test

.gitattributes

@@ -38,7 +38,3 @@ assets/scene_2.png filter=lfs diff=lfs merge=lfs -text
 assets/scene_3.png filter=lfs diff=lfs merge=lfs -text
 assets/scene_4.png filter=lfs diff=lfs merge=lfs -text
 assets/scene_5.png filter=lfs diff=lfs merge=lfs -text
-scene_assets/scene1_no_ceiling.glb filter=lfs diff=lfs merge=lfs -text
-scene_assets/scene2_no_ceiling.glb filter=lfs diff=lfs merge=lfs -text
-scene_assets/scene3_no_ceiling.glb filter=lfs diff=lfs merge=lfs -text
-scene_assets/scene4_no_ceiling.glb filter=lfs diff=lfs merge=lfs -text

.gradio/certificate.pem

@@ -1,31 +0,0 @@
------BEGIN CERTIFICATE-----
-MIIFazCCA1OgAwIBAgIRAIIQz7DSQONZRGPgu2OCiwAwDQYJKoZIhvcNAQELBQAw
-TzELMAkGA1UEBhMCVVMxKTAnBgNVBAoTIEludGVybmV0IFNlY3VyaXR5IFJlc2Vh
-cmNoIEdyb3VwMRUwEwYDVQQDEwxJU1JHIFJvb3QgWDEwHhcNMTUwNjA0MTEwNDM4
-WhcNMzUwNjA0MTEwNDM4WjBPMQswCQYDVQQGEwJVUzEpMCcGA1UEChMgSW50ZXJu
-ZXQgU2VjdXJpdHkgUmVzZWFyY2ggR3JvdXAxFTATBgNVBAMTDElTUkcgUm9vdCBY
-MTCCAiIwDQYJKoZIhvcNAQEBBQADggIPADCCAgoCggIBAK3oJHP0FDfzm54rVygc
-h77ct984kIxuPOZXoHj3dcKi/vVqbvYATyjb3miGbESTtrFj/RQSa78f0uoxmyF+
-0TM8ukj13Xnfs7j/EvEhmkvBioZxaUpmZmyPfjxwv60pIgbz5MDmgK7iS4+3mX6U
-A5/TR5d8mUgjU+g4rk8Kb4Mu0UlXjIB0ttov0DiNewNwIRt18jA8+o+u3dpjq+sW
-T8KOEUt+zwvo/7V3LvSye0rgTBIlDHCNAymg4VMk7BPZ7hm/ELNKjD+Jo2FR3qyH
-B5T0Y3HsLuJvW5iB4YlcNHlsdu87kGJ55tukmi8mxdAQ4Q7e2RCOFvu396j3x+UC
-B5iPNgiV5+I3lg02dZ77DnKxHZu8A/lJBdiB3QW0KtZB6awBdpUKD9jf1b0SHzUv
-KBds0pjBqAlkd25HN7rOrFleaJ1/ctaJxQZBKT5ZPt0m9STJEadao0xAH0ahmbWn
-OlFuhjuefXKnEgV4We0+UXgVCwOPjdAvBbI+e0ocS3MFEvzG6uBQE3xDk3SzynTn
-jh8BCNAw1FtxNrQHusEwMFxIt4I7mKZ9YIqioymCzLq9gwQbooMDQaHWBfEbwrbw
-qHyGO0aoSCqI3Haadr8faqU9GY/rOPNk3sgrDQoo//fb4hVC1CLQJ13hef4Y53CI
-rU7m2Ys6xt0nUW7/vGT1M0NPAgMBAAGjQjBAMA4GA1UdDwEB/wQEAwIBBjAPBgNV
-HRMBAf8EBTADAQH/MB0GA1UdDgQWBBR5tFnme7bl5AFzgAiIyBpY9umbbjANBgkq
-hkiG9w0BAQsFAAOCAgEAVR9YqbyyqFDQDLHYGmkgJykIrGF1XIpu+ILlaS/V9lZL
-ubhzEFnTIZd+50xx+7LSYK05qAvqFyFWhfFQDlnrzuBZ6brJFe+GnY+EgPbk6ZGQ
-3BebYhtF8GaV0nxvwuo77x/Py9auJ/GpsMiu/X1+mvoiBOv/2X/qkSsisRcOj/KK
-NFtY2PwByVS5uCbMiogziUwthDyC3+6WVwW6LLv3xLfHTjuCvjHIInNzktHCgKQ5
-ORAzI4JMPJ+GslWYHb4phowim57iaztXOoJwTdwJx4nLCgdNbOhdjsnvzqvHu7Ur
-TkXWStAmzOVyyghqpZXjFaH3pO3JLF+l+/+sKAIuvtd7u+Nxe5AW0wdeRlN8NwdC
-jNPElpzVmbUq4JUagEiuTDkHzsxHpFKVK7q4+63SM1N95R1NbdWhscdCb+ZAJzVc
-oyi3B43njTOQ5yOf+1CceWxG1bQVs5ZufpsMljq4Ui0/1lvh+wjChP4kqKOJ2qxq
-4RgqsahDYVvTH9w7jXbyLeiNdd8XM2w9U/t7y0Ff/9yi0GE44Za4rF2LN9d11TPA
-mRGunUHBcnWEvgJBQl9nJEiU0Zsnvgc/ubhPgXRR4Xq37Z0j4r7g1SgEEzwxA57d
-emyPxgcYxn/eR44/KJ4EBs+lVDR3veyJm+kXQ99b21/+jh5Xos1AnX5iItreGCc=
------END CERTIFICATE-----

