File size: 12,015 Bytes
e2b3a17
 
fff1d2d
017e9c7
7486fe7
fc9d15a
e2b3a17
 
fc9d15a
e2b3a17
 
 
3937110
e2b3a17
 
3937110
e2b3a17
7486fe7
 
4b541a7
7486fe7
aacba39
7486fe7
 
 
aacba39
 
7486fe7
 
4b541a7
7486fe7
e2b3a17
7486fe7
 
e2b3a17
4b541a7
e2b3a17
7486fe7
 
 
 
 
4b541a7
 
7486fe7
017e9c7
7486fe7
 
017e9c7
4b541a7
7486fe7
 
 
4b541a7
e2b3a17
7486fe7
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
4b541a7
7486fe7
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
f970c89
7486fe7
 
4b541a7
 
e2b3a17
4b541a7
7486fe7
1bec76b
e2b3a17
7486fe7
 
 
 
 
1bec76b
4b541a7
e2b3a17
7486fe7
 
 
e2b3a17
b2b6570
e2b3a17
7486fe7
 
 
 
 
 
 
 
e2b3a17
b2b6570
dbc06fe
7486fe7
 
 
 
 
 
 
 
 
dbc06fe
 
e2b3a17
7486fe7
 
 
 
 
 
 
 
 
 
aacba39
 
 
7486fe7
 
 
 
 
 
 
 
 
 
 
 
e2b3a17
b2b6570
7486fe7
e2b3a17
7486fe7
 
b2b6570
7486fe7
 
 
f970c89
 
7486fe7
 
 
e2b3a17
4469357
7486fe7
f970c89
7486fe7
 
f970c89
 
7486fe7
 
 
 
 
f970c89
7486fe7
 
f970c89
7486fe7
 
 
 
f970c89
7486fe7
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
f970c89
7486fe7
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
f970c89
 
34eae9c
7486fe7
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
import cors from 'cors';
import dotenv from 'dotenv';
import express from 'express';
import { createClient } from '@supabase/supabase-js';
import { BedrockRuntimeClient, ConverseStreamCommand } from "@aws-sdk/client-bedrock-runtime";
import { NodeHttpHandler } from "@smithy/node-http-handler";
import path from 'path';
import { fileURLToPath } from 'url';

dotenv.config();
const app = express();
const PORT = process.env.PORT || 7860;

const __filename = fileURLToPath(import.meta.url);
const __dirname = path.dirname(__filename);

app.use(cors());
app.use(express.json({ limit: '50mb' }));
app.use(express.static(path.join(__dirname, 'public'))); 

// --- LOGGER HELPER ---
const log = {
    info: (msg, id = "SYS") => console.log(`[${new Date().toISOString()}] [INFO] [${id}] ${msg}`),
    warn: (msg, id = "SYS") => console.warn(`[${new Date().toISOString()}] ⚠️ [WARN][${id}] ${msg}`),
    error: (msg, err, id = "SYS") => console.error(`[${new Date().toISOString()}] ❌ [ERROR] [${id}] ${msg}`, err?.message || err, err?.stack || "")
};

// --- SYSTEM PROMPT DEFINITIONS ---
const CLAUDE_SYSTEM_PROMPT = "You are a pro. Provide elite, high-level technical responses.";

// --- AI CLIENTS ---
const bedrockClient = new BedrockRuntimeClient({
    region: "us-east-1",
    requestHandler: new NodeHttpHandler({ http2Handler: undefined }) 
});

function getBedrockModelId(modelName) {
    switch(modelName) {
        case "haiku": return "arn:aws:bedrock:us-east-1:106774395747:inference-profile/global.anthropic.claude-haiku-4-5-20251001-v1:0";
        case "maverick": return "arn:aws:bedrock:us-east-1:106774395747:inference-profile/us.meta.llama4-maverick-17b-instruct-v1:0";
        case "claude": default: return "arn:aws:bedrock:us-east-1:106774395747:inference-profile/global.anthropic.claude-sonnet-4-6";
    }
}

