Get it back - Data schema spec fix: TSON instead of JSON (#3566) (#3569)

Co-authored-by: Jonathan Dobson <jon.m.dobson@gmail.com>

skyvern-frontend/src/components/DataSchemaInputGroup/WorkflowDataSchemaInputGroup.tsx

@@ -17,6 +17,7 @@ import { AutoResizingTextarea } from "../AutoResizingTextarea/AutoResizingTextar
 import { Button } from "../ui/button";
 import { AxiosError } from "axios";
 import { toast } from "../ui/use-toast";
+import { TSON } from "@/util/tson";
 import { cn } from "@/util/utils";
 
 type Props = {
@@ -36,42 +37,13 @@ function WorkflowDataSchemaInputGroup({
   const [generateWithAIActive, setGenerateWithAIActive] = useState(false);
   const [generateWithAIPrompt, setGenerateWithAIPrompt] = useState("");
 
-  function computeJsonError(
-    jsonText: string,
-  ): { message: string; line?: number; column?: number } | null {
-    try {
-      JSON.parse(jsonText);
-      return null;
-    } catch (e) {
-      const message = e instanceof Error ? e.message : "Invalid JSON";
-      // Try to extract position and compute line/column for friendlier feedback
-      const match = message.match(/position\s+(\d+)/i);
-      if (!match) {
-        return { message };
-      }
-      const pos = Number(match[1]);
-      if (Number.isNaN(pos)) {
-        return { message };
-      }
-      let line = 1;
-      let col = 1;
-      for (let i = 0; i < Math.min(pos, jsonText.length); i++) {
-        if (jsonText[i] === "\n") {
-          line += 1;
-          col = 1;
-        } else {
-          col += 1;
-        }
-      }
-      return { message, line, column: col };
-    }
-  }
-
-  const jsonError = useMemo(() => {
+  const tsonResult = useMemo(() => {
     if (value === "null") return null;
-    return computeJsonError(value);
+    return TSON.parse(value);
   }, [value]);
 
+  console.log({ tsonResult });
+
   const getDataSchemaSuggestionMutation = useMutation({
     mutationFn: async () => {
       const client = await getClient(credentialGetter);
@@ -161,7 +133,9 @@ function WorkflowDataSchemaInputGroup({
           <div
             className={cn(
               "rounded-md",
-              jsonError ? "ring-1 ring-red-500" : undefined,
+              tsonResult && !tsonResult.success
+                ? "ring-1 ring-red-500"
+                : undefined,
             )}
           >
             <CodeEditor
@@ -172,12 +146,8 @@ function WorkflowDataSchemaInputGroup({
               fontSize={8}
             />
           </div>
-          {jsonError && (
-            <div className="text-xs text-red-400">
-              {jsonError.line && jsonError.column
-                ? `Invalid JSON (${jsonError.line}:${jsonError.column}) — ${jsonError.message}`
-                : `Invalid JSON — ${jsonError.message}`}
-            </div>
+          {tsonResult !== null && !tsonResult.success && tsonResult.error && (
+            <div className="text-xs text-red-400">{tsonResult.error}</div>
           )}
         </div>
       )}

skyvern-frontend/src/routes/workflows/editor/workflowEditorUtils.ts

@@ -2,6 +2,9 @@ import Dagre from "@dagrejs/dagre";
 import type { Node } from "@xyflow/react";
 import { Edge } from "@xyflow/react";
 import { nanoid } from "nanoid";
+
+import { TSON } from "@/util/tson";
+
 import {
   WorkflowBlockType,
   WorkflowBlockTypes,
@@ -2188,10 +2191,12 @@ function getWorkflowErrors(nodes: Array<AppNode>): Array<string> {
     }
     // Validate Task data schema JSON when enabled (value different from "null")
     if (node.data.dataSchema && node.data.dataSchema !== "null") {
-      try {
-        JSON.parse(node.data.dataSchema);
-      } catch {
-        errors.push(`${node.data.label}: Data schema is not valid JSON.`);
+      const result = TSON.parse(node.data.dataSchema);
+
+      if (!result.success) {
+        errors.push(
+          `${node.data.label}: Data schema has invalid templated JSON: ${result.error ?? "-"}`,
+        );
       }
     }
   });
@@ -2227,10 +2232,12 @@ function getWorkflowErrors(nodes: Array<AppNode>): Array<string> {
     }
     // Validate Extraction data schema JSON when enabled (value different from "null")
     if (node.data.dataSchema && node.data.dataSchema !== "null") {
-      try {
-        JSON.parse(node.data.dataSchema);
-      } catch {
-        errors.push(`${node.data.label}: Data schema is not valid JSON.`);
+      const result = TSON.parse(node.data.dataSchema);
+
+      if (!result.success) {
+        errors.push(
+          `${node.data.label}: Data schema has invalid templated JSON: ${result.error ?? "-"}`,
+        );
       }
     }
   });

