@jsonstudio/llms 0.4.3 → 0.4.5

package/dist/conversion/codecs/anthropic-openai-codec.js

@@ -1,470 +1,8 @@
 import { runStandardChatRequestFilters } from '../index.js';
 import { FilterEngine, ResponseToolTextCanonicalizeFilter, ResponseToolArgumentsStringifyFilter, ResponseFinishInvariantsFilter } from '../../filters/index.js';
-function isObject(v) { return !!v && typeof v === 'object' && !Array.isArray(v); }
-export function mapAnthropicToolsToOpenAI(tools) {
-    const out = [];
-    for (const t of (Array.isArray(tools) ? tools : [])) {
-        try {
-            const name = typeof t?.name === 'string' ? String(t.name) : undefined;
-            const description = typeof t?.description === 'string' ? String(t.description) : undefined;
-            const params = t?.input_schema && isObject(t.input_schema) ? t.input_schema : { type: 'object', properties: {}, additionalProperties: true };
-            if (name) {
-                out.push({ type: 'function', function: { name, ...(description ? { description } : {}), parameters: params } });
-            }
-        }
-        catch { /* ignore invalid tool */ }
-    }
-    return out;
-}
-function safeJson(v) { try {
-    return JSON.stringify(v ?? {});
-}
-catch {
-    return '{}';
-} }
-function flattenAnthropicText(content) {
-    if (content == null)
-        return '';
-    if (typeof content === 'string')
-        return content;
-    if (Array.isArray(content))
-        return content.map(flattenAnthropicText).join('');
-    if (typeof content === 'object') {
-        const t = String(content.type || '').toLowerCase();
-        if (t === 'text' && typeof content.text === 'string')
-            return String(content.text);
-        if (Array.isArray(content.content))
-            return content.content.map(flattenAnthropicText).join('');
-        if (typeof content.content === 'string')
-            return String(content.content);
-    }
-    return '';
-}
-export function buildOpenAIChatFromAnthropic(payload) {
-    const newMessages = [];
-    // 支持 string 或 Anthropic content blocks 数组两种 system 形状
-    const systemField = payload?.system;
-    const systemBlocks = Array.isArray(systemField)
-        ? systemField
-        : (systemField !== undefined && systemField !== null ? [systemField] : []);
-    for (const block of systemBlocks) {
-        const text = flattenAnthropicText(block).trim();
-        if (text) {
-            newMessages.push({ role: 'system', content: text });
-        }
-    }
-    const msgs = Array.isArray(payload?.messages) ? payload.messages : [];
-    for (const m of msgs) {
-        if (!m || typeof m !== 'object')
-            continue;
-        const role = typeof m.role === 'string' ? String(m.role) : 'user';
-        const content = m.content;
-        if (!Array.isArray(content)) {
-            const text = flattenAnthropicText(content);
-            if (text)
-                newMessages.push({ role, content: text });
-            continue;
-        }
-        const textParts = [];
-        const toolCalls = [];
-        const toolResults = [];
-        for (const block of content) {
-            if (!block || typeof block !== 'object')
-                continue;
-            const t = String(block.type || '').toLowerCase();
-            if (t === 'text' && typeof block.text === 'string') {
-                const s = block.text.trim();
-                if (s)
-                    textParts.push(s);
-            }
-            else if (t === 'tool_use') {
-                const name = typeof block.name === 'string' ? String(block.name) : undefined;
-                const id = typeof block.id === 'string' ? String(block.id) : undefined;
-                const input = block.input ?? {};
-                if (name) {
-                    const args = safeJson(input);
-                    toolCalls.push({ id, type: 'function', function: { name, arguments: args } });
-                }
-            }
-            else if (t === 'tool_result') {
-                const callId = block.tool_call_id
-                    || block.call_id
-                    || block.tool_use_id
-                    || block.id
-                    || undefined;
-                let contentStr = '';
