@deepagents/text2sql 7.0.2 → 7.1.0

package/dist/lib/synthesis/index.js

@@ -24,31 +24,31 @@ async function toPairs(producer) {
 var { Parser } = nodeSqlParser;
 var parser = new Parser();
 var FilteredProducer = class extends PairProducer {
+  #producer;
+  #options;
   /**
    * @param producer - Source producer to filter
    * @param options - Filter configuration
    */
   constructor(producer, options = {}) {
     super();
-    this.producer = producer;
-    this.options = options;
+    this.#producer = producer;
+    this.#options = options;
   }
-  producer;
-  options;
   /**
    * Produces pairs filtered by success status, table usage, and custom predicates.
    * @returns Pairs matching all configured filter criteria
    */
   async *produce() {
-    for await (const chunk of this.producer.produce()) {
+    for await (const chunk of this.#producer.produce()) {
       const filtered = chunk.filter((pair) => {
-        if (this.options.successOnly !== false && !pair.success) {
+        if (this.#options.successOnly !== false && !pair.success) {
           return false;
         }
-        if (this.options.tables?.length) {
-          if (!this.matchesTables(pair.sql, this.options.tables)) return false;
+        if (this.#options.tables?.length) {
+          if (!this.matchesTables(pair.sql, this.#options.tables)) return false;
         }
-        if (this.options.filter && !this.options.filter(pair)) {
+        if (this.#options.filter && !this.#options.filter(pair)) {
           return false;
         }
         return true;
@@ -65,7 +65,7 @@ var FilteredProducer = class extends PairProducer {
     const filterNames = this.#filterNames;
     try {
       const refs = parser.tableList(sql, {
-        database: this.options.dialect
+        database: this.#options.dialect
       });
       const sqlNames = refs.map((r) => r.split("::").pop().toLowerCase());
       return sqlNames.some((t) => filterNames.has(t));
@@ -81,25 +81,25 @@ import nodeSqlParser2 from "node-sql-parser";
 var { Parser: Parser2 } = nodeSqlParser2;
 var parser2 = new Parser2();
 var DeduplicatedProducer = class extends PairProducer {
+  #producer;
+  #options;
   /**
    * @param producer - Source producer to deduplicate
    * @param options - Deduplication configuration
    */
   constructor(producer, options = {}) {
     super();
-    this.producer = producer;
-    this.options = options;
+    this.#producer = producer;
+    this.#options = options;
   }
-  producer;
-  options;
   /**
    * Produces pairs with duplicates removed based on the configured strategy.
    * @returns Unique pairs after deduplication
    */
   async *produce() {
-    const { strategy = "exact" } = this.options;
+    const { strategy = "exact" } = this.#options;
     const seen = /* @__PURE__ */ new Set();
-    for await (const chunk of this.producer.produce()) {
+    for await (const chunk of this.#producer.produce()) {
       const unique = [];
       for (const pair of chunk) {
         let key;
@@ -126,8 +126,8 @@ var DeduplicatedProducer = class extends PairProducer {
   }
   normalizeSQL(sql) {
     try {
-      const ast = parser2.astify(sql, { database: this.options.dialect });
-      return parser2.sqlify(ast, { database: this.options.dialect }).toLowerCase();
+      const ast = parser2.astify(sql, { database: this.#options.dialect });
+      return parser2.sqlify(ast, { database: this.#options.dialect }).toLowerCase();
     } catch {
       return sql.toLowerCase().replace(/\s+/g, " ").trim();
     }
@@ -136,6 +136,9 @@ var DeduplicatedProducer = class extends PairProducer {
 
 // packages/text2sql/src/lib/synthesis/decorators/validated-producer.ts
 var ValidatedProducer = class extends PairProducer {
+  #producer;
+  #adapter;
+  #options;
   /**
    * @param producer - Source producer to validate
    * @param adapter - Database adapter for SQL validation
@@ -143,24 +146,21 @@ var ValidatedProducer = class extends PairProducer {
    */
   constructor(producer, adapter, options = {}) {
     super();
-    this.producer = producer;
-    this.adapter = adapter;
-    this.options = options;
+    this.#producer = producer;
+    this.#adapter = adapter;
+    this.#options = options;
   }
-  producer;
-  adapter;
-  options;
   /**
    * Produces pairs with SQL validation applied, optionally executing queries.
