agentic-dataset-builder 0.1.0 → 0.2.0

package/dist/schemas/qwen35.d.ts

@@ -0,0 +1,338 @@
+import { z } from 'zod';
+export declare const Qwen35TextBlockSchema: z.ZodObject<{
+    type: z.ZodLiteral<"text">;
+    text: z.ZodString;
+}, z.core.$strip>;
+export declare const Qwen35ImageBlockSchema: z.ZodObject<{
+    type: z.ZodLiteral<"image">;
+    image_url: z.ZodOptional<z.ZodString>;
+    placeholder: z.ZodOptional<z.ZodBoolean>;
+    placeholder_token: z.ZodOptional<z.ZodString>;
+    source_kind: z.ZodOptional<z.ZodString>;
+    metadata: z.ZodOptional<z.ZodRecord<z.ZodString, z.ZodUnknown>>;
+}, z.core.$strip>;
+export declare const Qwen35VideoBlockSchema: z.ZodObject<{
+    type: z.ZodLiteral<"video">;
+    video_url: z.ZodOptional<z.ZodString>;
+    placeholder: z.ZodOptional<z.ZodBoolean>;
+    placeholder_token: z.ZodOptional<z.ZodString>;
+    source_kind: z.ZodOptional<z.ZodString>;
+    metadata: z.ZodOptional<z.ZodRecord<z.ZodString, z.ZodUnknown>>;
+}, z.core.$strip>;
+export declare const Qwen35ContentSchema: z.ZodUnion<readonly [z.ZodString, z.ZodArray<z.ZodUnion<readonly [z.ZodObject<{
+    type: z.ZodLiteral<"text">;
+    text: z.ZodString;
+}, z.core.$strip>, z.ZodObject<{
+    type: z.ZodLiteral<"image">;
+    image_url: z.ZodOptional<z.ZodString>;
+    placeholder: z.ZodOptional<z.ZodBoolean>;
+    placeholder_token: z.ZodOptional<z.ZodString>;
+    source_kind: z.ZodOptional<z.ZodString>;
+    metadata: z.ZodOptional<z.ZodRecord<z.ZodString, z.ZodUnknown>>;
+}, z.core.$strip>, z.ZodObject<{
+    type: z.ZodLiteral<"video">;
+    video_url: z.ZodOptional<z.ZodString>;
+    placeholder: z.ZodOptional<z.ZodBoolean>;
+    placeholder_token: z.ZodOptional<z.ZodString>;
+    source_kind: z.ZodOptional<z.ZodString>;
+    metadata: z.ZodOptional<z.ZodRecord<z.ZodString, z.ZodUnknown>>;
+}, z.core.$strip>]>>]>;
+export declare const Qwen35ToolCallSchema: z.ZodObject<{
+    type: z.ZodDefault<z.ZodLiteral<"function">>;
+    id: z.ZodOptional<z.ZodString>;
+    function: z.ZodObject<{
+        name: z.ZodString;
+        arguments: z.ZodDefault<z.ZodRecord<z.ZodString, z.ZodUnknown>>;
+    }, z.core.$strip>;
+}, z.core.$strip>;
+export declare const Qwen35ToolSpecSchema: z.ZodObject<{
+    name: z.ZodString;
+    description: z.ZodOptional<z.ZodString>;
+    parameters: z.ZodOptional<z.ZodRecord<z.ZodString, z.ZodUnknown>>;
+}, z.core.$strip>;
+export declare const Qwen35MessageSchema: z.ZodDiscriminatedUnion<[z.ZodObject<{
+    role: z.ZodLiteral<"system">;
+    content: z.ZodUnion<readonly [z.ZodString, z.ZodArray<z.ZodUnion<readonly [z.ZodObject<{
+        type: z.ZodLiteral<"text">;
+        text: z.ZodString;
+    }, z.core.$strip>, z.ZodObject<{
+        type: z.ZodLiteral<"image">;
+        image_url: z.ZodOptional<z.ZodString>;
+        placeholder: z.ZodOptional<z.ZodBoolean>;
+        placeholder_token: z.ZodOptional<z.ZodString>;
+        source_kind: z.ZodOptional<z.ZodString>;
+        metadata: z.ZodOptional<z.ZodRecord<z.ZodString, z.ZodUnknown>>;
+    }, z.core.$strip>, z.ZodObject<{
+        type: z.ZodLiteral<"video">;
+        video_url: z.ZodOptional<z.ZodString>;
+        placeholder: z.ZodOptional<z.ZodBoolean>;
+        placeholder_token: z.ZodOptional<z.ZodString>;
+        source_kind: z.ZodOptional<z.ZodString>;
+        metadata: z.ZodOptional<z.ZodRecord<z.ZodString, z.ZodUnknown>>;
+    }, z.core.$strip>]>>]>;
+}, z.core.$strip>, z.ZodObject<{
