@goodandready/dsh-image-gen 0.8.7 → 0.8.8

package/lib/index.js

@@ -16,6 +16,7 @@ import z from '@deepseek-ai/schemastery'
 import { defineTool } from '@deepseek-ai/dsh-tools'
 import { credentialRef } from '@deepseek-ai/dsh-credentials'
 import { mkdir, readFile, writeFile } from 'node:fs/promises'
+import os from 'node:os'
 import path from 'node:path'
 import {
   IMAGE_SIZES,
@@ -128,6 +129,22 @@ export const Config = z.object({
     .number()
     .description('provider=local: CFG scale.')
     .default(7),
+  seedreamKeyEnv: z
+    .string()
+    .description('provider=seedream: credential reference / env var holding the API key.')
+    .default('SEEDREAM_API_KEY'),
+  seedreamModel: z
+    .string()
+    .description('provider=seedream: model id, e.g. seedream-4.0.')
+    .default('seedream-4.0'),
+  geminiKeyEnv: z
+    .string()
+    .description('provider=gemini: credential reference / env var holding the Google API key.')
+    .default('GEMINI_API_KEY'),
+  geminiModel: z
+    .string()
+    .description('provider=gemini: model id, e.g. gemini-2.0-flash-exp-image-generation.')
+    .default('gemini-2.0-flash-exp-image-generation'),
   outputDir: z
     .string()
     .description('Where generated images are saved. A relative path resolves against the session working directory; an absolute path is used as given.')
@@ -148,6 +165,14 @@ export const Config = z.object({
     .number()
     .description('Only enhance prompts shorter than this many characters.')
     .default(200),
+  stylePreset: z
+    .string()
+    .description('Optional style suffix appended to the prompt before generation. Empty (default) means no style is applied and the prompt is used as-is.')
+    .default(''),
+  cacheBySeed: z
+    .boolean()
+    .description('If the same seed+prompt was already generated (present in history), return the cached result instead of generating again. Off by default.')
+    .default(false),
 })
 
 /** Keep a file stem safe for the filesystem. */
@@ -214,8 +239,52 @@ function migrateLegacySettings(sctx, scope) {
   }
 }
 
-/** In-memory история генераций (пополняется в execute, где доступен cwd). */
-const history = []
+/** Каталог истории: общий для всех устройств, переживает рестарт. */
+export function historyFile() {
+  return path.join(process.env.DSH_HOME || path.join(os.homedir(), '.dsh'), 'dsh-image-gen', 'history.json')
+}
+
+/** Прочитать историю из файла (пусто, если файла нет). */
+/** Найти запись истории по seed+prompt, если файл ещё существует. */
+export async function findCached(entries, seed, prompt) {
+  if (seed === undefined) return undefined
+  return entries.find((e) => e.seed === seed && e.prompt === prompt)
+}
+
+/** Вернуть запись из кэша, если файл существует; иначе undefined. */
+export async function cachedResult(entry) {
+  if (!entry || !entry.path) return undefined
+  try { require('node:fs').accessSync(entry.path) } catch (e) { return undefined }
+  return {
+    path: entry.path,
+    url: entry.attachmentId ? `/dsh-image-gen/image?id=${encodeURIComponent(entry.attachmentId)}` : '',
+    width: 0,
+    height: 0,
+    seed: entry.seed,
+    prompt: entry.prompt,
+    format: 'png',
+    cached: true,
+    attachment: entry.attachmentId ? { attachmentId: entry.attachmentId, mediaType: 'image/png', bytes: 0, width: 0, height: 0, name: '' } : undefined,
+  }
+}
+
+export async function readHistory() {
+  try {
+    const raw = await readFile(historyFile(), 'utf-8')
+    const parsed = JSON.parse(raw)
+    return Array.isArray(parsed) ? parsed : []
+  } catch (e) {
+    return []
+  }
+}
+
+/** Записать историю в файл (перезапись целиком). */
+export async function writeHistory(entries) {
+  try {
+    await mkdir(path.dirname(historyFile()), { recursive: true })
+    await writeFile(historyFile(), JSON.stringify(entries, null, 2))
+  } catch (e) { /* история не критична */ }
+}
 