app.py

@@ -1,52 +1,17 @@
 # main.py
 # 主入口文件，负责启动 Gradio UI
 import gradio as gr
-from config import SCENE_CONFIGS, MODEL_CHOICES, MODE_CHOICES, EPISODE_CONFIGS
+from config import SCENE_CONFIGS, MODEL_CHOICES, MODE_CHOICES
 from backend_api import submit_to_backend, get_task_status, get_task_result
 from logging_utils import log_access, log_submission, is_request_allowed
 from simulation import stream_simulation_results, convert_to_h264, create_final_video_from_oss_images
-from ui_components import update_history_display, update_scene_display, update_episode_display, update_log_display, get_scene_instruction
+from ui_components import update_history_display, update_scene_display, update_log_display, get_scene_instruction
 from oss_utils import download_oss_file, get_user_tmp_dir, cleanup_user_tmp_dir, oss_file_exists, clean_oss_result_path
 import os
 from datetime import datetime
 
 SESSION_TASKS = {}
 
-# 添加版本检查函数
-def get_gradio_versions():
-    """获取Gradio相关包的版本信息"""
-    try:
-        import gradio
-        gradio_version = gradio.__version__
-    except:
-        gradio_version = "Unknown"
-    
-    try:
-        import gradio_client
-        gradio_client_version = gradio_client.__version__
-    except:
-        gradio_client_version = "Not installed"
-    
-    return gradio_version, gradio_client_version
-
-def display_version_info():
-    """显示版本信息"""
-    gradio_version, gradio_client_version = get_gradio_versions()
-    
-    # 在控制台输出版本信息
-    print("=" * 50)
-    print("📦 Package Versions:")
-    print(f"   Gradio: {gradio_version}")
-    print(f"   Gradio Client: {gradio_client_version}")
-    print("=" * 50)
-    
-    # 返回格式化的版本信息字符串
-    version_info = f"""**📦 Package Versions:**
-- **Gradio**: `{gradio_version}`
-- **Gradio Client**: `{gradio_client_version}`"""
-    return version_info
-
-
 def run_simulation(scene, model, mode, prompt, history, request: gr.Request):
     timestamp = datetime.now().strftime("%Y-%m-%d %H:%M:%S")
     scene_desc = SCENE_CONFIGS.get(scene, {}).get("description", scene)
@@ -171,25 +136,6 @@ custom_css = """
 .history-accordion {
     margin-bottom: 10px;
 }
-.scene-preview {
-    height: 400px;
-    border: 1px solid #ddd;
-    border-radius: 8px;
-    overflow: hidden;
-}
-.version-info {
-    background: #e8f4fd;
-    border: 1px solid #bee5eb;
-    border-radius: 6px;
-    padding: 10px;
-    margin: 10px 0;
-    font-size: 0.9em;
-}
-.dark .version-info {
-    background: #1a2a3a;
-    border-color: #3a5a7a;
-    color: #e0e0e0;
-}
 """
 
