update t2v

.gitattributes

@@ -36,4 +36,6 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.ggml filter=lfs diff=lfs merge=lfs -text
 *.llamafile* filter=lfs diff=lfs merge=lfs -text
 test_imgs/* filter=lfs diff=lfs merge=lfs -text
-demo/gallery/* filter=lfs diff=lfs merge=lfs -text
+demo/gallery/* filter=lfs diff=lfs merge=lfs -text
+resource/* filter=lfs diff=lfs merge=lfs -text
+demo/style_img/* filter=lfs diff=lfs merge=lfs -text

app.py

@@ -1,13 +1,88 @@
 import cv2, os
 import gradio as gr
 import numpy as np
-from demo.generation import call_generation
-from demo.instructions import INSTRUCTIONS_VECTORIZE_SIMPLIFY
+from demo.generation import call_generation, call_generation_t2v
+from demo.instructions import INSTRUCTIONS_VECTORIZE_SIMPLIFY, INSTRUCTIONS_T2V
 
 VERSION = 'v0.1'
 
 GALLERY_LIST = [os.path.join('demo/gallery',path) for path in os.listdir('demo/gallery')]
 
+PROMPT_EXAMPLE=[['a cute dog'],
+    ['a cute cat'],
+    ['a beautiful woman'],
+    ['a bus'],]
+
+STYLE_MAPPING = {"0": "商业扁平插画",
+                 "1": "矢量风扁平插画",
+                 "2": "高级时尚系列插画",
+                 "3": "s矢量插画",
+                 "4": "彩色插画",
+                 "5": "扁平海报插画",
+                 "6": "扁平职人风插画",
+                 "7": "矢量风插画",
+                 "8": "轮廓插画",
+                 "9": "F矢量插画",
+                 "10": "Material图标",
+                 "11": "图标元素",
+                 "12": "矢量风插画0",
+                 "13": "Logo风格图标",
+                 "14": "sticker风图标",
+                 "15": "纹理插画",
+                 "16": "矢量风插画2",
+                 "17": "矢量风插画M"}
+
+STYLE_MAPPING_EN = {
+    "0": "Commercial flat illustration",
+    "1": "Vector-style flat illustration",
+    "2": "High-end fashion series illustration",
+    "3": "s vector illustration",
+    "4": "Color illustration",
+    "5": "Flat poster illustration",
+    "6": "Flat artisan-style illustration",
+    "7": "Vector-style illustration",
+    "8": "Outline illustration",
+    "9": "F vector illustration",
+    "10": "Material icon",
+    "11": "Icon elements",
+    "12": "Vector-style illustration 0",
+    "13": "Logo-style icon",
+    "14": "Sticker-style icon",
+    "15": "Texture illustration",
+    "16": "Vector-style illustration 2",
+    "17": "Vector-style illustration M"
+}
+
+
+STYLE_MAPPING_REVERT = {v: k for k, v in STYLE_MAPPING_EN.items()}
+
+
+STYLE_IMG_DICT = {"demo/style_img/商业扁平插画.webp": "0", 
+                  "demo/style_img/矢量风扁平插画.webp":"1",
+                  "demo/style_img/高级时尚系列插画.webp":"2",
+                  "demo/style_img/s矢量插画.webp": "3", 
+                  "demo/style_img/彩色插画.webp": "4", 
+                  "demo/style_img/扁平海报插画.webp": "5",
+                  "demo/style_img/扁平职人.webp": "6", 
+                  "demo/style_img/vectorart.webp": "7", 
+                  "demo/style_img/outline.webp": "8", 
+                  "demo/style_img/fllust.webp": "9", 
+                  "demo/style_img/icon.webp": "11", 
+                  "demo/style_img/icon_material.webp": "10",
+                  "demo/style_img/last.webp": "12", 
+                  "demo/style_img/logo.webp": "13", 
+                  "demo/style_img/sdmai.webp": "14", 
+                  "demo/style_img/texture.webp": "15", 
+                  "demo/style_img/vector_illustration.webp":"16",
+                  "demo/style_img/vectorizeM.webp": "17", 
+                  }
+
+STYLE_IMG_DICT_REVERT = {v: k for k, v in STYLE_IMG_DICT.items()}
+
+details_slider_dict = {"Minimal 简约":"minimal",
+                       "Medium 中等":"medium",
+                       "Rich 丰富": "rich"
+                    }
 
 def resize_image(image, size):
     # find the minimal size of the image, resize it to size
@@ -55,6 +130,31 @@ def process_vector(input_image, upsample_method, svg_simplify, svg_optimize, tra
         return file_list
 