 /** Отфильтровать записи, чьи файлы ещё существуют; новые первыми. */
 export function filterHistory(entries, exists) {
@@ -349,8 +418,13 @@ export function apply(ctx, config) {
         return
       }
       const exists = (p) => { try { return require('node:fs').existsSync(p) } catch { return false } }
+      const entries = await readHistory()
+      const withThumbs = filterHistory(entries, exists).map((e) => ({
+        ...e,
+        thumbnailUrl: e.attachmentId ? `/dsh-image-gen/image?id=${encodeURIComponent(e.attachmentId)}` : '',
+      }))
       res.writeHead(200, { 'Content-Type': 'application/json' })
-      res.end(JSON.stringify(filterHistory(history, exists)))
+      res.end(JSON.stringify(withThumbs))
     },
   }), 'dsh-image-gen: history route')
 
@@ -390,6 +464,11 @@ export function apply(ctx, config) {
           type: 'integer',
           description: 'Number of variations to generate in one call, 1-4 (default 1). Cost scales with count.',
         },
+        prompts: {
+          type: 'array',
+          items: { type: 'string' },
+          description: 'Optional list of prompts; one image is generated per prompt. When set, overrides count and the single prompt.',
+        },
         negative_prompt: {
           type: 'string',
           description: 'Optional text describing what NOT to draw; sent to providers that support it (ignored otherwise).',
@@ -398,6 +477,11 @@ export function apply(ctx, config) {
           type: 'number',
           description: 'Optional prompt adherence (e.g. 1-20); sent to providers that support it (ignored otherwise).',
         },
+        quality: {
+          type: 'string',
+          enum: ['low', 'medium', 'high'],
+          description: 'Optional quality for providers that support it (seedream, gemini). Ignored otherwise.',
+        },
         source_image: {
           type: 'string',
           description: 'Optional path to an image file or an attachment id (sha256:...) to edit instead of drawing from scratch. Only providers that support image editing accept it; others refuse with a clear reason.',
@@ -488,22 +572,23 @@ export function apply(ctx, config) {
         const source = args.source_image ? await resolveSource(ctx, exec, args.source_image) : undefined
         const mask = args.mask ? await resolveSource(ctx, exec, args.mask) : undefined
         const enhanced = await enhancePrompt(ctx, cfg, args.prompt, exec.signal)
-        const effectivePrompt = enhanced.prompt
+        const styleSuffix = cfg.stylePreset ? `, ${cfg.stylePreset}` : ''
+        const effectivePrompt = enhanced.prompt + styleSuffix
         const providers = makeProviders(
           { fetchImpl: fetch, resolveKey: (ref) => resolveApiKey(ctx, ref), cfg, subscriptionImages },
-          { prompt: effectivePrompt, size, format, seed: args.seed, signal: exec.signal, negativePrompt: args.negative_prompt, guidanceScale: args.guidance_scale, source, mask, strength: args.strength },
+          { prompt: effectivePrompt, size, format, seed: args.seed, signal: exec.signal, negativePrompt: args.negative_prompt, guidanceScale: args.guidance_scale, source, mask, strength: args.strength, quality: args.quality },
         )
         // Подписочные провайдеры (codex/grok) отдают {ok:false, reason} вместо исключения:
         // отказ должен дойти до модели текстом. Без проверки execute шёл дальше с пустыми
         // байтами, и пользователь получал битую карточку вместо внятного отказа.
         const guard = (generated) => { if (generated && generated.ok === false) throw new Error(generated.reason) }
-        const one = async (jobSeed, providerKey) => {
-          const gen = await providers[providerKey](jobSeed)
+        const one = async (jobSeed, providerKey, promptArg = effectivePrompt) => {
+          const gen = await providers[providerKey](jobSeed, promptArg)
           guard(gen)
           const bytes = gen.bytes
           const mediaType = gen.mediaType
           const extension = mediaType === 'image/jpeg' ? 'jpg' : mediaType === 'image/webp' ? 'webp' : 'png'
-          const stem = `${slugify(args.output_name || effectivePrompt)}-${Date.now().toString(36)}-${jobSeed}`
+          const stem = `${slugify(args.output_name || promptArg)}-${Date.now().toString(36)}-${jobSeed}`
           const name = `${stem}.${extension}`
 
           const attachment = await ctx.attachments.saveImage({
@@ -529,7 +614,7 @@ export function apply(ctx, config) {
           await writeFile(
             path.join(outDir, `${stem}.json`),
             JSON.stringify(buildSidecar({
-              prompt: effectivePrompt,
+              prompt: promptArg,
               size,
               format,
               seed: gen.seed,
@@ -544,17 +629,20 @@ export function apply(ctx, config) {
             }), null, 2),
           )
 