 header_html = """
@@ -224,40 +170,22 @@ header_html = """
 with gr.Blocks(title="InternNav Model Inference Demo", css=custom_css) as demo:
     gr.HTML(header_html)
     
-    # 添加版本信息显示
-    version_display = gr.Markdown(
-        value="",
-        elem_classes=["version-info"],
-        visible=True
-    )
-    
     history_state = gr.State([])
     with gr.Row():
         with gr.Column(elem_id="simulation-panel"):
             gr.Markdown("### Simulation Settings")
-            with gr.Row():
-                scene_dropdown = gr.Dropdown(
-                    label="Choose a scene",
-                    choices=list(SCENE_CONFIGS.keys()),
-                    value="demo1",
-                    interactive=True
-                )
-                episode_dropdown = gr.Dropdown(
-                    label="Select Start Position",
-                    choices=list(EPISODE_CONFIGS.keys()),
-                    value="episode_1",
-                    interactive=True
-                )
-            
-            with gr.Row():
-                scene_preview = gr.Model3D(
-                    elem_classes=["scene-preview"],
-                    camera_position=(90.0, 120, 20000.0)
-                )
-                fps_preview = gr.Image(label="FPS Preview")
-            
-            scene_description = gr.Markdown("### Scene preview")
-            
+            scene_dropdown = gr.Dropdown(
+                label="Choose a scene",
+                choices=list(SCENE_CONFIGS.keys()),
+                value="demo1",
+                interactive=True
+            )
+            scene_description = gr.Markdown("")
+            scene_preview = gr.Image(
+                label="Scene Preview",
+                elem_classes=["scene-preview"],
+                interactive=False
+            )
             prompt_input = gr.Textbox(
                 label="Navigation Prompt",
                 value="Walk past the left side of the bed and stop in the doorway.",
@@ -281,16 +209,6 @@ with gr.Blocks(title="InternNav Model Inference Demo", css=custom_css) as demo:
                 fn=lambda scene: [update_scene_display(scene)[0], update_scene_display(scene)[1], get_scene_instruction(scene)],
                 inputs=scene_dropdown,
                 outputs=[scene_description, scene_preview, prompt_input]
-            ).then(
-                update_episode_display,
-                inputs=[scene_dropdown, episode_dropdown],
-                outputs=[fps_preview]
-            )
-            
-            episode_dropdown.change(
-                update_episode_display,
-                inputs=[scene_dropdown, episode_dropdown],
-                outputs=[fps_preview]
             )
             
             submit_btn = gr.Button("Start Navigation Simulation", variant="primary")
@@ -337,11 +255,8 @@ with gr.Blocks(title="InternNav Model Inference Demo", css=custom_css) as demo:
         queue=True
     )
     demo.load(
-        fn=lambda: [display_version_info(), update_scene_display("demo1")[0], update_scene_display("demo1")[1]],
-        outputs=[version_display, scene_description, scene_preview]
-    ).then(
-        fn=lambda: update_episode_display("demo1", "episode_1"),
-        outputs=[fps_preview]
+        fn=lambda: update_scene_display("demo1"),
+        outputs=[scene_description, scene_preview]
     )
     demo.load(
         fn=record_access,
@@ -353,9 +268,6 @@ with gr.Blocks(title="InternNav Model Inference Demo", css=custom_css) as demo:
     demo.unload(fn=cleanup_session)
 
 if __name__ == "__main__":
-    # 在启动时显示版本信息
-    display_version_info()
-    
     demo.launch(
         server_name="0.0.0.0", 
         server_port=7860,  # Hugging Face Space默认端口