// --- DB & MEMORY MANAGEMENT (SUPABASE) ---
const supabase = createClient(
    process.env.SUPABASE_URL || '',
    process.env.SUPABASE_KEY || ''
);

let memoryChats = {}; 
let dirtyChats = new Set(); 
const activeGenerations = new Map(); 

async function initDB() {
    try {
        log.info("Connecting to Supabase...");
        const { data: dbChats, error } = await supabase.from('chats').select('*');
        if (error) throw error;

        if (dbChats) {
            dbChats.forEach(c => {
                memoryChats[c.id] = { 
                    ...c, 
                    inputTokens: c.inputTokens || 0,
                    outputTokens: c.outputTokens || 0,
                    isGenerating: false 
                };
            });
            log.info(`Hydrated ${dbChats.length} chats from DB.`);
        }

        setInterval(async () => {
            if (dirtyChats.size === 0) return;
            const toSync = Array.from(dirtyChats);
            dirtyChats.clear();
            
            const rowsToUpsert = toSync.map(id => {
                const chat = memoryChats[id];
                chat.updatedAt = new Date().toISOString();
                return {
                    id: chat.id, title: chat.title, totalTokens: chat.totalTokens,
                    inputTokens: chat.inputTokens, outputTokens: chat.outputTokens,
                    messages: chat.messages, updatedAt: chat.updatedAt
                };
            });

            if (rowsToUpsert.length > 0) {
                const { error } = await supabase.from('chats').upsert(rowsToUpsert);
                if (error) log.error(`Supabase Sync Error.`, error);
                else log.info(`Synced ${rowsToUpsert.length} chats to Supabase.`);
            }
        }, 15000);

    } catch (err) {
        log.error('Supabase Initialization Error.', err);
    }
}
initDB();

// --- API ENDPOINTS ---

app.get('/api/chats', (req, res) => {
    const chatsList = Object.values(memoryChats).map(c => ({
        id: c.id, title: c.title, totalTokens: c.totalTokens, 
        inputTokens: c.inputTokens, outputTokens: c.outputTokens, updatedAt: c.updatedAt
    })).sort((a, b) => new Date(b.updatedAt) - new Date(a.updatedAt));
    res.json(chatsList);
});

app.get('/api/chats/:id', (req, res) => {
    const chat = memoryChats[req.params.id];
    if (!chat) return res.status(404).json({ error: "Chat not found" });
    res.json(chat);
});

app.post('/api/chats', (req, res) => {
    const newId = Date.now().toString();
    memoryChats[newId] = {
        id: newId, title: "New Chat", totalTokens: 0, inputTokens: 0, outputTokens: 0,
        messages:[], isGenerating: false, updatedAt: new Date().toISOString()
    };
    dirtyChats.add(newId);
    log.info("Created new chat.", newId);
    res.json(memoryChats[newId]);
});

app.put('/api/chats/:id/title', (req, res) => {
    const { id } = req.params;
    const { title } = req.body;
    if (!memoryChats[id]) return res.status(404).json({ error: "Chat not found" });
    if (!title || typeof title !== 'string') return res.status(400).json({ error: "Invalid title" });
    
    memoryChats[id].title = title.trim();
    dirtyChats.add(id);
    log.info(`Title updated to: "${title.trim()}"`, id);
    res.json({ success: true, title: memoryChats[id].title });
});

app.delete('/api/chats/:id', async (req, res) => {
    const { id } = req.params;
    if (activeGenerations.has(id)) {
        activeGenerations.get(id).abort();
        activeGenerations.delete(id);
    }
    delete memoryChats[id];
    dirtyChats.delete(id);
    await supabase.from('chats').delete().eq('id', id);
    log.info("Deleted chat permanently.", id);
    res.json({ success: true });
});

app.post('/api/chats/:id/stop', (req, res) => {
    const { id } = req.params;
    log.info("User requested to stop generation.", id);
    
    if (activeGenerations.has(id)) {
        activeGenerations.get(id).abort();
        activeGenerations.delete(id);
    }
    if (memoryChats[id]) {
        memoryChats[id].isGenerating = false;
        dirtyChats.add(id);
    }
    res.json({ success: true });
});

// --- STREAM ENDPOINT ---
app.post('/api/chats/:id/stream', async (req, res) => {
    const { id } = req.params;
    const { model, prompt, system_prompt, images } = req.body; 

    if (!memoryChats[id]) return res.status(404).send("Chat not found");
    if (!prompt || typeof prompt !== 'string' || prompt.trim() === '') {
        return res.status(400).send("Prompt cannot be empty");
    }

    if (memoryChats[id].isGenerating) {
        log.warn("Attempted concurrent generation. Rejecting request.", id);
        return res.status(409).json({ error: "Chat is currently generating." });
    }

    log.info(`Starting stream. Model: ${model} | Prompt length: ${prompt.length} | Images: ${images?.length || 0}`, id);

    if (memoryChats[id].messages.length === 0 && memoryChats[id].title === "New Chat") {
        memoryChats[id].title = prompt.substring(0, 30) + (prompt.length > 30 ? '...' : '');
    }

    memoryChats[id].messages.push({ role: "user", content: prompt });
    const aiMessage = { role: "assistant", content: "", reasoning: "" };
    memoryChats[id].messages.push(aiMessage);
    memoryChats[id].isGenerating = true;
    dirtyChats.add(id);