-                const c = block.content;
-                if (typeof c === 'string')
-                    contentStr = c;
-                else if (c != null) {
-                    try {
-                        contentStr = JSON.stringify(c);
-                    }
-                    catch {
-                        contentStr = String(c);
-                    }
-                }
-                toolResults.push({ role: 'tool', tool_call_id: callId, content: contentStr });
-            }
-        }
-        if (textParts.length > 0 || toolCalls.length > 0) {
-            const msg = { role, content: textParts.join('\n') };
-            if (toolCalls.length)
-                msg.tool_calls = toolCalls;
-            newMessages.push(msg);
-        }
-        for (const tr of toolResults)
-            newMessages.push(tr);
-    }
-    const request = { messages: newMessages };
-    if (payload && typeof payload === 'object') {
-        if (typeof payload.model === 'string')
-            request.model = payload.model;
-        if (typeof payload.max_tokens === 'number')
-            request.max_tokens = payload.max_tokens;
-        if (typeof payload.temperature === 'number')
-            request.temperature = payload.temperature;
-        if (typeof payload.top_p === 'number')
-            request.top_p = payload.top_p;
-        if (typeof payload.stream === 'boolean')
-            request.stream = payload.stream;
-        if (payload.tool_choice !== undefined)
-            request.tool_choice = payload.tool_choice;
-        if (Array.isArray(payload.tools)) {
-            request.tools = mapAnthropicToolsToOpenAI(payload.tools);
-        }
-    }
-    return request;
-}
-function buildAnthropicFromOpenAIChat(oa) {
-    // 将 OpenAI finish_reason 映射为 Anthropic stop_reason
-    const mapFinishReason = (reason) => {
-        const r = (reason || '').toString();
-        const mapping = {
-            stop: 'end_turn',
-            length: 'max_tokens',
-            tool_calls: 'tool_use',
-            content_filter: 'stop_sequence',
-            function_call: 'tool_use'
-        };
-        return mapping[r] || 'end_turn';
-    };
-    const choices = Array.isArray(oa?.choices) ? oa.choices : [];
-    const primary = (choices[0] && typeof choices[0] === 'object') ? choices[0] : {};
-    const msg = primary?.message || {};
-    const role = String(msg?.role || 'assistant');
-    const blocks = [];
-    const content = msg?.content;
-    // 保持与请求侧对称：即便 content 为空字符串也显式生成 text block，
-    // 这样 AnthropicMessage → Chat → AnthropicMessage 闭环时，空文本块不会被静默丢弃
-    // （例如 provider 返回的 [{type:'text',text:''},{type:'tool_use',...}] 应在闭环后保持同样结构）。
-    if (typeof content === 'string') {
-        blocks.push({ type: 'text', text: content });
-    }
-    if (Array.isArray(content)) {
-        const text = content.map((p) => (typeof p?.text === 'string' ? p.text : (typeof p === 'string' ? p : ''))).filter(Boolean).join('');
-        if (text)
-            blocks.push({ type: 'text', text: text });
-    }
-    const tool_calls = Array.isArray(msg?.tool_calls) ? msg.tool_calls : [];
-    for (const tc of tool_calls) {
-        try {
-            const id = String(tc?.id || `call_${Math.random().toString(36).slice(2, 10)}`);
-            const fn = tc?.function || {};
-            const name = String(fn?.name || 'tool');
-            const argsRaw = fn?.arguments;
-            let input;
-            if (typeof argsRaw === 'string') {
-                try {
-                    input = JSON.parse(argsRaw);
-                }
-                catch {
-                    input = { _raw: argsRaw };
-                }
-            }
-            else {
-                input = (argsRaw ?? {});
-            }
-            blocks.push({ type: 'tool_use', id, name, input });
-        }
-        catch { /* ignore */ }
-    }
-    const u = (oa?.usage ?? {});
-    const input_tokens = Number(u.prompt_tokens ?? u.input_tokens ?? 0);
-    const output_tokens = Number(u.completion_tokens ?? u.output_tokens ?? 0);
-    let stop_reason = mapFinishReason(oa?.choices?.[0]?.finish_reason);
-    // 对于“仅选择工具”的响应，即便 Chat finish_reason 被归一到 'stop'，
-    // 只要存在 tool_calls，就应在 Anthropic 侧显式标记为 tool_use，