app_old.py

@@ -1,277 +0,0 @@
-# main.py
-# 主入口文件，负责启动 Gradio UI
-import gradio as gr
-from config import SCENE_CONFIGS, MODEL_CHOICES, MODE_CHOICES
-from backend_api import submit_to_backend, get_task_status, get_task_result
-from logging_utils import log_access, log_submission, is_request_allowed
-from simulation import stream_simulation_results, convert_to_h264, create_final_video_from_oss_images
-from ui_components import update_history_display, update_scene_display, update_log_display, get_scene_instruction
-from oss_utils import download_oss_file, get_user_tmp_dir, cleanup_user_tmp_dir, oss_file_exists, clean_oss_result_path
-import os
-from datetime import datetime
-
-SESSION_TASKS = {}
-
-def run_simulation(scene, model, mode, prompt, history, request: gr.Request):
-    timestamp = datetime.now().strftime("%Y-%m-%d %H:%M:%S")
-    scene_desc = SCENE_CONFIGS.get(scene, {}).get("description", scene)
-    user_ip = request.client.host if request else "unknown"
-    session_id = request.session_hash
-    
-    if not is_request_allowed(user_ip):
-        log_submission(scene, prompt, model, user_ip, "IP blocked temporarily")
-        raise gr.Error("Too many requests from this IP. Please wait and try again one minute later.")
-    
-    # 提交任务到后端
-    submission_result = submit_to_backend(scene, prompt, mode, model, user_ip)
-    if submission_result.get("status") != "pending":
-        log_submission(scene, prompt, model, user_ip, "Submission failed")
-        raise gr.Error(f"Submission failed: {submission_result.get('message', 'unknown issue')}")
-    
-    try:
-        task_id = submission_result["task_id"]
-        SESSION_TASKS[session_id] = task_id
-        gr.Info(f"Simulation started, task_id: {task_id}")
-        
-        import time
-        time.sleep(5)
-        status = get_task_status(task_id)
-        # OSS上的结果文件夹路径，不再检查本地路径是否存在
-        result_folder = clean_oss_result_path(status.get("result", f"gradio_demo/tasks/{task_id}"), task_id)
-        
-    except Exception as e:
-        log_submission(scene, prompt, model, user_ip, str(e))
-        raise gr.Error(f"error occurred when parsing submission result from backend: {str(e)}")
-
-    # 流式输出视频片段（从OSS读取）
-    try:
-        for video_path in stream_simulation_results(result_folder, task_id, request):
-            if video_path:
-                yield video_path, history
-    except Exception as e:
-        log_submission(scene, prompt, model, user_ip, str(e))
-        raise gr.Error(f"流式输出过程中出错: {str(e)}")
-
-    # 获取最终任务状态
-    status = get_task_status(task_id)
-    if status.get("status") == "completed":
-        try:
-            # 从OSS上的所有图片拼接成最终视频（6帧每秒）
-            gr.Info("Creating final video from all OSS images...")
-            video_path = create_final_video_from_oss_images(result_folder, task_id, request, fps=6)
-            gr.Info(f"Final video created successfully with 6 fps!")
-                
-        except Exception as e:
-            print(f"Error creating final video from OSS images: {e}")
-            log_submission(scene, prompt, model, user_ip, f"Final video creation failed: {str(e)}")
-            video_path = None
-        new_entry = {
-            "timestamp": timestamp,
-            "scene": scene,
-            "model": model,
-            "mode": mode,
-            "prompt": prompt,
-            "video_path": video_path
-        }
-        updated_history = history + [new_entry]
-        if len(updated_history) > 10:
-            updated_history = updated_history[:10]
-        log_submission(scene, prompt, model, user_ip, "success")
-        gr.Info("Simulation completed successfully!")
-        yield None, updated_history
-    elif status.get("status") == "failed":
-        log_submission(scene, prompt, model, user_ip, status.get('result', 'backend error'))
-        raise gr.Error(f"任务执行失败: {status.get('result', 'backend 未知错误')}")
-        yield None, history
-    elif status.get("status") == "terminated":
-        log_submission(scene, prompt, model, user_ip, "terminated")
-        # 对于终止的任务，不再检查本地文件
-        yield None, history