    * @returns Validated pairs with error/rowCount metadata attached
    */
   async *produce() {
-    for await (const chunk of this.producer.produce()) {
+    for await (const chunk of this.#producer.produce()) {
       const validated = [];
       for (const pair of chunk) {
-        const error = await this.adapter.validate(pair.sql);
+        const error = await this.#adapter.validate(pair.sql);
         if (error) {
-          if (!this.options.removeInvalid) {
+          if (!this.#options.removeInvalid) {
             validated.push({
               ...pair,
               success: false,
@@ -170,9 +170,9 @@ var ValidatedProducer = class extends PairProducer {
           continue;
         }
         let rowCount;
-        if (this.options.execute) {
+        if (this.#options.execute) {
           try {
-            const result = await this.adapter.execute(pair.sql);
+            const result = await this.#adapter.execute(pair.sql);
             rowCount = Array.isArray(result) ? result.length : void 0;
           } catch {
           }
@@ -197,7 +197,7 @@ import { getToolName as getToolName2, isToolUIPart as isToolUIPart2 } from "ai";
 import { groq } from "@ai-sdk/groq";
 import { getToolName, isTextUIPart, isToolUIPart } from "ai";
 import dedent from "dedent";
-import z from "zod";
+import z2 from "zod";
 import {
   ContextEngine,
   InMemoryContextStore,
@@ -206,8 +206,18 @@ import {
   structuredOutput,
   user
 } from "@deepagents/context";
-var contextResolverSchema = z.object({
-  question: z.string().describe(
+
+// packages/text2sql/src/lib/synthesis/extractors/tool-input.ts
+import z from "zod";
+var sqlToolInput = z.object({ sql: z.string() });
+function toolInputSql(input) {
+  const parsed = sqlToolInput.safeParse(input);
+  return parsed.success ? parsed.data.sql : void 0;
+}
+
+// packages/text2sql/src/lib/synthesis/extractors/base-contextual-extractor.ts
+var contextResolverSchema = z2.object({
+  question: z2.string().describe(
     "A standalone natural language question that the SQL query answers"
   )
 });
@@ -320,8 +330,8 @@ var BaseContextualExtractor = class extends PairProducer {
       if (getToolName(part) !== toolName) {
         continue;
       }
-      const toolInput = "input" in part ? part.input : void 0;
-      if (!toolInput?.sql) {
+      const sql = toolInputSql("input" in part ? part.input : void 0);
+      if (!sql) {
         continue;
       }
       const success = part.state === "output-available";
@@ -337,7 +347,7 @@ var BaseContextualExtractor = class extends PairProducer {
         continue;
       }
       this.results.push({
-        sql: toolInput.sql,
+        sql,
         success,
         conversationContext: snapshot
       });
@@ -402,8 +412,8 @@ var MessageExtractor = class extends PairProducer {
           if (getToolName2(part) !== toolName) {
             continue;
           }
-          const toolInput = "input" in part ? part.input : void 0;
-          if (!toolInput?.sql) {
+          const sql = toolInputSql("input" in part ? part.input : void 0);
+          if (!sql) {
             continue;
           }
           const success = part.state === "output-available";
@@ -421,7 +431,7 @@ var MessageExtractor = class extends PairProducer {
           yield [
             {
               question,
-              sql: toolInput.sql,
+              sql,
               success
             }
           ];
@@ -434,7 +444,7 @@ var MessageExtractor = class extends PairProducer {
 // packages/text2sql/src/lib/synthesis/extractors/sql-extractor.ts
 import { groq as groq2 } from "@ai-sdk/groq";
 import dedent2 from "dedent";
-import z2 from "zod";
+import z3 from "zod";
 import {
   ContextEngine as ContextEngine2,
   InMemoryContextStore as InMemoryContextStore2,
@@ -443,8 +453,8 @@ import {
   structuredOutput as structuredOutput2,
   user as user2
 } from "@deepagents/context";
-var outputSchema = z2.object({
-  question: z2.string().describe("A natural language question that the SQL query answers")
+var outputSchema = z3.object({
+  question: z3.string().describe("A natural language question that the SQL query answers")
 });
 var SqlExtractor = class extends PairProducer {
   #sqls;
@@ -579,7 +589,7 @@ var WindowedContextExtractor = class extends BaseContextualExtractor {
 // packages/text2sql/src/lib/synthesis/extractors/segmented-context-extractor.ts
 import { groq as groq3 } from "@ai-sdk/groq";
 import dedent3 from "dedent";
-import z3 from "zod";
+import z4 from "zod";
 import {
   ContextEngine as ContextEngine3,
   InMemoryContextStore as InMemoryContextStore3,
@@ -588,9 +598,9 @@ import {
   structuredOutput as structuredOutput3,
   user as user3
 } from "@deepagents/context";