 
+def process_t2v(prompt, num_imgs, image_resolution_h, image_resolution_w, details_slider, style_slider, vectorize, upsample_method, svg_simplify, svg_optimize, trace_mode, subsample_ratio, speckle_removal,sorting_method, sorting_order, use_gpu):
+    print("Processing t2v:",upsample_method, svg_simplify, svg_optimize, trace_mode)
+    if prompt is not None:
+        ## save input_image to a temp file
+        details = details_slider_dict[details_slider]
+        style = STYLE_MAPPING_REVERT[style_slider]
+        ## process the image
+        file_list = call_generation_t2v(prompt,
+                        num_imgs,
+                        image_resolution_h,
+                        image_resolution_w,
+                        details,
+                        style,
+                        vectorize,
+                        preprocess=upsample_method, 
+                        simplify=svg_simplify, 
+                        optimize=svg_optimize, 
+                        mode=trace_mode, 
+                        subsample_ratio=subsample_ratio, 
+                        speckle_removal=speckle_removal, 
+                        sorting_method=sorting_method,
+                        sorting_order=sorting_order,
+                        use_gpu=use_gpu)
+        
+        return file_list
 
 
 
@@ -85,6 +185,61 @@ with block:
     #     with gr.Row():
     #         gr.Gallery(label='图像生成结果', value=GALLERY_LIST,show_label=False, elem_id="Gallery", columns=5, height=1000)
 
+    tab_1 = gr.Tab(label="TEXT to Vector (文生矢量图)")
+    
+    with tab_1:
+        with gr.Accordion('🕹Usage (操作说明)', open=True,):
+            with gr.Tabs():
+                gr.HTML(INSTRUCTIONS_T2V)
+        with gr.Row():
+            with gr.Column():
+                with gr.Row():
+                    lora_radio_t2v = gr.Radio(choices=STYLE_MAPPING_REVERT.keys(), type="value", value='Commercial flat illustration', label="Style Selection (风格选择)")
+                    style_image_t2v = gr.Image(value=STYLE_IMG_DICT_REVERT["0"],type="numpy", image_mode="RGBA", label="Style Image (风格图片)")
+                prompt_t2v = gr.Textbox(label="Prompt (提示词)", value="a cute dog")
+                details_slider_t2v = gr.Radio(label="图形细节 Details", choices=["Minimal 简约","Medium 中等","Rich 丰富" ], type="value", value="Medium 中等")
+                num_samples_t2v = gr.Slider(label="Images (图像数量)", minimum=1, maximum=2, value=1, step=1, visible=False)
+                vectorize = gr.Checkbox(label='generate svg file', value=True, visible=True)
+                run_button_t2v = gr.Button(value="Generation Vector (生成矢量)",elem_id="btnSEG")
+
+                with gr.Accordion("Image options", open=False):
+                    image_resolution_h = gr.Slider(label="Image height (高)", minimum=256, maximum=1024, value=512, step=64, visible=True)
+                    image_resolution_w = gr.Slider(label="Image width (宽)", minimum=256, maximum=1024, value=512, step=64, visible=True)
+                     
+
+                with gr.Accordion("Vector options", open=False):
+                    upsample_method_t2v = gr.Dropdown(choices=["None", "x4", "x2"], type="value", value="None", label="Upsample Method (超分倍数)[无, 4倍, 2倍], 默认无")
+                    sorting_method_t2v = gr.Dropdown(choices=["brightness","area"], type="value", value="brightness", label="Sorting Method (排序方法)[亮度,面积], 默认亮度")
+                    sorting_order_t2v = gr.Dropdown(choices=["ascend","descend"], type="value", value="descend", label="Sorting Order (排序顺序)[递增,递减], 默认递减")
+                    trace_mode_t2v = gr.Radio(choices=["overlap", "cutout"], type="value", value="overlap", label="Trace Mode (追踪模式)[堆叠,分开]")
+                    use_gpu_t2v = gr.Checkbox(label='use GPU (是否使用GPU加速[针对复杂图形较为有效]), 默认关闭', value=False, visible=True)
+                    svg_simplify_t2v = gr.Checkbox(label='Simplify SVG (简化矢量图, 默认关闭)', value=False, visible=True)
+                    svg_optimize_t2v = gr.Checkbox(label='Optimize SVG (优化矢量图, 默认关闭)', value=False, visible=True)
+                    speckle_removal_t2v = gr.Checkbox(label='Remove small speckle[是否移除面积过小的图形]', value=False)
+                    subsample_ratio_t2v = gr.Slider(label="Subsample Ratio", minimum=1, maximum=10000, value=12, step=1, visible=False)
+
+                with gr.Tab("Prompt Examples"):
+                    t2v_examples = gr.Examples(
+                        PROMPT_EXAMPLE,
+                        [prompt_t2v],
+                        examples_per_page=5,
+                        label=''
+                    )
+
+            with gr.Column():
+                result_gallery_t2v = gr.Gallery(label='图像生成结果 Generation Results', show_label=False, elem_id="Gallery")
+
+            def update_lora_image(lora_radio: gr.Radio):
+                return gr.Image(value=STYLE_IMG_DICT_REVERT[STYLE_MAPPING_REVERT[lora_radio]],)
+            
+            lora_radio_t2v.change(fn=update_lora_image, inputs=[lora_radio_t2v], outputs=[style_image_t2v])
+
+
+
+            t2v_ips = [prompt_t2v, num_samples_t2v, image_resolution_h, image_resolution_w, details_slider_t2v, lora_radio_t2v, vectorize, upsample_method_t2v, svg_simplify_t2v, svg_optimize_t2v, trace_mode_t2v, subsample_ratio_t2v, speckle_removal_t2v,sorting_method_t2v, sorting_order_t2v, use_gpu_t2v]
+
+
+            run_button_t2v.click(fn=process_t2v, inputs=t2v_ips, outputs=result_gallery_t2v)
 