+    role: z.ZodLiteral<"user">;
+    content: z.ZodUnion<readonly [z.ZodString, z.ZodArray<z.ZodUnion<readonly [z.ZodObject<{
+        type: z.ZodLiteral<"text">;
+        text: z.ZodString;
+    }, z.core.$strip>, z.ZodObject<{
+        type: z.ZodLiteral<"image">;
+        image_url: z.ZodOptional<z.ZodString>;
+        placeholder: z.ZodOptional<z.ZodBoolean>;
+        placeholder_token: z.ZodOptional<z.ZodString>;
+        source_kind: z.ZodOptional<z.ZodString>;
+        metadata: z.ZodOptional<z.ZodRecord<z.ZodString, z.ZodUnknown>>;
+    }, z.core.$strip>, z.ZodObject<{
+        type: z.ZodLiteral<"video">;
+        video_url: z.ZodOptional<z.ZodString>;
+        placeholder: z.ZodOptional<z.ZodBoolean>;
+        placeholder_token: z.ZodOptional<z.ZodString>;
+        source_kind: z.ZodOptional<z.ZodString>;
+        metadata: z.ZodOptional<z.ZodRecord<z.ZodString, z.ZodUnknown>>;
+    }, z.core.$strip>]>>]>;
+}, z.core.$strip>, z.ZodObject<{
+    role: z.ZodLiteral<"assistant">;
+    content: z.ZodUnion<readonly [z.ZodString, z.ZodArray<z.ZodUnion<readonly [z.ZodObject<{
+        type: z.ZodLiteral<"text">;
+        text: z.ZodString;
+    }, z.core.$strip>, z.ZodObject<{
+        type: z.ZodLiteral<"image">;
+        image_url: z.ZodOptional<z.ZodString>;
+        placeholder: z.ZodOptional<z.ZodBoolean>;
+        placeholder_token: z.ZodOptional<z.ZodString>;
+        source_kind: z.ZodOptional<z.ZodString>;
+        metadata: z.ZodOptional<z.ZodRecord<z.ZodString, z.ZodUnknown>>;
+    }, z.core.$strip>, z.ZodObject<{
+        type: z.ZodLiteral<"video">;
+        video_url: z.ZodOptional<z.ZodString>;
+        placeholder: z.ZodOptional<z.ZodBoolean>;
+        placeholder_token: z.ZodOptional<z.ZodString>;
+        source_kind: z.ZodOptional<z.ZodString>;
+        metadata: z.ZodOptional<z.ZodRecord<z.ZodString, z.ZodUnknown>>;
+    }, z.core.$strip>]>>]>;
+    reasoning_content: z.ZodOptional<z.ZodString>;
+    tool_calls: z.ZodOptional<z.ZodArray<z.ZodObject<{
+        type: z.ZodDefault<z.ZodLiteral<"function">>;
+        id: z.ZodOptional<z.ZodString>;
+        function: z.ZodObject<{
+            name: z.ZodString;
+            arguments: z.ZodDefault<z.ZodRecord<z.ZodString, z.ZodUnknown>>;
+        }, z.core.$strip>;
+    }, z.core.$strip>>>;
+}, z.core.$strip>, z.ZodObject<{
+    role: z.ZodLiteral<"tool">;
+    content: z.ZodUnion<readonly [z.ZodString, z.ZodArray<z.ZodUnion<readonly [z.ZodObject<{
+        type: z.ZodLiteral<"text">;
+        text: z.ZodString;
+    }, z.core.$strip>, z.ZodObject<{
+        type: z.ZodLiteral<"image">;
+        image_url: z.ZodOptional<z.ZodString>;
+        placeholder: z.ZodOptional<z.ZodBoolean>;
+        placeholder_token: z.ZodOptional<z.ZodString>;
+        source_kind: z.ZodOptional<z.ZodString>;
+        metadata: z.ZodOptional<z.ZodRecord<z.ZodString, z.ZodUnknown>>;
+    }, z.core.$strip>, z.ZodObject<{
+        type: z.ZodLiteral<"video">;
+        video_url: z.ZodOptional<z.ZodString>;
+        placeholder: z.ZodOptional<z.ZodBoolean>;
+        placeholder_token: z.ZodOptional<z.ZodString>;
+        source_kind: z.ZodOptional<z.ZodString>;
+        metadata: z.ZodOptional<z.ZodRecord<z.ZodString, z.ZodUnknown>>;
+    }, z.core.$strip>]>>]>;
+    tool_call_id: z.ZodOptional<z.ZodString>;
+    name: z.ZodOptional<z.ZodString>;
+}, z.core.$strip>], "role">;
+export declare const Qwen35MetaSchema: z.ZodObject<{