skyvern-frontend/src/util/tson.test.ts

@@ -0,0 +1,91 @@
+// TSON.parse.test.ts
+import { TSON } from "./tson";
+import { describe, test, expect } from "vitest";
+
+describe("TSON.parse", () => {
+  test("single top-level template works", () => {
+    const input = "{{ hello }}";
+    const result = TSON.parse(input);
+
+    expect(result.success).toBe(true);
+    expect(result.data).toEqual("<STUB>");
+  });
+
+  test("preserves double braces inside quoted strings", () => {
+    const input = '{"a": "{{ hello }}"}';
+    const result = TSON.parse(input);
+
+    expect(result.success).toBe(true);
+    expect(result.data).toEqual({ a: "{{ hello }}" });
+  });
+
+  test("replaces double braces outside strings with stub", () => {
+    const input = '{"a": {{ hello }} }';
+    const result = TSON.parse(input);
+
+    expect(result.success).toBe(true);
+    expect(result.data).toEqual({ a: "<STUB>" });
+  });
+
+  test("handles double braces in keys and values", () => {
+    const input = `
+{
+    "hello": "world",
+    {{foo}}: "bar",
+    "baz": {{quux}}
+}`;
+    const result = TSON.parse(input);
+
+    expect(result.success).toBe(true);
+    expect(result.data).toEqual({
+      hello: "world",
+      "<STUB>": "bar",
+      baz: "<STUB>",
+    });
+  });
+
+  test("does not allow trailing commas", () => {
+    const input = `
+{
+    "hello": "world",
+    {{foo}}: "bar",
+    "baz": {{quux}},
+}`;
+    const result = TSON.parse(input);
+
+    expect(result.success).toBe(false);
+    expect(result.error).toContain("Expected double-quoted property name");
+  });
+
+  test("detects unclosed double braces", () => {
+    const input = "{{ unclosed";
+    const result = TSON.parse(input);
+
+    expect(result.success).toBe(false);
+    expect(result.error).toContain("Unclosed");
+  });
+
+  test("detects unmatched closing double braces", () => {
+    const input = "closed }}";
+    const result = TSON.parse(input);
+
+    expect(result.success).toBe(false);
+    expect(result.error).toContain("Unmatched");
+  });
+
+  test("handles nested double braces", () => {
+    const input = "{{ {{ nested }} }}";
+    const result = TSON.parse(input);
+
+    expect(result.success).toBe(true);
+    expect(result.data).toEqual("<STUB>");
+  });
+
+  test("handles double braces in arrays", () => {
+    const input = '[{{ }}, {{ }}, "normal"]';
+    const result = TSON.parse(input);
+
+    expect(result.success).toBe(true);
+    expect(result.data).toEqual(["<STUB>", "<STUB>", "normal"]);
+  });
+});