-    else:
-        log_submission(scene, prompt, model, user_ip, "missing task's status from backend")
-        raise gr.Error("missing task's status from backend")
-        yield None, history
-
-def cleanup_session(request: gr.Request):
-    session_id = request.session_hash
-    task_id = SESSION_TASKS.pop(session_id, None)
-    from config import BACKEND_URL
-    import requests
-    if task_id:
-        try:
-            requests.post(f"{BACKEND_URL}/predict/terminate/{task_id}", timeout=3)
-        except Exception:
-            pass
-    
-    # 清理用户临时目录
-    cleanup_user_tmp_dir(session_id)
-
-def record_access(request: gr.Request):
-    user_ip = request.client.host if request else "unknown"
-    user_agent = request.headers.get("user-agent", "unknown")
-    log_access(user_ip, user_agent)
-    return update_log_display()
-
-
-
-custom_css = """
-#simulation-panel {
-    border-radius: 8px;
-    padding: 20px;
-    background: #f9f9f9;
-    box-shadow: 0 2px 4px rgba(0,0,0,0.1);
-}
-#result-panel {
-    border-radius: 8px;
-    padding: 20px;
-    background: #f0f8ff;
-}
-.dark #simulation-panel { background: #2a2a2a; }
-.dark #result-panel { background: #1a2a3a; }
-.history-container {
-    max-height: 600px;
-    overflow-y: auto;
-    margin-top: 20px;
-}
-.history-accordion {
-    margin-bottom: 10px;
-}
-"""
-
-header_html = """
-<div style="display: flex; justify-content: space-between; align-items: center; width: 100%; margin-bottom: 20px; padding: 20px; background: linear-gradient(135deg, #e0e5ec 0%, #a7b5d0 100%); border-radius: 8px; box-shadow: 0 2px 8px rgba(0,0,0,0.1);">
-    <div style="display: flex; align-items: center;">
-        <img src="https://www.shlab.org.cn/static/img/index_14.685f6559.png" alt="Institution Logo" style="height: 60px; margin-right: 20px;">
-        <div>
-            <h1 style="margin: 0; color: #2c3e50; font-weight: 600;">🤖 InternNav Model Inference Demo</h1>
-            <p style="margin: 4px 0 0 0; color: #5d6d7e; font-size: 0.9em;">Model trained on InternNav framework</p>
-        </div>
-    </div>
-    <div style="display: flex; gap: 15px; align-items: center;">
-        <a href="https://github.com/OpenRobotLab" target="_blank" style="text-decoration: none; transition: transform 0.2s;" onmouseover="this.style.transform='scale(1.1)'" onmouseout="this.style.transform='scale(1)'">
-            <img src="https://github.githubassets.com/images/modules/logos_page/GitHub-Mark.png" alt="GitHub" style="height: 30px;">
-        </a>
-        <a href="https://huggingface.co/OpenRobotLab" target="_blank" style="text-decoration: none; transition: transform 0.2s;" onmouseover="this.style.transform='scale(1.1)'" onmouseout="this.style.transform='scale(1)'">
-            <img src="https://huggingface.co/front/assets/huggingface_logo-noborder.svg" alt="HuggingFace" style="height: 30px;">
-        </a>
-        <a href="https://huggingface.co/spaces/OpenRobotLab/InternManip-eval-demo" target="_blank">
-            <button style="padding: 8px 15px; background: #3498db; color: white; border: none; border-radius: 4px; cursor: pointer; font-weight: 500; transition: all 0.2s;" 
-                    onmouseover="this.style.backgroundColor='#2980b9'; this.style.transform='scale(1.05)'" 
-                    onmouseout="this.style.backgroundColor='#3498db'; this.style.transform='scale(1)'">
-                Go to InternManip Demo
-            </button>
-        </a>
-    </div>
-</div>
-"""
-
-
-
-with gr.Blocks(title="InternNav Model Inference Demo", css=custom_css) as demo:
-    gr.HTML(header_html)
-    
-    history_state = gr.State([])
-    with gr.Row():
-        with gr.Column(elem_id="simulation-panel"):
-            gr.Markdown("### Simulation Settings")
-            scene_dropdown = gr.Dropdown(
-                label="Choose a scene",
-                choices=list(SCENE_CONFIGS.keys()),
-                value="demo1",
-                interactive=True
-            )
-            scene_description = gr.Markdown("")
-            scene_preview = gr.Image(
-                label="Scene Preview",