+    endpoint: z.ZodString;
+    status: z.ZodNumber;
+    ts: z.ZodString;
+    key: z.ZodOptional<z.ZodString>;
+    source: z.ZodOptional<z.ZodString>;
+    requested_model: z.ZodOptional<z.ZodNullable<z.ZodString>>;
+    actual_model: z.ZodOptional<z.ZodNullable<z.ZodString>>;
+    stream: z.ZodOptional<z.ZodBoolean>;
+    thinking_level: z.ZodOptional<z.ZodNullable<z.ZodString>>;
+    reasoning_summary_mode: z.ZodOptional<z.ZodUnion<readonly [z.ZodString, z.ZodArray<z.ZodUnknown>, z.ZodRecord<z.ZodString, z.ZodUnknown>]>>;
+    thinking_type: z.ZodOptional<z.ZodNullable<z.ZodString>>;
+    thinking_budget_tokens: z.ZodOptional<z.ZodNullable<z.ZodNumber>>;
+    max_output_tokens: z.ZodOptional<z.ZodNullable<z.ZodNumber>>;
+    tool_spec_count: z.ZodOptional<z.ZodNumber>;
+    tool_choice: z.ZodOptional<z.ZodUnion<readonly [z.ZodString, z.ZodArray<z.ZodUnknown>, z.ZodRecord<z.ZodString, z.ZodUnknown>]>>;
+    request_contains_non_text_content: z.ZodDefault<z.ZodBoolean>;
+    request_image_block_count: z.ZodDefault<z.ZodNumber>;
+    request_video_block_count: z.ZodDefault<z.ZodNumber>;
+    request_tool_call_block_count: z.ZodDefault<z.ZodNumber>;
+    request_tool_result_block_count: z.ZodDefault<z.ZodNumber>;
+    request_thinking_block_count: z.ZodDefault<z.ZodNumber>;
+    response_contains_non_text_content: z.ZodDefault<z.ZodBoolean>;
+    response_image_block_count: z.ZodDefault<z.ZodNumber>;
+    response_video_block_count: z.ZodDefault<z.ZodNumber>;
+    response_tool_call_block_count: z.ZodDefault<z.ZodNumber>;
+    response_tool_result_block_count: z.ZodDefault<z.ZodNumber>;
+    response_thinking_block_count: z.ZodDefault<z.ZodNumber>;
+    request_truncated: z.ZodDefault<z.ZodBoolean>;
+    response_truncated: z.ZodDefault<z.ZodBoolean>;
+    lossy_source: z.ZodDefault<z.ZodBoolean>;
+    lossy_reasons: z.ZodDefault<z.ZodArray<z.ZodString>>;
+    dataset_label: z.ZodOptional<z.ZodString>;
+    dataset_source_system: z.ZodOptional<z.ZodString>;
+    dataset_source_bucket: z.ZodOptional<z.ZodString>;
+    dataset_source_file: z.ZodOptional<z.ZodString>;
+    dataset_has_reasoning: z.ZodOptional<z.ZodBoolean>;
+    dataset_reasoning_chars: z.ZodOptional<z.ZodNumber>;
+}, z.core.$strip>;
+export declare const Qwen35RecordSchema: z.ZodObject<{
+    id: z.ZodString;
+    request_id: z.ZodOptional<z.ZodString>;
+    messages: z.ZodArray<z.ZodDiscriminatedUnion<[z.ZodObject<{
+        role: z.ZodLiteral<"system">;
+        content: z.ZodUnion<readonly [z.ZodString, z.ZodArray<z.ZodUnion<readonly [z.ZodObject<{
+            type: z.ZodLiteral<"text">;
+            text: z.ZodString;
+        }, z.core.$strip>, z.ZodObject<{
+            type: z.ZodLiteral<"image">;
+            image_url: z.ZodOptional<z.ZodString>;
+            placeholder: z.ZodOptional<z.ZodBoolean>;
+            placeholder_token: z.ZodOptional<z.ZodString>;
+            source_kind: z.ZodOptional<z.ZodString>;
+            metadata: z.ZodOptional<z.ZodRecord<z.ZodString, z.ZodUnknown>>;
+        }, z.core.$strip>, z.ZodObject<{
+            type: z.ZodLiteral<"video">;
+            video_url: z.ZodOptional<z.ZodString>;
+            placeholder: z.ZodOptional<z.ZodBoolean>;
+            placeholder_token: z.ZodOptional<z.ZodString>;
+            source_kind: z.ZodOptional<z.ZodString>;
+            metadata: z.ZodOptional<z.ZodRecord<z.ZodString, z.ZodUnknown>>;
+        }, z.core.$strip>]>>]>;