-          history.unshift({
+          const entry = {
             path: filePath,
-            prompt: effectivePrompt,
+            prompt: promptArg,
             provider,
             size,
             format,
             seed: gen.seed,
             createdAt: new Date().toISOString(),
             attachmentId: attachment.attachmentId,
-          })
-          if (history.length > (cfg.historyLimit || 50)) history.length = cfg.historyLimit || 50
+          }
+          const current = await readHistory()
+          current.unshift(entry)
+          if (current.length > (cfg.historyLimit || 50)) current.length = cfg.historyLimit || 50
+          await writeHistory(current)
 
           return {
             path: filePath,
@@ -562,7 +650,7 @@ export function apply(ctx, config) {
             width: gen.width || attachment.width,
             height: gen.height || attachment.height,
             seed: gen.seed,
-            prompt: effectivePrompt,
+            prompt: promptArg,
             originalPrompt: enhanced.enhanced ? args.prompt : undefined,
             cost: gen.cost,
             format: mediaType.replace('image/', ''),
@@ -577,15 +665,25 @@ export function apply(ctx, config) {
           }
         }
 
-        const count = normalizeCount(args.count)
         const seedBase = args.seed ?? Math.floor(Math.random() * 100000)
         // Fallback-цепочка: пробуем текущий провайдер, при отказе — следующий
         // по порядку (fal → custom → codex → grok), собирая причины отказов.
         const order = fallbackOrder(provider)
-        const generators = Object.fromEntries(PROVIDER_KEYS.map((k) => [k, (s) => one(s, k)]))
+        const generators = Object.fromEntries(PROVIDER_KEYS.map((k) => [k, (s, p) => one(s, k, p)]))
         const images = []
-        for (let i = 0; i < count; i += 1) {
-          images.push(await tryGenerate(generators, order, seedBase + i))
+        const historyEntries = cfg.cacheBySeed ? await readHistory() : []
+        const batchPrompts = Array.isArray(args.prompts) && args.prompts.length ? args.prompts : null
+        if (batchPrompts) {
+          for (let i = 0; i < batchPrompts.length; i += 1) {
+            const cached = cfg.cacheBySeed ? await cachedResult(findCached(historyEntries, seedBase + i, batchPrompts[i])) : undefined
+            images.push(cached || await tryGenerate(generators, order, seedBase + i, batchPrompts[i]))
+          }
+        } else {
+          const count = normalizeCount(args.count)
+          for (let i = 0; i < count; i += 1) {
+            const cached = cfg.cacheBySeed ? await cachedResult(findCached(historyEntries, seedBase + i, effectivePrompt)) : undefined
+            images.push(cached || await tryGenerate(generators, order, seedBase + i))
+          }
         }
         const first = images[0]
         return { ...first, images }

package/lib/providers.js

@@ -7,7 +7,7 @@
 // Сеть приходит параметром (fetchImpl), ключ — через resolveKey, поэтому оба
 // провайдера проверяются юнит-тестами без единого реального запроса.
 
-export const PROVIDER_KEYS = ['fal', 'custom', 'codex', 'grok', 'local']
+export const PROVIDER_KEYS = ['fal', 'custom', 'codex', 'grok', 'local', 'seedream', 'gemini']
 
 /** Привести count из аргумента инструмента к диапазону 1..4. */
 /** Порядок провайдеров для fallback: основной первым, остальные по PROVIDER_KEYS. */
@@ -21,11 +21,11 @@ export function fallbackOrder(primary) {
  * @param generators - массив функций (key, seed) => Promise<generated>.
  * @param order - порядок ключей, длина = generators.length.
  */
-export async function tryGenerate(generators, order, seed) {
+export async function tryGenerate(generators, order, seed, promptArg) {
   const refusals = []
   for (const key of order) {
     try {
-      const produced = await generators[key](seed)
+      const produced = await generators[key](seed, promptArg)
       return produced
     } catch (e) {
       refusals.push(`${key}: ${e.message || String(e)}`)
@@ -179,11 +179,11 @@ export function buildEditForm({ source, mask, prompt, size, strength }) {
 
 export function makeProviders(deps, job) {
   const { fetchImpl, resolveKey, cfg } = deps
-  const { prompt, size, format, seed, signal, negativePrompt, guidanceScale, source, mask, strength } = job
+  const { prompt, size, format, seed, signal, negativePrompt, guidanceScale, source, mask, strength, quality } = job
 