     tab_3 = gr.Tab(label="IMG to SVG")
     with tab_3:   

demo/generation.py

@@ -6,7 +6,7 @@ import gradio as gr
 import requests
 
 from demo.log import logger
-from demo.util import download_svgs, upload_np_2_oss
+from demo.util import download_svgs, upload_np_2_oss, download_images
 
 API_KEY = os.getenv("API_KEY_GENERATION")
 
@@ -84,7 +84,7 @@ def call_generation(input_path,
             res = json.loads(res_.content.decode())
             request_id = res['request_id']
             task_id = res['output']['task_id']
-            logger.info(f"task_id: {task_id}: Create Poster Imitation request success. Params: {data}")
+            logger.info(f"task_id: {task_id}: Create Vectorization I2V request success. Params: {data}")
 
             # 异步查询
             is_running = True
@@ -127,5 +127,146 @@ def call_generation(input_path,
         raise gr.Error("Fail to Generation.")
     return all_image_data[-1:]
 
+def call_generation_t2v(prompt,
+                    num_imgs,
+                    image_resolution_h,
+                    image_resolution_w,
+                    details,
+                    style,
+                    vectorize,
+                    preprocess, 
+                    simplify, 
+                    optimize, 
+                    mode, 
+                    subsample_ratio, 
+                    speckle_removal, 
+                    sorting_method,
+                    sorting_order,
+                    use_gpu):
+    
+    ## generate image name based on time stamp
+    time_str = time.strftime("%Y%m%d%H%M%S", time.localtime())
+    # img_name = f"upload_{time_str}.png"
+    svg_name = f"result_{time_str}"
+    generate_img_name = f"generate_{time_str}"
+    BATCH_SIZE = 1
+    count = 1
+    start_ind = 0
+    if simplify:
+        BATCH_SIZE += 1
+        count +=1
+        start_ind += 1
+    if optimize:
+        BATCH_SIZE += 1
+        start_ind += 1
+        count +=1
+
+    BATCH_SIZE *= num_imgs
+    # img_url = upload_np_2_oss(input_path, name=img_name)
+    # simplify = convert_bool_to_str(simplify)
+    # optimize = convert_bool_to_str(optimize)
+    # speckle_removal = convert_bool_to_str(speckle_removal)
+    # use_gpu = convert_bool_to_str(use_gpu)
+
+    headers = {
+        "Content-Type": "application/json",
+        "Accept": "application/json",
+        "Authorization": f"Bearer {API_KEY}",
+        "X-DashScope-Async": "enable",
+    }
+    data = {
+        "model": "pre-vectorize_anything_t2v-2352",
+        "input": {
+            "prompt": prompt
+        },
+        "parameters":{
+            "num_imgs" : num_imgs,
+            "image_resolution_h": image_resolution_h,
+            "image_resolution_w": image_resolution_w,
+            "details" : details,
+            "style" : style,
+            "vectorize" : vectorize,
+            "preprocess": preprocess,
+            "mode": mode,
+            "simplify": simplify, 
+            "optimize": optimize,
+            "sorting_method": sorting_method,
+            "sorting_order": sorting_order,
+            "subsample_ratio": subsample_ratio,
+            "speckle_removal": speckle_removal, 
+            "use_GPU": use_gpu
+        }
+    }
+    url_create_task = 'https://poc-dashscope.aliyuncs.com/api/v1/services/aigc/text2image/image-synthesis'
+
+    all_res_ = []
+    REPEAT = 1
+    for _ in range(REPEAT):
+        try:
+            res_ = requests.post(url_create_task, data=json.dumps(data), headers=headers, timeout=120)
+            print(json.dumps(data))
+            all_res_.append(res_)
+        except requests.Timeout:
+            # back off and retry
+            raise gr.Error("网络波动，请求失败，请再次尝试")
+
+    all_image_data = []
+    for res_ in all_res_:
+        respose_code = res_.status_code
+        if 200 == respose_code:
+            res = json.loads(res_.content.decode())
+            request_id = res['request_id']
+            task_id = res['output']['task_id']