+    }, z.core.$strip>, z.ZodObject<{
+        role: z.ZodLiteral<"user">;
+        content: z.ZodUnion<readonly [z.ZodString, z.ZodArray<z.ZodUnion<readonly [z.ZodObject<{
+            type: z.ZodLiteral<"text">;
+            text: z.ZodString;
+        }, z.core.$strip>, z.ZodObject<{
+            type: z.ZodLiteral<"image">;
+            image_url: z.ZodOptional<z.ZodString>;
+            placeholder: z.ZodOptional<z.ZodBoolean>;
+            placeholder_token: z.ZodOptional<z.ZodString>;
+            source_kind: z.ZodOptional<z.ZodString>;
+            metadata: z.ZodOptional<z.ZodRecord<z.ZodString, z.ZodUnknown>>;
+        }, z.core.$strip>, z.ZodObject<{
+            type: z.ZodLiteral<"video">;
+            video_url: z.ZodOptional<z.ZodString>;
+            placeholder: z.ZodOptional<z.ZodBoolean>;
+            placeholder_token: z.ZodOptional<z.ZodString>;
+            source_kind: z.ZodOptional<z.ZodString>;
+            metadata: z.ZodOptional<z.ZodRecord<z.ZodString, z.ZodUnknown>>;
+        }, z.core.$strip>]>>]>;
+    }, z.core.$strip>, z.ZodObject<{
+        role: z.ZodLiteral<"assistant">;
+        content: z.ZodUnion<readonly [z.ZodString, z.ZodArray<z.ZodUnion<readonly [z.ZodObject<{
+            type: z.ZodLiteral<"text">;
+            text: z.ZodString;
+        }, z.core.$strip>, z.ZodObject<{
+            type: z.ZodLiteral<"image">;
+            image_url: z.ZodOptional<z.ZodString>;
+            placeholder: z.ZodOptional<z.ZodBoolean>;
+            placeholder_token: z.ZodOptional<z.ZodString>;
+            source_kind: z.ZodOptional<z.ZodString>;
+            metadata: z.ZodOptional<z.ZodRecord<z.ZodString, z.ZodUnknown>>;
+        }, z.core.$strip>, z.ZodObject<{
+            type: z.ZodLiteral<"video">;
+            video_url: z.ZodOptional<z.ZodString>;
+            placeholder: z.ZodOptional<z.ZodBoolean>;
+            placeholder_token: z.ZodOptional<z.ZodString>;
+            source_kind: z.ZodOptional<z.ZodString>;
+            metadata: z.ZodOptional<z.ZodRecord<z.ZodString, z.ZodUnknown>>;
+        }, z.core.$strip>]>>]>;
+        reasoning_content: z.ZodOptional<z.ZodString>;
+        tool_calls: z.ZodOptional<z.ZodArray<z.ZodObject<{
+            type: z.ZodDefault<z.ZodLiteral<"function">>;
+            id: z.ZodOptional<z.ZodString>;
+            function: z.ZodObject<{
+                name: z.ZodString;
+                arguments: z.ZodDefault<z.ZodRecord<z.ZodString, z.ZodUnknown>>;
+            }, z.core.$strip>;
+        }, z.core.$strip>>>;
+    }, z.core.$strip>, z.ZodObject<{
+        role: z.ZodLiteral<"tool">;
+        content: z.ZodUnion<readonly [z.ZodString, z.ZodArray<z.ZodUnion<readonly [z.ZodObject<{
+            type: z.ZodLiteral<"text">;
+            text: z.ZodString;
+        }, z.core.$strip>, z.ZodObject<{
+            type: z.ZodLiteral<"image">;
+            image_url: z.ZodOptional<z.ZodString>;
+            placeholder: z.ZodOptional<z.ZodBoolean>;
+            placeholder_token: z.ZodOptional<z.ZodString>;
+            source_kind: z.ZodOptional<z.ZodString>;
+            metadata: z.ZodOptional<z.ZodRecord<z.ZodString, z.ZodUnknown>>;
+        }, z.core.$strip>, z.ZodObject<{
+            type: z.ZodLiteral<"video">;
+            video_url: z.ZodOptional<z.ZodString>;
+            placeholder: z.ZodOptional<z.ZodBoolean>;
+            placeholder_token: z.ZodOptional<z.ZodString>;