-var topicChangeSchema = z3.object({
-  isTopicChange: z3.boolean().describe("Whether the new message represents a topic change"),
-  reason: z3.string().describe("Brief explanation for the decision")
+var topicChangeSchema = z4.object({
+  isTopicChange: z4.boolean().describe("Whether the new message represents a topic change"),
+  reason: z4.string().describe("Brief explanation for the decision")
 });
 async function detectTopicChange(params) {
   const context = new ContextEngine3({
@@ -740,7 +750,7 @@ var LastQueryExtractor = class extends BaseContextualExtractor {
 import { groq as groq4 } from "@ai-sdk/groq";
 import dedent4 from "dedent";
 import pLimit from "p-limit";
-import z4 from "zod";
+import z5 from "zod";
 import "@deepagents/agent";
 import {
   ContextEngine as ContextEngine4,
@@ -786,9 +796,9 @@ var styleInstructions = {
 };
 
 // packages/text2sql/src/lib/synthesis/synthesizers/breadth-evolver.ts
-var paraphraserOutputSchema = z4.object({
-  paraphrases: z4.array(
-    z4.string().describe("A paraphrased version of the original question")
+var paraphraserOutputSchema = z5.object({
+  paraphrases: z5.array(
+    z5.string().describe("A paraphrased version of the original question")
   ).min(1).describe("List of paraphrased questions that would produce the same SQL")
 });
 async function paraphraseQuestion(params) {
@@ -865,33 +875,33 @@ async function paraphraseQuestion(params) {
   return paraphraserOutput.generate();
 }
 var BreadthEvolver = class extends PairProducer {
+  #source;
+  #options;
+  #limit;
   /**
    * @param source - Source pairs or producer to evolve
    * @param options - Evolution options including count, persona, and concurrency
    */
   constructor(source, options) {
     super();
-    this.source = source;
-    this.options = options;
-    this.#limit = pLimit(this.options.concurrency ?? 4);
+    this.#source = source;
+    this.#options = options;
+    this.#limit = pLimit(this.#options.concurrency ?? 4);
   }
-  source;
-  options;
-  #limit;
   /**
    * Batch pairs within each chunk for concurrent processing.
    * Uses pLimit for concurrency control, yields results per pair after chunk completes.
    */
   async *produce() {
-    for await (const chunk of this.from(this.source)) {
+    for await (const chunk of this.from(this.#source)) {
       const tasks = chunk.map(
         (pair) => this.#limit(async () => {
           const result = await paraphraseQuestion({
             question: pair.question,
             sql: pair.sql,
-            count: this.options.count,
-            persona: this.options.persona,
-            model: this.options.model
+            count: this.#options.count,
+            persona: this.#options.persona,
+            model: this.#options.model
           });
           return result.paraphrases.map((paraphrase) => ({
             question: paraphrase,
@@ -913,7 +923,7 @@ import { NoObjectGeneratedError as NoObjectGeneratedError2, NoOutputGeneratedErr
 import dedent6 from "dedent";
 import pLimit2 from "p-limit";
 import pRetry2 from "p-retry";
-import z6 from "zod";
+import z7 from "zod";
 import "@deepagents/agent";
 import {
   ContextEngine as ContextEngine6,
@@ -993,7 +1003,7 @@ import {
 } from "ai";
 import dedent5 from "dedent";
 import pRetry from "p-retry";
-import z5 from "zod";
+import z6 from "zod";
 import "@deepagents/agent";
 import {
   ContextEngine as ContextEngine5,
@@ -1285,14 +1295,14 @@ Question: ${options.input}
       const sqlOutput = structuredOutput5({
         model,
         context,
-        schema: z5.object({
-          result: z5.union([
-            z5.object({
-              sql: z5.string().describe("The SQL query that answers the question"),
-              reasoning: z5.string().describe("The reasoning steps taken to generate the SQL")
+        schema: z6.object({
+          result: z6.union([
+            z6.object({
+              sql: z6.string().describe("The SQL query that answers the question"),
+              reasoning: z6.string().describe("The reasoning steps taken to generate the SQL")
             }),
-            z5.object({
-              error: z5.string().describe(
+            z6.object({
+              error: z6.string().describe(
                 "Error message explaining why the question cannot be answered with the given schema"
               )
             })
@@ -1343,11 +1353,11 @@ Question: ${options.input}
         const forcedSqlOutput = structuredOutput5({
           model,
           context,
-          schema: z5.object({
-            sql: z5.string().describe(
+          schema: z6.object({
+            sql: z6.string().describe(
               "Best-effort SQL query that answers the core intent using only available schema entities."