-    // 与 Responses 路径保持一致（Responses 会在 required_action 中保留 tool 调用）。
-    if (tool_calls.length > 0 && stop_reason === 'end_turn') {
-        stop_reason = 'tool_use';
-    }
-    return {
-        id: `resp_${Date.now()}`,
-        type: 'message',
-        role,
-        model: String(oa?.model || 'unknown'),
-        created: Math.floor(Date.now() / 1000),
-        content: blocks,
-        usage: (input_tokens || output_tokens) ? { input_tokens, output_tokens } : undefined,
-        stop_reason
-    };
-}
-/**
- * ChatRequest → AnthropicRequest 编码器（请求侧）
- *
- * 用于在 v3 管线中，将 canonical Chat 请求编码为 Anthropic /v1/messages wire 形状：
- *  - messages: Chat messages → Anthropic content blocks（仅 text，工具调用由 tools 列表承载）
- *  - tools: Chat tools[].function → Anthropic { name, description, input_schema }
- *  - 参数：max_tokens/temperature/top_p/stop → Anthropic 对应字段
- */
-export function buildAnthropicRequestFromOpenAIChat(chatReq) {
-    const model = String(chatReq?.model || 'unknown');
-    const messages = [];
-    const collectText = (val) => {
-        if (!val)
-            return '';
-        if (typeof val === 'string')
-            return val;
-        if (Array.isArray(val))
-            return val.map(collectText).join('');
-        if (typeof val === 'object') {
-            if (typeof val.text === 'string')
-                return String(val.text);
-            if (Array.isArray(val.content))
-                return collectText(val.content);
-        }
-        return '';
-    };
-    const msgs = Array.isArray(chatReq?.messages) ? chatReq.messages : [];
-    // 预扫描 Chat 历史中的 tool_calls，只有存在配对的 tool_use/id 时，
-    // 才允许在后续请求里生成对应的 tool_result 块，避免“凭空”制造
-    // 未配对的 tool_result（上一轮未执行工具却发送 tool_result）。
-    const knownToolCallIds = new Set();
-    for (const m of msgs) {
-        if (!m || typeof m !== 'object')
-            continue;
-        const role = String(m.role || 'user');
-        if (role !== 'assistant')
-            continue;
-        const toolCalls = m.tool_calls;
-        if (!Array.isArray(toolCalls))
-            continue;
-        for (const tc of toolCalls) {
-            if (!tc || typeof tc !== 'object')
-                continue;
-            const id = tc.id;
-            if (typeof id === 'string' && id) {
-                knownToolCallIds.add(id);
-            }
-        }
-    }
-    // 兼容已有 Anthropic 形状：如果顶层已经存在 system[]，优先保留其文本，
-    // 后续再追加来自 Chat system 消息的文本，避免信息丢失。
-    const systemBlocks = [];
-    const pushSystemBlock = (text) => {
-        const trimmed = text.trim();
-        if (trimmed)
-            systemBlocks.push({ type: 'text', text: trimmed });
-    };
-    try {
-        const sys = chatReq?.system;
-        const ingestSystem = (val) => {
-            if (!val)
-                return;
-            if (typeof val === 'string') {
-                pushSystemBlock(val);
-                return;
-            }
-            if (Array.isArray(val)) {
-                for (const entry of val)
-                    ingestSystem(entry);
-                return;
-            }
-            if (typeof val === 'object') {
-                if (typeof val.text === 'string') {
-                    pushSystemBlock(String(val.text));
-                    return;
-                }
-                if (Array.isArray(val.content)) {
-                    ingestSystem(val.content);
-                    return;
-                }
-            }
-        };
-        ingestSystem(sys);
-    }
-    catch {
-        // ignore system pre-scan errors
-    }
-    for (const m of msgs) {
-        if (!m || typeof m !== 'object')
-            continue;
-        const role = String(m.role || 'user');
-        const text = collectText(m.content).trim();
-        // Chat 中的 system 消息应折叠到顶层 system[]，而不是作为 messages[0].role='system' 发送，
-        // 以满足 GLM Anthropic 兼容端“messages[*].role 只能是 user/assistant”的约束。
-        if (role === 'system') {
-            if (text)
-                pushSystemBlock(text);
-            continue;
-        }
-        // Chat 中的 role:'tool' 消息在 Anthropic 中表示为 user 侧的 tool_result block。
-        // 为避免出现 role:'tool' 的顶层消息（会触发 422），在这里专门映射：
-        if (role === 'tool') {
-            const rawToolCallId = m.tool_call_id ||
-                m.call_id ||
-                m.tool_use_id ||