+            source_kind: z.ZodOptional<z.ZodString>;
+            metadata: z.ZodOptional<z.ZodRecord<z.ZodString, z.ZodUnknown>>;
+        }, z.core.$strip>]>>]>;
+        tool_call_id: z.ZodOptional<z.ZodString>;
+        name: z.ZodOptional<z.ZodString>;
+    }, z.core.$strip>], "role">>;
+    tools: z.ZodDefault<z.ZodArray<z.ZodObject<{
+        name: z.ZodString;
+        description: z.ZodOptional<z.ZodString>;
+        parameters: z.ZodOptional<z.ZodRecord<z.ZodString, z.ZodUnknown>>;
+    }, z.core.$strip>>>;
+    meta: z.ZodObject<{
+        endpoint: z.ZodString;
+        status: z.ZodNumber;
+        ts: z.ZodString;
+        key: z.ZodOptional<z.ZodString>;
+        source: z.ZodOptional<z.ZodString>;
+        requested_model: z.ZodOptional<z.ZodNullable<z.ZodString>>;
+        actual_model: z.ZodOptional<z.ZodNullable<z.ZodString>>;
+        stream: z.ZodOptional<z.ZodBoolean>;
+        thinking_level: z.ZodOptional<z.ZodNullable<z.ZodString>>;
+        reasoning_summary_mode: z.ZodOptional<z.ZodUnion<readonly [z.ZodString, z.ZodArray<z.ZodUnknown>, z.ZodRecord<z.ZodString, z.ZodUnknown>]>>;
+        thinking_type: z.ZodOptional<z.ZodNullable<z.ZodString>>;
+        thinking_budget_tokens: z.ZodOptional<z.ZodNullable<z.ZodNumber>>;
+        max_output_tokens: z.ZodOptional<z.ZodNullable<z.ZodNumber>>;
+        tool_spec_count: z.ZodOptional<z.ZodNumber>;
+        tool_choice: z.ZodOptional<z.ZodUnion<readonly [z.ZodString, z.ZodArray<z.ZodUnknown>, z.ZodRecord<z.ZodString, z.ZodUnknown>]>>;
+        request_contains_non_text_content: z.ZodDefault<z.ZodBoolean>;
+        request_image_block_count: z.ZodDefault<z.ZodNumber>;
+        request_video_block_count: z.ZodDefault<z.ZodNumber>;
+        request_tool_call_block_count: z.ZodDefault<z.ZodNumber>;
+        request_tool_result_block_count: z.ZodDefault<z.ZodNumber>;
+        request_thinking_block_count: z.ZodDefault<z.ZodNumber>;
+        response_contains_non_text_content: z.ZodDefault<z.ZodBoolean>;
+        response_image_block_count: z.ZodDefault<z.ZodNumber>;
+        response_video_block_count: z.ZodDefault<z.ZodNumber>;
+        response_tool_call_block_count: z.ZodDefault<z.ZodNumber>;
+        response_tool_result_block_count: z.ZodDefault<z.ZodNumber>;
+        response_thinking_block_count: z.ZodDefault<z.ZodNumber>;
+        request_truncated: z.ZodDefault<z.ZodBoolean>;
+        response_truncated: z.ZodDefault<z.ZodBoolean>;
+        lossy_source: z.ZodDefault<z.ZodBoolean>;
+        lossy_reasons: z.ZodDefault<z.ZodArray<z.ZodString>>;
+        dataset_label: z.ZodOptional<z.ZodString>;
+        dataset_source_system: z.ZodOptional<z.ZodString>;
+        dataset_source_bucket: z.ZodOptional<z.ZodString>;
+        dataset_source_file: z.ZodOptional<z.ZodString>;
+        dataset_has_reasoning: z.ZodOptional<z.ZodBoolean>;
+        dataset_reasoning_chars: z.ZodOptional<z.ZodNumber>;
+    }, z.core.$strip>;
+    label: z.ZodOptional<z.ZodString>;
+    source_system: z.ZodOptional<z.ZodString>;
+    source_bucket: z.ZodOptional<z.ZodString>;
+    source_file: z.ZodOptional<z.ZodString>;
+    agentic_label: z.ZodOptional<z.ZodObject<{
+        label: z.ZodString;
+        tool_call_count: z.ZodOptional<z.ZodNumber>;
+        tool_message_count: z.ZodOptional<z.ZodNumber>;
+        dialogue_rounds_est: z.ZodOptional<z.ZodNumber>;
+        reasoning_chars: z.ZodOptional<z.ZodNumber>;
+        has_reasoning: z.ZodOptional<z.ZodBoolean>;