+            logger.info(f"task_id: {task_id}: Create Vectorize T2V request success. Params: {data}")
+
+            # 异步查询
+            is_running = True
+            while is_running:
+                # url_query = f'https://dashscope.aliyuncs.com/api/v1/tasks/{task_id}'
+                url_query = f'https://poc-dashscope.aliyuncs.com/api/v1/tasks/{task_id}'
+                try:
+                    res_ = requests.post(url_query, headers=headers, timeout=120)
+                except requests.Timeout:
+                    # back off and retry
+                    raise gr.Error("网络波动，请求失败，请再次尝试")
+                respose_code = res_.status_code
+                if 200 == respose_code:
+                    res = json.loads(res_.content.decode())
+                    if "SUCCEEDED" == res['output']['task_status']:
+                        logger.info(f"task_id: {task_id}: Generation task query success.")
+                        results = res['output']
+                        img_urls = results['output_img']
+                        logger.info(f"task_id: {task_id}: {res}")
+                        break
+                    elif "FAILED" != res['output']['task_status']:
+                        logger.debug(f"task_id: {task_id}: query result...")
+                        time.sleep(1)
+                    else:
+                        raise gr.Error('Fail to get results from Generation task.')
+
+                else:
+                    logger.error(f'task_id: {task_id}: Fail to query task result: {res_.content}')
+                    raise gr.Error("Fail to query task result.")
+
+            logger.info(f"task_id: {task_id}: download generated images.")
+            if vectorize:
+                img_data = download_svgs(img_urls, BATCH_SIZE, svg_name)
+            else:
+                img_data = download_images(img_urls, num_imgs, generate_img_name)
+
+            logger.info(f"task_id: {task_id}: Generate done.")
+            all_image_data += img_data
+        else:
+            logger.error(f'Fail to create Generation task: {res_.content}')
+            raise gr.Error("Fail to create Generation task.")
+    if vectorize:
+        if len(all_image_data) != REPEAT * BATCH_SIZE:
+            raise gr.Error("Fail to Generation.")
+    else:
+        if len(all_image_data) != REPEAT * num_imgs:
+            raise gr.Error("Fail to Generation.")
+    return all_image_data[start_ind::BATCH_SIZE//num_imgs]
+
+
 if __name__ == "__main__":
     call_generation()

demo/instructions.py

@@ -28,4 +28,15 @@ INSTRUCTIONS_VECTORIZE_SIMPLIFY='''
 </ul>
 <p>Step3: 点击转化矢量图 Press the button to vectorize</p>
 </body>
+'''
+
+INSTRUCTIONS_T2V='''
+
+<p>Step 1: Choose a vector style</p>
+<p>Step 2: Input Prompt
+<p>Step 3: Adjust the details levels of shape</p>
+<p>Step 4: Generate vectors</p>
+
+<p> More parameters can be found in Image Options and Vector Options</p>
+
 '''

demo/util.py

@@ -20,23 +20,24 @@ oss_path_img_gallery = "VectorizeAnything"
 
 bucket = oss2.Bucket(oss2.Auth(access_key_id, access_key_secret), endpoint, bucket_name)
 
-def download_img_pil(index, img_url):
+def download_img_pil(index, img_url, img_name):
     # print(img_url)
     r = requests.get(img_url, stream=True)
     if r.status_code == 200:
         img = Image.open(io.BytesIO(r.content))
+        upload_svg_2_oss(io.BytesIO(r.content), f"{img_name}_{index}.png")
         return (index, img)
     else:
         logger.error(f"Fail to download: {img_url}")
 
 
-def download_images(img_urls, batch_size):
+def download_images(img_urls, batch_size, img_name):
     imgs_pil = [None] * batch_size
     # worker_results = []
     with concurrent.futures.ThreadPoolExecutor(max_workers=4) as executor:
         to_do = []
         for i, url in enumerate(img_urls):
-            future = executor.submit(download_img_pil, i, url)
+            future = executor.submit(download_img_pil, i, url, img_name)
             to_do.append(future)
 
         for future in concurrent.futures.as_completed(to_do):