             ),
-            reasoning: z5.string().describe("Reasoning steps for best-effort schema mapping.")
+            reasoning: z6.string().describe("Reasoning steps for best-effort schema mapping.")
           })
         });
         try {
@@ -1472,8 +1482,8 @@ var techniqueInstructions = {
     - "Get costs" → "What would be the impact of a 10% discount on profit margins?"
   `
 };
-var evolverOutputSchema = z6.object({
-  evolvedQuestion: z6.string().describe("The evolved, more complex version of the original question")
+var evolverOutputSchema = z7.object({
+  evolvedQuestion: z7.string().describe("The evolved, more complex version of the original question")
 });
 async function evolveQuestion(params) {
   const context = new ContextEngine6({
@@ -1542,6 +1552,10 @@ var ALL_TECHNIQUES = [
   "hypothetical"
 ];
 var DepthEvolver = class extends PairProducer {
+  #source;
+  #adapter;
+  #options;
+  #limit;
   /**
    * @param source - Source pairs or producer to evolve
    * @param adapter - Database adapter for SQL generation
@@ -1549,28 +1563,24 @@ var DepthEvolver = class extends PairProducer {
    */
   constructor(source, adapter, options) {
     super();
-    this.source = source;
-    this.adapter = adapter;
-    this.options = options;
-    this.#limit = pLimit2(this.options?.concurrency ?? 4);
+    this.#source = source;
+    this.#adapter = adapter;
+    this.#options = options;
+    this.#limit = pLimit2(this.#options?.concurrency ?? 4);
   }
-  source;
-  adapter;
-  options;
-  #limit;
   /**
    * Yields evolved pairs as each completes (streaming pattern).
    * Removes batch barrier - no longer waits for all evolutions before yielding.
    */
   async *produce() {
     const introspection = "";
-    const count = this.options?.count ?? 1;
-    const techniques = this.options?.techniques ?? ALL_TECHNIQUES;
+    const count = this.#options?.count ?? 1;
+    const techniques = this.#options?.techniques ?? ALL_TECHNIQUES;
     let pairIndex = 0;
-    for await (const chunk of this.from(this.source)) {
+    for await (const chunk of this.from(this.#source)) {
       for (const pair of chunk) {
         const tasks = Array.from({ length: count }, (_, i) => {
-          const technique = this.options?.techniques ? techniques[i % techniques.length] : techniques[(pairIndex * count + i) % techniques.length];
+          const technique = this.#options?.techniques ? techniques[i % techniques.length] : techniques[(pairIndex * count + i) % techniques.length];
           return this.#limit(
             () => this.#processTask(pair, technique, introspection)
           );
@@ -1589,16 +1599,16 @@ var DepthEvolver = class extends PairProducer {
         schema: introspection,
         technique,
         techniqueInstruction: techniqueInstructions[technique],
-        model: this.options?.model
+        model: this.#options?.model
       })
     );
     const evolvedQuestion = output.evolvedQuestion;
     try {
       const sqlResult = await toSql({
         input: evolvedQuestion,
-        adapter: this.adapter,
+        adapter: this.#adapter,
         fragments: [],
-        model: this.options?.model
+        model: this.#options?.model
       });
       return {
         question: evolvedQuestion,
@@ -1648,7 +1658,7 @@ async function withRetry2(computation) {
 // packages/text2sql/src/lib/synthesis/synthesizers/persona-generator.ts
 import { groq as groq7 } from "@ai-sdk/groq";
 import dedent7 from "dedent";
-import z7 from "zod";
+import z8 from "zod";
 import "@deepagents/agent";
 import {
   ContextEngine as ContextEngine7,
@@ -1660,14 +1670,14 @@ import {
   structuredOutput as structuredOutput7,
   user as user7
 } from "@deepagents/context";
-var outputSchema2 = z7.object({
-  personas: z7.array(
-    z7.object({
-      role: z7.string().describe("The job title or role of this persona"),
-      perspective: z7.string().describe(
+var outputSchema2 = z8.object({
+  personas: z8.array(
+    z8.object({
+      role: z8.string().describe("The job title or role of this persona"),