+        lossy_source: z.ZodOptional<z.ZodBoolean>;
+        lossy_reasons: z.ZodOptional<z.ZodArray<z.ZodString>>;
+    }, z.core.$strip>>;
+}, z.core.$strip>;
+export type Qwen35Record = z.infer<typeof Qwen35RecordSchema>;

package/dist/schemas/qwen35.js

@@ -0,0 +1,139 @@
+import { z } from 'zod';
+export const Qwen35TextBlockSchema = z.object({
+    type: z.literal('text'),
+    text: z.string(),
+});
+export const Qwen35ImageBlockSchema = z.object({
+    type: z.literal('image'),
+    image_url: z.string().optional(),
+    placeholder: z.boolean().optional(),
+    placeholder_token: z.string().optional(),
+    source_kind: z.string().optional(),
+    metadata: z.record(z.string(), z.unknown()).optional(),
+});
+export const Qwen35VideoBlockSchema = z.object({
+    type: z.literal('video'),
+    video_url: z.string().optional(),
+    placeholder: z.boolean().optional(),
+    placeholder_token: z.string().optional(),
+    source_kind: z.string().optional(),
+    metadata: z.record(z.string(), z.unknown()).optional(),
+});
+export const Qwen35ContentSchema = z.union([
+    z.string(),
+    z.array(z.union([Qwen35TextBlockSchema, Qwen35ImageBlockSchema, Qwen35VideoBlockSchema])),
+]);
+export const Qwen35ToolCallSchema = z.object({
+    type: z.literal('function').default('function'),
+    id: z.string().optional(),
+    function: z.object({
+        name: z.string(),
+        arguments: z.record(z.string(), z.unknown()).default({}),
+    }),
+});
+export const Qwen35ToolSpecSchema = z.object({
+    name: z.string(),
+    description: z.string().optional(),
+    parameters: z.record(z.string(), z.unknown()).optional(),
+});
+export const Qwen35MessageSchema = z.discriminatedUnion('role', [
+    z.object({ role: z.literal('system'), content: Qwen35ContentSchema }),
+    z.object({ role: z.literal('user'), content: Qwen35ContentSchema }),
+    z.object({
+        role: z.literal('assistant'),
+        content: Qwen35ContentSchema,
+        reasoning_content: z.string().optional(),
+        tool_calls: z.array(Qwen35ToolCallSchema).optional(),
+    }),
+    z.object({
+        role: z.literal('tool'),
+        content: Qwen35ContentSchema,
+        tool_call_id: z.string().optional(),
+        name: z.string().optional(),
+    }),
+]);
+export const Qwen35MetaSchema = z.object({
+    endpoint: z.string(),
+    status: z.number().int().min(100).max(599),
+    ts: z.string(),
+    key: z.string().optional(),
+    source: z.string().optional(),
+    requested_model: z.string().nullable().optional(),
+    actual_model: z.string().nullable().optional(),
+    stream: z.boolean().optional(),
+    thinking_level: z.string().nullable().optional(),
+    reasoning_summary_mode: z.union([z.string(), z.array(z.unknown()), z.record(z.string(), z.unknown())]).optional(),
+    thinking_type: z.string().nullable().optional(),
+    thinking_budget_tokens: z.number().int().nonnegative().nullable().optional(),
+    max_output_tokens: z.number().int().nonnegative().nullable().optional(),
+    tool_spec_count: z.number().int().nonnegative().optional(),
+    tool_choice: z.union([z.string(), z.array(z.unknown()), z.record(z.string(), z.unknown())]).optional(),
+    request_contains_non_text_content: z.boolean().default(false),
+    request_image_block_count: z.number().int().nonnegative().default(0),
+    request_video_block_count: z.number().int().nonnegative().default(0),