-  async function fal(seedArg = seed) {
+  async function fal(seedArg = seed, promptArg = prompt) {
     const key = await resolveKey(cfg.apiKeyEnv)
-    const body = { prompt, image_size: size, num_images: 1 }
+    const body = { prompt: promptArg, image_size: size, num_images: 1 }
     if (seedArg !== undefined) body.seed = seedArg
     if (negativePrompt !== undefined) body.negative_prompt = negativePrompt
     if (guidanceScale !== undefined) body.guidance_scale = guidanceScale
@@ -218,7 +218,7 @@ export function makeProviders(deps, job) {
   }
 
   // Любой OpenAI-совместимый API картинок. Один запрос вместо очереди FAL.
-  async function custom(seedArg = seed) {
+  async function custom(seedArg = seed, promptArg = prompt) {
     const base = String(cfg.customBaseURL || '').replace(/\/+$/, '')
     if (!base) throw new Error('Custom image provider: base URL is not configured (Settings → Image generation)')
     if (!cfg.customModel) throw new Error('Custom image provider: model is not configured')
@@ -236,10 +236,10 @@ export function makeProviders(deps, job) {
       method: 'POST',
       headers,
       body: source
-        ? buildEditForm({ source, mask, prompt, size: cfg.customSize || SIZE_PIXELS[size] || size, strength })
+        ? buildEditForm({ source, mask, prompt: promptArg, size: cfg.customSize || SIZE_PIXELS[size] || size, strength })
         : JSON.stringify({
             model: cfg.customModel,
-            prompt,
+            prompt: promptArg,
             n: 1,
             size: cfg.customSize || SIZE_PIXELS[size] || size,
             ...(negativePrompt !== undefined ? { negative_prompt: negativePrompt } : {}),
@@ -291,7 +291,7 @@ export function makeProviders(deps, job) {
   // только запрос и разбор ответа. Токен сюда не попадает вовсе — служба
   // отдаёт готовую картинку, а не ключ доступа.
   function subscription(provider) {
-    return async function generate(seedArg = seed) {
+    return async function generate(seedArg = seed, promptArg = prompt) {
       if (source) {
         return { ok: false, provider, reason: `${provider}: не умеет править изображения — используйте fal, custom или local` }
       }
@@ -307,7 +307,7 @@ export function makeProviders(deps, job) {
       try {
         produced = await images.generate({
           provider,
-          prompt,
+          prompt: promptArg,
           size: SUBSCRIPTION_SIZES[size] || '1024x1024',
           quality: cfg.subscriptionQuality || undefined,
           signal,
@@ -333,7 +333,7 @@ export function makeProviders(deps, job) {
   }
 
   // Локальная генерация: ComfyUI (очередь + опрос) или Automatic1111 (txt2img).
-  async function local(seedArg = seed) {
+  async function local(seedArg = seed, promptArg = prompt) {
     const base = String(cfg.localBaseURL || '').replace(/\/+$/, '')
     if (!base) throw new Error('Local image provider: server address is not configured (Settings → Image generation)')
     const [width, height] = sizeToPixels(size)
@@ -341,7 +341,7 @@ export function makeProviders(deps, job) {
 
     if (kind === 'a1111') {
       const body = {
-        prompt,
+        prompt: promptArg,
         negative_prompt: negativePrompt,
         width,
         height,
@@ -379,7 +379,7 @@ export function makeProviders(deps, job) {
         '3': { class_type: 'KSampler', inputs: { seed: seedArg ?? 0, steps: cfg.localSteps ?? 20, cfg: cfg.localCfg ?? 7, sampler_name: 'euler', scheduler: 'normal', denoise: 1, model: ['4', 0], positive: ['6', 0], negative: ['7', 0], latent_image: ['5', 0] } },
         '4': { class_type: 'CheckpointLoaderSimple', inputs: { ckpt_name: cfg.localModel || 'v1-5-pruned-emaonly.safetensors' } },
         '5': { class_type: 'EmptyLatentImage', inputs: { width, height, batch_size: 1 } },
-        '6': { class_type: 'CLIPTextEncode', inputs: { text: prompt, clip: ['4', 1] } },
+        '6': { class_type: 'CLIPTextEncode', inputs: { text: promptArg, clip: ['4', 1] } },
         '7': { class_type: 'CLIPTextEncode', inputs: { text: negativePrompt || '', clip: ['4', 1] } },
         '8': { class_type: 'VAEDecode', inputs: { samples: ['3', 0], vae: ['4', 2] } },
         '9': { class_type: 'SaveImage', inputs: { filename_prefix: 'dsh', images: ['8', 0] } },
@@ -429,5 +429,61 @@ export function makeProviders(deps, job) {
     }
   }
 