skyvern-frontend/src/util/tson.ts

@@ -0,0 +1,223 @@
+type JSONValue =
+  | string
+  | number
+  | boolean
+  | null
+  | JSONValue[]
+  | { [key: string]: JSONValue };
+
+interface ParseResult {
+  success: boolean;
+  data?: JSONValue;
+  error?: string;
+}
+
+const placeholder = () => "<STUB>";
+
+/**
+ * TSON ("templated JSON") is a superset of JSON, where double curly braces {{...}} can:
+ *  - exist anywhere outside of string literals, and
+ *  - are treated as placeholders
+ */
+const TSON = {
+  parse(input: string): ParseResult {
+    try {
+      const balanceCheck = checkDoubleBraceBalance(input);
+
+      if (!balanceCheck.balanced) {
+        return {
+          success: false,
+          error: balanceCheck.error,
+        };
+      }
+
+      const pipeline = [
+        replaceBracesOutsideQuotes,
+        // removeTrailingCommas,
+        JSON.parse,
+      ];
+
+      const parsed = pipeline.reduce((acc, fn) => fn(acc), input) as JSONValue;
+
+      return {
+        success: true,
+        data: parsed,
+      };
+    } catch (error) {
+      return {
+        success: false,
+        error: error instanceof Error ? error.message : String(error),
+      };
+    }
+  },
+};
+
+function checkDoubleBraceBalance(input: string): {
+  balanced: boolean;
+  error?: string;
+} {
+  let inString = false;
+  let escapeNext = false;
+  let depth = 0;
+
+  for (let i = 0; i < input.length; i++) {
+    const char = input[i];
+    const nextChar = input[i + 1];
+
+    // handle escape sequences
+    if (escapeNext) {
+      escapeNext = false;
+      continue;
+    }
+
+    if (char === "\\") {
+      escapeNext = true;
+      continue;
+    }
+
+    // inside-string tracking
+    if (char === '"') {
+      inString = !inString;
+      continue;
+    }
+
+    // double braces counts (only outside strings)
+    if (!inString) {
+      if (char === "{" && nextChar === "{") {
+        depth++;
+        i++; // skip next char
+      } else if (char === "}" && nextChar === "}") {
+        depth--;
+        if (depth < 0) {
+          return {
+            balanced: false,
+            error: `Unmatched closing }} at position ${i}`,
+          };
+        }
+        i++; // skip next char
+      }
+    }
+  }
+
+  if (depth > 0) {
+    return {
+      balanced: false,
+      error: `Unclosed {{ - missing ${depth} closing }}`,
+    };
+  }
+
+  return { balanced: true };
+}
+
+function replaceBracesOutsideQuotes(input: string): string {
+  let result = "";
+  let inString = false;
+  let escapeNext = false;
+  let inDoubleBrace = 0; // track nesting depth of {{...}}
+
+  for (let i = 0; i < input.length; i++) {
+    const char = input[i];
+    const nextChar = input[i + 1];
+
+    // escape sequences
+    if (escapeNext) {
+      if (inDoubleBrace === 0) {
+        result += char;
+      }
+      escapeNext = false;
+      continue;
+    }
+
+    if (char === "\\") {
+      if (inDoubleBrace === 0) {
+        result += char;
+      }
+      escapeNext = true;
+      continue;
+    }
+
+    // inside-string tracking
+    if (char === '"') {
+      inString = !inString;
+      if (inDoubleBrace === 0) {
+        result += char;
+      }
+      continue;
+    }
+
+    // double braces (only outside strings)
+    if (!inString) {
+      if (char === "{" && nextChar === "{") {
+        if (inDoubleBrace === 0) {
+          result += `"${placeholder()}"`;
+        }
+        inDoubleBrace++;
+        i++; // skip next char
+        continue;
+      } else if (char === "}" && nextChar === "}") {
+        inDoubleBrace--;
+        i++; // skip next char
+        continue;
+      }
+    }
+
+    // add characters when we're not inside double braces
+    if (inDoubleBrace === 0) {
+      result += char;
+    }
+  }
+
+  return result;
+}
+
+export function removeTrailingCommas(input: string): string {
+  let result = "";
+  let inString = false;
+  let escapeNext = false;
+
+  for (let i = 0; i < input.length; i++) {
+    const char = input[i];
+
+    // escape sequences
+    if (escapeNext) {
+      result += char;
+      escapeNext = false;
+      continue;
+    }
+
+    if (char === "\\") {
+      result += char;
+      escapeNext = true;
+      continue;
+    }
+
+    // inside-string tracking
+    if (char === '"') {
+      inString = !inString;
+      result += char;
+      continue;
+    }
+
+    // check for trailing commas (outside strings)
+    if (!inString && char === ",") {
+      // look-ahead for the next non-whitespace character
+      let j = i + 1;
+      // eslint-disable-next-line @typescript-eslint/ban-ts-comment
+      // @ts-ignore
+      while (j < input.length && /\s/.test(input[j])) {
+        j++;
+      }
+
+      // if next non-whitespace is } or ], skip the comma
+      if (j < input.length && (input[j] === "}" || input[j] === "]")) {
+        continue; // Skip this comma
+      }
+    }
+
+    result += char;
+  }
+
+  return result;
+}
+
+export { TSON };