-                m.id ||
-                undefined;
-            const toolCallId = rawToolCallId ? String(rawToolCallId) : undefined;
-            const hasPair = toolCallId ? knownToolCallIds.has(toolCallId) : false;
-            // 如果当前请求历史中不存在匹配的 tool_use/id，则视为未配对的
-            // tool_result —— 不再编码为 Anthropic 的 tool_result，而是退化为
-            // 普通 user 文本，避免向上游发送“凭空”的工具结果。
-            if (!hasPair) {
-                if (text) {
-                    messages.push({
-                        role: 'user',
-                        content: [{ type: 'text', text }]
-                    });
-                }
-                continue;
-            }
-            const block = {
-                type: 'tool_result',
-                content: text
-            };
-            if (toolCallId) {
-                block.tool_use_id = toolCallId;
-            }
-            messages.push({
-                role: 'user',
-                content: [block]
-            });
-            continue;
-        }
-        const blocks = [];
-        if (text) {
-            blocks.push({ type: 'text', text });
-        }
-        // 将 Chat 段中的 tool_calls 映射回 Anthropic 的 tool_use block，保持
-        // Anthropic→Chat→Anthropic 闭环时 tool_use 的数量和分组不变：
-        //   - role 保持为 assistant；
-        //   - 每个 tool_call 生成一个 {type:'tool_use', id,name,input} block。
-        const toolCalls = Array.isArray(m.tool_calls) ? m.tool_calls : [];
-        for (const tc of toolCalls) {
-            if (!tc || typeof tc !== 'object')
-                continue;
-            const id = String(tc.id || `call_${Math.random().toString(36).slice(2, 10)}`);
-            const fn = tc.function || {};
-            const name = typeof fn.name === 'string' ? String(fn.name) : 'tool';
-            const argsRaw = fn.arguments;
-            let input;
-            if (typeof argsRaw === 'string') {
-                try {
-                    input = JSON.parse(argsRaw);
-                }
-                catch {
-                    input = { _raw: argsRaw };
-                }
-            }
-            else {
-                input = (argsRaw ?? {});
-            }
-            blocks.push({ type: 'tool_use', id, name, input });
-        }
-        if (blocks.length > 0) {
-            messages.push({ role, content: blocks });
-        }
-    }
-    const toolsArr = [];
-    const tools = Array.isArray(chatReq?.tools) ? chatReq.tools : [];
-    for (const t of tools) {
-        if (!t || typeof t !== 'object')
-            continue;
-        const fn = t.function || {};
-        const chatName = typeof fn.name === 'string' ? fn.name : undefined;
-        const anthName = !chatName && typeof t.name === 'string' ? String(t.name) : undefined;
-        // Chat 形状：tools[].function.parameters
-        if (chatName) {
-            const description = typeof fn.description === 'string' ? fn.description : undefined;
-            const params = fn.parameters || {};
-            const input_schema = {
-                type: 'object',
-                properties: params.properties || {}
-            };
-            if (Array.isArray(params.required)) {
-                input_schema.required = params.required;
-            }
-            if (typeof params.additionalProperties === 'boolean') {
-                input_schema.additionalProperties = params.additionalProperties;
-            }
-            const toolDef = { name: chatName, input_schema };
-            if (description)
-                toolDef.description = description;
-            toolsArr.push(toolDef);
-            continue;
-        }
-        // 已是 Anthropic 形状：{ name, input_schema }
-        if (anthName) {
-            const description = typeof t.description === 'string' ? String(t.description) : undefined;
-            const schemaSrc = t.input_schema;
-            let input_schema;
-            if (schemaSrc && typeof schemaSrc === 'object') {
-                try {
-                    input_schema = JSON.parse(JSON.stringify(schemaSrc));
-                }
-                catch {
-                    input_schema = schemaSrc;
-                }
-            }
-            else {
-                input_schema = { type: 'object', properties: {}, additionalProperties: true };