+    request_tool_call_block_count: z.number().int().nonnegative().default(0),
+    request_tool_result_block_count: z.number().int().nonnegative().default(0),
+    request_thinking_block_count: z.number().int().nonnegative().default(0),
+    response_contains_non_text_content: z.boolean().default(false),
+    response_image_block_count: z.number().int().nonnegative().default(0),
+    response_video_block_count: z.number().int().nonnegative().default(0),
+    response_tool_call_block_count: z.number().int().nonnegative().default(0),
+    response_tool_result_block_count: z.number().int().nonnegative().default(0),
+    response_thinking_block_count: z.number().int().nonnegative().default(0),
+    request_truncated: z.boolean().default(false),
+    response_truncated: z.boolean().default(false),
+    lossy_source: z.boolean().default(false),
+    lossy_reasons: z.array(z.string()).default([]),
+    dataset_label: z.string().optional(),
+    dataset_source_system: z.string().optional(),
+    dataset_source_bucket: z.string().optional(),
+    dataset_source_file: z.string().optional(),
+    dataset_has_reasoning: z.boolean().optional(),
+    dataset_reasoning_chars: z.number().int().nonnegative().optional(),
+});
+export const Qwen35RecordSchema = z.object({
+    id: z.string(),
+    request_id: z.string().optional(),
+    messages: z.array(Qwen35MessageSchema).min(1),
+    tools: z.array(Qwen35ToolSpecSchema).default([]),
+    meta: Qwen35MetaSchema,
+    label: z.string().optional(),
+    source_system: z.string().optional(),
+    source_bucket: z.string().optional(),
+    source_file: z.string().optional(),
+    agentic_label: z
+        .object({
+        label: z.string(),
+        tool_call_count: z.number().int().nonnegative().optional(),
+        tool_message_count: z.number().int().nonnegative().optional(),
+        dialogue_rounds_est: z.number().int().nonnegative().optional(),
+        reasoning_chars: z.number().int().nonnegative().optional(),
+        has_reasoning: z.boolean().optional(),
+        lossy_source: z.boolean().optional(),
+        lossy_reasons: z.array(z.string()).optional(),
+    })
+        .optional(),
+}).superRefine((record, ctx) => {
+    const seenUser = record.messages.some((message) => message.role === 'user');
+    if (!seenUser) {
+        ctx.addIssue({ code: 'custom', message: 'at least one user message is required' });
+    }
+    let seenNonSystem = false;
+    for (const message of record.messages) {
+        if (message.role !== 'system') {
+            seenNonSystem = true;
+        }
+        else if (seenNonSystem) {
+            ctx.addIssue({ code: 'custom', message: 'system messages must appear only at the beginning' });
+            break;
+        }
+        if (message.role === 'assistant' && typeof message.reasoning_content === 'string') {
+            if (message.reasoning_content.includes('<think>') || message.reasoning_content.includes('</think>')) {
+                ctx.addIssue({ code: 'custom', message: 'reasoning_content must not include <think> wrappers' });
+            }
+        }
+    }
+    if (record.meta.lossy_source && record.meta.lossy_reasons.length === 0) {
+        ctx.addIssue({ code: 'custom', message: 'lossy_source requires lossy_reasons' });
+    }
+});

package/dist/sources/claude.d.ts

@@ -0,0 +1,2 @@
+import { type Qwen35Record } from '../schemas/qwen35.js';