-                elem_classes=["scene-preview"],
-                interactive=False
-            )
-            prompt_input = gr.Textbox(
-                label="Navigation Prompt",
-                value="Walk past the left side of the bed and stop in the doorway.",
-                placeholder="e.g.: 'Walk past the left side of the bed and stop in the doorway.'",
-                lines=2,
-                max_lines=4
-            )
-            model_dropdown = gr.Dropdown(
-                label="Chose a pretrained model",
-                choices=MODEL_CHOICES,
-                value=MODEL_CHOICES[0],
-                interactive=True
-            )
-            mode_dropdown = gr.Dropdown(
-                label="Select Mode",
-                choices=MODE_CHOICES,
-                value=MODE_CHOICES[0],
-                interactive=True
-            )
-            scene_dropdown.change(
-                fn=lambda scene: [update_scene_display(scene)[0], update_scene_display(scene)[1], get_scene_instruction(scene)],
-                inputs=scene_dropdown,
-                outputs=[scene_description, scene_preview, prompt_input]
-            )
-            
-            submit_btn = gr.Button("Start Navigation Simulation", variant="primary")
-        with gr.Column(elem_id="result-panel"):
-            gr.Markdown("### Latest Simulation Result")
-            video_output = gr.Video(
-                label="Live",
-                interactive=False,
-                format="mp4",
-                autoplay=True,
-                streaming=True
-            )
-            with gr.Column() as history_container:
-                gr.Markdown("### History")
-                gr.Markdown("#### History will be reset after refresh")
-                history_slots = []
-                for i in range(10):
-                    with gr.Column(visible=False) as slot:
-                        with gr.Accordion(visible=False, open=False) as accordion:
-                            video = gr.Video(interactive=False)
-                            detail_md = gr.Markdown()
-                    history_slots.append((slot, accordion, video, detail_md))
-    gr.Examples(
-        examples=[
-            ["demo1", "rdp", "vlnPE", "Walk past the left side of the bed and stop in the doorway."],
-            ["demo2", "rdp", "vlnPE", "Walk through the bathroom, past the sink and toilet. Stop in front of the counter with the two suitcase."],
-            ["demo3", "rdp", "vlnPE", "Do a U-turn. Walk forward through the kitchen, heading to the black door. Walk out of the door and take a right onto the deck. Walk out on to the deck and stop."],
-            ["demo4", "rdp", "vlnPE", "Walk out of bathroom and stand on white bath mat."],
-            ["demo5", "rdp", "vlnPE", "Walk straight through the double wood doors, follow the red carpet straight to the next doorway and stop where the carpet splits off."]
-        ],
-        inputs=[scene_dropdown, model_dropdown, mode_dropdown, prompt_input],
-        label="Navigation Task Examples"
-    )
-    submit_btn.click(
-        fn=run_simulation,
-        inputs=[scene_dropdown, model_dropdown, mode_dropdown, prompt_input, history_state],
-        outputs=[video_output, history_state],
-        queue=True,
-        api_name="run_simulation"
-    ).then(
-        fn=update_history_display,
-        inputs=history_state,
-        outputs=[comp for slot in history_slots for comp in slot],
-        queue=True
-    )
-    demo.load(
-        fn=lambda: update_scene_display("demo1"),
-        outputs=[scene_description, scene_preview]
-    )
-    demo.load(
-        fn=record_access,
-        inputs=None,
-        outputs=None,
-        queue=False
-    )
-    demo.queue(default_concurrency_limit=8)
-    demo.unload(fn=cleanup_session)
-
-if __name__ == "__main__":
-    demo.launch(
-        server_name="0.0.0.0", 
-        server_port=7860,  # Hugging Face Space默认端口
-        share=False, 
-        debug=False,  # 生产环境建议关闭debug
-        allowed_paths=["./assets", "./logs", "./tmp"]  # 添加临时目录到允许路径
-    )