    const abortController = new AbortController();
    activeGenerations.set(id, abortController);

    res.setHeader('Content-Type', 'text/plain; charset=utf-8');
    res.setHeader('Transfer-Encoding', 'chunked');
    res.setHeader('X-Accel-Buffering', 'no'); 
    res.flushHeaders();

    const safeWrite = (data) => {
        if (!req.socket.destroyed && !res.writableEnded) {
            try { res.write(data); } catch (e) { log.warn("Socket disconnected during write.", id); }
        }
    };
    const safeEnd = () => {
        if (!req.socket.destroyed && !res.writableEnded) {
            try { res.end(); } catch (e) {}
        }
    };

    let streamInputTokens = 0;
    let streamOutputTokens = 0;
    let streamTotalTokens = 0;

    try {
        const bedrockModelId = getBedrockModelId(model);
        let contentBlock = [{ text: prompt }];
        
        if (images && images.length > 0) {
            const imageBlocks = images.map(imgStr => {
                const base64Data = imgStr.replace(/^data:image\/\w+;base64,/, "");
                return { image: { format: 'png', source: { bytes: Buffer.from(base64Data, 'base64') } } };
            });
            contentBlock =[...imageBlocks, ...contentBlock];
        }

        const historicalMessages = memoryChats[id].messages.slice(0, -2).map(m => {
            let safeText = m.content;
            if (!safeText || safeText.trim() === "") {
                safeText = "[System Note: The model failed to generate a response here previously.]";
            }
            return { role: m.role, content:[{ text: safeText }] };
        });
        
        historicalMessages.push({ role: "user", content: contentBlock });

        // THE FIX: Uncapped Token limit for Claude 3.7 to allow massive reasoning + coding
        const commandMaxTokens = model.includes("claude") ? 64000 : 8192;

        const command = new ConverseStreamCommand({
            modelId: bedrockModelId,
            system:[{ text: system_prompt || CLAUDE_SYSTEM_PROMPT }],
            messages: historicalMessages,
            inferenceConfig: { maxTokens: commandMaxTokens, temperature: 1 },
           /* additionalModelRequestFields: model.includes("claude") ? { 
                thinking: { type: "adaptive" } 
            } : undefined
            */
        });

        const response = await bedrockClient.send(command, { abortSignal: abortController.signal });
        log.info(`Bedrock connected successfully (Max Tokens: ${commandMaxTokens}), streaming chunks...`, id);
        
        for await (const chunk of response.stream) {
            if (chunk.contentBlockDelta) {
                const delta = chunk.contentBlockDelta.delta;
                if (delta.reasoningContent && delta.reasoningContent.text) {
                    aiMessage.reasoning += delta.reasoningContent.text;
                    safeWrite(`__THINK__${delta.reasoningContent.text}`);
                } else if (delta.text) {
                    aiMessage.content += delta.text;
                    safeWrite(delta.text);
                }
            }
            // Log exactly why it stopped (e.g. max_tokens vs end_turn)
            if (chunk.messageStop) {
                log.info(`Stream stopped. Reason: ${chunk.messageStop.stopReason}`, id);
            }
            if (chunk.metadata && chunk.metadata.usage) {
                streamInputTokens = chunk.metadata.usage.inputTokens || 0;
                streamOutputTokens = chunk.metadata.usage.outputTokens || 0;
                streamTotalTokens = streamInputTokens + streamOutputTokens;
            }
        }
        
        log.info(`Stream completed normally. (In: ${streamInputTokens}, Out: ${streamOutputTokens})`, id);

    } catch (err) {
        if (err.name === 'AbortError' || err.name === 'TimeoutError') {
            log.warn("Generation aborted by user or timeout.", id);
            aiMessage.content += "\n\n*[Generation stopped by user]*";
            safeWrite("\n\n*[Generation stopped by user]*");
        } else {
            log.error("Generation failed during stream processing.", err, id);
            aiMessage.content += `\n\n**[Error]**: ${err.message}`;
            safeWrite(`\n\n**ERROR**: ${err.message}`);
        }
    } finally {
        activeGenerations.delete(id);
        
        memoryChats[id].inputTokens += streamInputTokens;
        memoryChats[id].outputTokens += streamOutputTokens;
        memoryChats[id].totalTokens += streamTotalTokens;
        memoryChats[id].isGenerating = false;
        dirtyChats.add(id);
        
        safeWrite(`__USAGE__${JSON.stringify({ 
            inputTokens: streamInputTokens, 
            outputTokens: streamOutputTokens, 
            totalTokens: streamTotalTokens 
        })}`);
        safeEnd();
    }
});

app.listen(PORT, '0.0.0.0', () => log.info(`AI Server live on http://localhost:${PORT}`));