+      perspective: z8.string().describe(
         "Rich description of what this persona cares about when querying the database"
       ),
-      styles: z7.array(z7.enum(ALL_STYLES)).min(1).max(3).describe(
+      styles: z8.array(z8.enum(ALL_STYLES)).min(1).max(3).describe(
         "Typical communication styles for this persona (1-3 styles)"
       )
     })
@@ -1767,7 +1777,7 @@ import pLimit3 from "p-limit";
 // packages/text2sql/src/lib/agents/question.agent.ts
 import { groq as groq8 } from "@ai-sdk/groq";
 import dedent8 from "dedent";
-import z8 from "zod";
+import z9 from "zod";
 import "@deepagents/agent";
 import {
   ContextEngine as ContextEngine8,
@@ -1816,8 +1826,8 @@ var complexityInstructions = {
     Examples: "What is the running total of sales per month?", "Rank customers by their purchase frequency within each region"
   `
 };
-var outputSchema3 = z8.object({
-  questions: z8.array(z8.string().describe("A natural language question about the data")).min(1).describe("List of natural language questions a user might ask")
+var outputSchema3 = z9.object({
+  questions: z9.array(z9.string().describe("A natural language question about the data")).min(1).describe("List of natural language questions a user might ask")
 });
 async function generateQuestions(params) {
   const { introspection, complexity, count, prompt, model } = params;
@@ -1878,23 +1888,23 @@ async function generateQuestions(params) {
 
 // packages/text2sql/src/lib/synthesis/synthesizers/schema-synthesizer.ts
 var SchemaSynthesizer = class extends PairProducer {
+  #adapter;
+  #options;
+  #complexities = [];
+  #personas = [];
+  #limit;
   /**
    * @param adapter - Database adapter for schema introspection and SQL validation
    * @param options - Synthesis configuration including count, complexity, and concurrency
    */
   constructor(adapter, options) {
     super();
-    this.adapter = adapter;
-    this.options = options;
-    this.#complexities = Array.isArray(this.options.complexity) ? this.options.complexity : [this.options.complexity ?? "moderate"];
-    this.#personas = this.options.personas ?? [void 0];
-    this.#limit = pLimit3(this.options.concurrency ?? 5);
+    this.#adapter = adapter;
+    this.#options = options;
+    this.#complexities = Array.isArray(this.#options.complexity) ? this.#options.complexity : [this.#options.complexity ?? "moderate"];
+    this.#personas = this.#options.personas ?? [void 0];
+    this.#limit = pLimit3(this.#options.concurrency ?? 5);
   }
-  adapter;
-  options;
-  #complexities = [];
-  #personas = [];
-  #limit;
   /**
    * Generates question-SQL pairs by iterating through all persona × complexity combinations.
    * Uses parallel processing bounded by the configured concurrency limit.
@@ -1927,14 +1937,14 @@ var SchemaSynthesizer = class extends PairProducer {
 Generate questions this persona would ask.` : void 0;
     const prompt = personaContext ? `${personaContext}
 
-Generate ${this.options.count} questions at ${complexity} complexity.` : void 0;
+Generate ${this.#options.count} questions at ${complexity} complexity.` : void 0;
     const { questions } = await this.#limit(
       () => generateQuestions({
         introspection,
         complexity,
-        count: this.options.count,
+        count: this.#options.count,
         prompt,
-        model: this.options.model
+        model: this.#options.model
       })
     );
     const pairs = await Promise.all(
@@ -1943,9 +1953,9 @@ Generate ${this.options.count} questions at ${complexity} complexity.` : void 0;
           try {
             return await toSql({
               input: question,
-              adapter: this.adapter,
-              fragments: this.options.teachings ?? [],
-              model: this.options.model
+              adapter: this.#adapter,
+              fragments: this.#options.teachings ?? [],
+              model: this.#options.model
             });
           } catch (error) {
             if (UnanswerableSQLError.isInstance(error)) {