-  return { fal, custom, codex: subscription('codex'), grok: subscription('grok'), local }
+  // Seedream (ByteDance): OpenAI-совместимый API картинок.
+  async function seedream(seedArg = seed, promptArg = prompt) {
+    const key = await resolveKey(cfg.seedreamKeyEnv)
+    const base = 'https://api.bytedanceapi.com/v1'
+    const res = await fetchImpl(`${base}/images/generations`, {
+      method: 'POST',
+      headers: { 'Content-Type': 'application/json', Authorization: `Bearer ${key}` },
+      body: JSON.stringify({
+        model: cfg.seedreamModel || 'seedream-4.0',
+        prompt: promptArg,
+        n: 1,
+        size: SIZE_PIXELS[size] || size,
+        ...(negativePrompt !== undefined ? { negative_prompt: negativePrompt } : {}),
+        ...(quality !== undefined ? { quality } : {}),
+      }),
+      signal,
+    })
+    const data = await res.json().catch(() => ({}))
+    if (!res.ok) {
+      const detail = data?.error?.message || JSON.stringify(data).slice(0, 600)
+      throw new Error(`Seedream failed (HTTP ${res.status}): ${detail}`)
+    }
+    const item = data?.data?.[0]
+    if (!item) throw new Error('Seedream returned no images')
+    if (item.b64_json) {
+      return { bytes: Buffer.from(item.b64_json, 'base64'), mediaType: normalizeMediaType('image/png', format), width: 0, height: 0, seed: seedArg ?? 0, sourceUrl: '' }
+    }
+    if (!item.url) throw new Error('Seedream returned neither b64_json nor url')
+    const dl = await fetchImpl(item.url, { signal })
+    if (!dl.ok) throw new Error(`Seedream download failed (HTTP ${dl.status})`)
+    return { bytes: Buffer.from(await dl.arrayBuffer()), mediaType: normalizeMediaType('image/png', format), width: 0, height: 0, seed: seedArg ?? 0, sourceUrl: item.url }
+  }
+
+  // Gemini (Google): images.generate через GenAI API.
+  async function gemini(seedArg = seed, promptArg = prompt) {
+    const key = await resolveKey(cfg.geminiKeyEnv)
+    const model = cfg.geminiModel || 'gemini-2.0-flash-exp-image-generation'
+    const res = await fetchImpl(`https://generativelanguage.googleapis.com/v1beta/models/${model}:generateContent?key=${encodeURIComponent(key)}`, {
+      method: 'POST',
+      headers: { 'Content-Type': 'application/json' },
+      body: JSON.stringify({
+        contents: [{ parts: [{ text: promptArg }] }],
+        generationConfig: { responseModalities: ['IMAGE'], ...(quality !== undefined ? { imageConfig: { imageQuality: quality } } : {}) },
+      }),
+      signal,
+    })
+    const data = await res.json().catch(() => ({}))
+    if (!res.ok) {
+      const detail = data?.error?.message || JSON.stringify(data).slice(0, 600)
+      throw new Error(`Gemini failed (HTTP ${res.status}): ${detail}`)
+    }
+    const part = data?.candidates?.[0]?.content?.parts?.find((p) => p.inlineData?.data)
+    if (!part) throw new Error('Gemini returned no image')
+    return { bytes: Buffer.from(part.inlineData.data, 'base64'), mediaType: normalizeMediaType(part.inlineData.mimeType || 'image/png', format), width: 0, height: 0, seed: seedArg ?? 0, sourceUrl: '' }
+  }
+
+  return { fal, custom, codex: subscription('codex'), grok: subscription('grok'), local, seedream, gemini }
 }

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@goodandready/dsh-image-gen",
-  "version": "0.8.7",
+  "version": "0.8.8",
   "description": "Image generation for DeepSeek Harness: a generate_image tool with pluggable providers — the FAL queue, any OpenAI-compatible images API, or a ChatGPT/Grok subscription with no API key at all. The picture is shown inline in the conversation; the model receives either a link (works with any chat model) or the image itself (needs dsh-vision-bridge or a vision-capable model).",
   "keywords": [
     "deepseek-harness",