-            }
-            const toolDef = { name: anthName, input_schema };
-            if (description)
-                toolDef.description = description;
-            toolsArr.push(toolDef);
-        }
-    }
-    const out = { model };
-    // 从 ChatRequest 中恢复 Anthropic system[]（某些 Anthropic 兼容端要求 messages[*].role 只能是 user/assistant）
-    if (systemBlocks.length) {
-        out.system = systemBlocks;
-    }
-    out.messages = messages;
-    if (toolsArr.length > 0) {
-        out.tools = toolsArr;
-    }
-    // 透传 metadata（如果存在），避免上游诊断/路由信息丢失
-    try {
-        if (chatReq && typeof chatReq === 'object' && chatReq.metadata && typeof chatReq.metadata === 'object') {
-            out.metadata = JSON.parse(JSON.stringify(chatReq.metadata));
-        }
-    }
-    catch {
-        // best-effort
-    }
-    // 参数映射：尽量保持与 AnthropicInputNode.extractParameters 对称
-    const mt = Number(chatReq?.max_tokens ?? chatReq?.maxTokens ?? NaN);
-    if (Number.isFinite(mt) && mt > 0)
-        out.max_tokens = mt;
-    if (typeof chatReq?.temperature === 'number')
-        out.temperature = chatReq.temperature;
-    if (typeof chatReq?.top_p === 'number')
-        out.top_p = chatReq.top_p;
-    if (typeof chatReq?.stream === 'boolean')
-        out.stream = chatReq.stream;
-    const stop = chatReq?.stop;
-    if (typeof stop === 'string' && stop.trim()) {
-        out.stop_sequences = [stop.trim()];
-    }
-    else if (Array.isArray(stop) && stop.length > 0) {
-        out.stop_sequences = stop.map((s) => String(s)).filter(Boolean);
-    }
-    return out;
-}
+import { buildAnthropicFromOpenAIChat, buildOpenAIChatFromAnthropic, mapAnthropicToolsToChat } from '../shared/anthropic-message-utils.js';
+export { buildAnthropicFromOpenAIChat, buildOpenAIChatFromAnthropic };
+export { buildAnthropicRequestFromOpenAIChat } from '../shared/anthropic-message-utils.js';
 export class AnthropicOpenAIConversionCodec {
     _dependencies;
     id = 'anthropic-openai';
@@ -475,26 +13,32 @@ export class AnthropicOpenAIConversionCodec {
     }
     async initialize() { this.initialized = true; }
     async convertRequest(payload, _profile, context) {
-        if (!this.initialized)
+        if (!this.initialized) {
             await this.initialize();
+        }
         const model = String(payload?.model || 'unknown');
         const { messages } = buildOpenAIChatFromAnthropic(payload);
         const out = { model, messages };
         // 最小必要的形状转换：Anthropic tools → OpenAI Chat function 工具
         try {
-            if (Array.isArray(payload?.tools)) {
-                out.tools = mapAnthropicToolsToOpenAI(payload.tools);
+            const normalizedTools = mapAnthropicToolsToChat(payload?.tools);
+            if (normalizedTools) {
+                out.tools = normalizedTools;
             }
         }
         catch { /* keep tools undefined on failure */ }
-        if (Array.isArray(payload?.stop_sequences))
+        if (Array.isArray(payload?.stop_sequences)) {
             out.stop = payload.stop_sequences;
-        if (typeof payload?.temperature === 'number')
+        }
+        if (typeof payload?.temperature === 'number') {
             out.temperature = payload.temperature;
-        if (typeof payload?.top_p === 'number')
+        }
+        if (typeof payload?.top_p === 'number') {
             out.top_p = payload.top_p;
-        if (typeof payload?.max_tokens === 'number')
+        }
+        if (typeof payload?.max_tokens === 'number') {
             out.max_tokens = payload.max_tokens;
+        }
         // 不设置默认 tool_choice；统一由 Chat 后半段治理
         // 统一流控：非流式由核心过滤器处理；此处不设置 stream
         // 统一走 Chat 请求过滤链路，使 /v1/messages 与 /v1/chat 在工具治理上保持一致
@@ -507,8 +51,9 @@ export class AnthropicOpenAIConversionCodec {
         return runStandardChatRequestFilters(out, _profile, ctxForFilters);
     }
     async convertResponse(payload, _profile, context) {
-        if (!this.initialized)
+        if (!this.initialized) {
             await this.initialize();
+        }
         // 如果是 /v1/messages 且请求端声明了流式（或需要伪流式），合成 Anthropic SSE 流并透传给上层
         // 对非流式 JSON 响应，先在 OpenAI Chat canonical 上运行统一的响应工具治理，
         // 再映射回 Anthropic Messages 形状，保证与 /v1/chat 路径的行为一致。