSafeFakerRules accepts known faker commands with literal args', ()
   expect(result.ok).toBe(true);
 });
 
+test('validateSafeFakerRules accepts pict-style inline faker commands', () => {
+  const result = validateSafeFakerRules('Name: person.firstName("female")\nStatus: enum(active,inactive)');
+  expect(result.ok).toBe(true);
+});
+
 test('validateSafeFakerRules accepts js-style object literal faker args', () => {
   const result = validateSafeFakerRules('Template\nhelpers.mustache("{{name}}", { name: "Ada" })');
   expect(result.ok).toBe(true);

packages/core/src/tests/data_generation/schema-rules-adapter.test.js

@@ -84,6 +84,30 @@ describe('schema rules adapter', () => {
     expect(rendered.text).toBe('t1\nliteral("")\nt2\nliteral(   123)');
   });
 
+  test('round-trips pict-style inline schema tokens', () => {
+    const schemaText = `Priority: enum(high,medium,low)
+Status: person.jobTitle`;
+
+    const parsed = schemaTextToDataRules({
+      schemaText,
+      faker,
+      RandExp,
+    });
+
+    expect(parsed.errors).toEqual([]);
+    expect(parsed.schemaTokens).toEqual([
+      expect.objectContaining({ kind: 'rule', inline: true }),
+      expect.objectContaining({ kind: 'rule', inline: true }),
+    ]);
+
+    const rendered = dataRulesToSchemaText({
+      dataRules: parsed.dataRules,
+      schemaTokens: parsed.schemaTokens,
+    });
+
+    expect(rendered.text).toBe(schemaText);
+  });
+
   test('prefers schema tokens when rendering so blank lines are preserved', () => {
     const rendered = dataRulesToSchemaText({
       dataRules: [

packages/core/src/tests/data_generation/unit/rulesParser.test.js

@@ -18,6 +18,22 @@ person.fullName`;
     expect(parser.testDataRules.rules[0].ruleSpec).toBe('person.fullName');
   });
 
+  test('can parse inline pict-style column definitions into rules', () => {
+    const inputText = `Browser: Chrome,Firefox,Safari
+Status: enum("Open","Closed")
+Name: person.fullName`;
+
+    const parser = new RulesParser(faker, RandExp);
+    parser.parseText(inputText);
+
+    expect(parser.isValid()).toBe(true);
+    expect(parser.testDataRules.rules).toHaveLength(3);
+    expect(parser.testDataRules.rules[0]).toMatchObject({ name: 'Browser', ruleSpec: 'Chrome,Firefox,Safari' });
+    expect(parser.testDataRules.rules[1]).toMatchObject({ name: 'Status', ruleSpec: 'enum("Open","Closed")' });
+    expect(parser.testDataRules.rules[2]).toMatchObject({ name: 'Name', ruleSpec: 'person.fullName' });
+    expect(parser.getSchemaTokens().every((token) => token.kind !== 'rule' || token.inline === true)).toBe(true);
+  });
+
   test('flags an empty rule definition line', () => {
     const inputText = `Name
 `;
@@ -113,6 +129,17 @@ enum(active,inactive,pending)`;
     expect(output).toBe(inputText);
   });
 
+  test('preserves inline pict-style rules when rebuilding from parsed tokens', () => {
+    const inputText = `# compact
+Priority: enum(high,medium,low)
+Status: person.jobTitle`;
+
+    const parser = new RulesParser(faker, RandExp);
+    parser.parseText(inputText);
+
+    expect(parser.renderSpecFromRulesWithTokens(parser.testDataRules.rules)).toBe(inputText);
+  });
+
   test('preserves comments and blank lines when rebuilding from rule comments', () => {
     const inputText = `# one
 
@@ -167,6 +194,21 @@ enum(open,closed)`;
     expect(parser.testDataRules.constraints).toHaveLength(0);
   });
 
+  test('does not treat non-rule colon lines as inline pict definitions', () => {
+    const inputText = `Environment: Browser
+enum(chrome,firefox)`;
+
+    const parser = new RulesParser(faker, RandExp);
+    parser.parseText(inputText);
+
+    expect(parser.isValid()).toBe(true);
+    expect(parser.testDataRules.rules).toHaveLength(1);
+    expect(parser.testDataRules.rules[0]).toMatchObject({
+      name: 'Environment: Browser',
+      ruleSpec: 'enum(chrome,firefox)',
+    });
+  });
+
   test('does not treat ENDIF inside a parameter reference as the constraint terminator', () => {
     const inputText = `ENDIF
 enum(yes,no)