+export declare function collectClaudePromptOnlyRecords(root: string): Promise<Qwen35Record[]>;

package/dist/sources/claude.js

@@ -0,0 +1,64 @@
+import fg from 'fast-glob';
+import { z } from 'zod';
+import { Qwen35RecordSchema } from '../schemas/qwen35.js';
+import { readJsonl } from '../utils/jsonl.js';
+const EntrySchema = z.record(z.string(), z.unknown());
+export async function collectClaudePromptOnlyRecords(root) {
+    const files = await fg('**/*.jsonl', { cwd: root, absolute: true, onlyFiles: true });
+    const records = [];
+    for (const file of files.sort()) {
+        const entries = (await readJsonl(file)).map((row) => EntrySchema.parse(row));
+        for (const entry of entries) {
+            if (entry.type !== 'user')
+                continue;
+            const message = isRecord(entry.message) ? entry.message : {};
+            const content = asString(message.content);
+            if (!content)
+                continue;
+            records.push(Qwen35RecordSchema.parse({
+                id: `${asString(entry.sessionId) ?? file}:${asString(entry.promptId) ?? asString(entry.uuid) ?? 'prompt'}`,
+                request_id: asString(entry.promptId) ?? asString(entry.uuid) ?? undefined,
+                messages: [{ role: 'user', content }],
+                tools: [],
+                meta: {
+                    endpoint: 'claude/prompt_history',
+                    status: 200,
+                    ts: asString(entry.timestamp) ?? '',
+                    key: asString(entry.sessionId) ?? undefined,
+                    source: `claude:session=${asString(entry.sessionId)}:cwd=${asString(entry.cwd)}:entrypoint=${asString(entry.entrypoint)}`,
+                    requested_model: undefined,
+                    actual_model: undefined,
+                    stream: false,
+                    thinking_level: undefined,
+                    reasoning_summary_mode: 'claude_prompt_only',
+                    thinking_type: 'prompt_history_only',
+                    tool_spec_count: 0,
+                    tool_choice: { mode: 'prompt_only' },
+                    request_contains_non_text_content: false,
+                    request_image_block_count: 0,
+                    request_video_block_count: 0,
+                    request_tool_call_block_count: 0,
+                    request_tool_result_block_count: 0,
+                    request_thinking_block_count: 0,
+                    response_contains_non_text_content: false,
+                    response_image_block_count: 0,
+                    response_video_block_count: 0,
+                    response_tool_call_block_count: 0,
+                    response_tool_result_block_count: 0,
+                    response_thinking_block_count: 0,
+                    request_truncated: false,
+                    response_truncated: false,
+                    lossy_source: true,
+                    lossy_reasons: ['prompt_history_only', 'assistant_trace_unavailable'],
+                },
+            }));
+        }
+    }
+    return records;
+}
+function asString(value) {
+    return typeof value === 'string' ? value : undefined;
+}
+function isRecord(value) {
+    return Boolean(value) && typeof value === 'object' && !Array.isArray(value);
+}

package/dist/sources/codex.d.ts

@@ -0,0 +1,2 @@
+import { type Qwen35Record } from '../schemas/qwen35.js';
+export declare function collectCodexRecords(root: string): Promise<Qwen35Record[]>;