check_paths.py

@@ -1,17 +0,0 @@
-from config import SCENE_CONFIGS
-import os
-
-print('GLB文件路径检查:')
-for scene, config in SCENE_CONFIGS.items():
-    glb_path = config.get('glb_path', '')
-    exists = os.path.exists(glb_path)
-    status = '存在' if exists else '不存在'
-    print(f'{scene}: {glb_path} - {status}')
-
-print('\nEpisode图片路径检查 (以episode_1为例):')
-for scene, config in SCENE_CONFIGS.items():
-    scene_name = config.get('scene_name', scene)
-    image_path = os.path.join('scene_assets', f'{scene_name}_0.jpg')
-    exists = os.path.exists(image_path)
-    status = '存在' if exists else '不存在'
-    print(f'{scene} -> {scene_name}: {image_path} - {status}')

config.py

@@ -14,58 +14,33 @@ SCENE_CONFIGS = {
         "description": "Demo 1",
         "objects": ["bedroom", "kitchen", "living room", ""],
         "preview_image": "./assets/scene_1.png",
-        "glb_path": "scene_assets/scene1_no_ceiling.glb",
-        "scene_name": "17DRP5sb8fy",
         "default_instruction": "Walk past the left side of the bed and stop in the doorway."
     },
     "demo2": {
-        "description": "Demo 2", 
+        "description": "Demo 2",
         "objects": ["office", "meeting room", "corridor"],
         "preview_image": "./assets/scene_2.png",
-        "glb_path": "scene_assets/scene2_no_ceiling.glb",
-        "scene_name": "r1Q1Z4BcV1o",
         "default_instruction": "Walk through the bathroom, past the sink and toilet. Stop in front of the counter with the two suitcase."
     },
     "demo3": {
         "description": "Demo 3",
         "objects": ["garage", "workshop", "storage"],
         "preview_image": "./assets/scene_3.png",
-        "glb_path": "scene_assets/scene3_no_ceiling.glb",
-        "scene_name": "dhjEzFoUFzH",
         "default_instruction": "Do a U-turn. Walk forward through the kitchen, heading to the black door. Walk out of the door and take a right onto the deck. Walk out on to the deck and stop."
     },
     "demo4": {
         "description": "Demo 4",
         "objects": ["garden", "patio", "pool"],
         "preview_image": "./assets/scene_4.png",
-        "glb_path": "scene_assets/scene4_no_ceiling.glb",
-        "scene_name": "demo4",  # 没有对应的图片文件，保持原名
         "default_instruction": "Walk out of bathroom and stand on white bath mat."
     },
     "demo5": {
         "description": "Demo 5",
         "objects": ["library", "hall", "lounge"],
         "preview_image": "./assets/scene_5.png",
-        "glb_path": "scene_assets/scene4_no_ceiling.glb",  # 只有scene1-4的GLB文件
-        "scene_name": "demo5",  # 没有对应的图片文件，保持原名
         "default_instruction": "Walk straight through the double wood doors, follow the red carpet straight to the next doorway and stop where the carpet splits off."
     },
 }
 
-EPISODE_CONFIGS = {
-    "episode_1": {
-        "description": "1",
-    },
-    "episode_2": {
-        "description": "2",
-    },
-    "episode_3": {
-        "description": "3",
-    },
-    "episode_4": {
-        "description": "4",
-    }
-}
-
 MODEL_CHOICES = ["rdp", "cma"]
 MODE_CHOICES = ["vlnPE", "vlnCE"]

navigation_ui.py

@@ -1,575 +0,0 @@
-import base64
-import json
-import logging
-import os
-import subprocess
-import sys
-import threading
-import time
-import uuid
-from datetime import datetime, timedelta
-from typing import Dict, List, Optional
-
-import gradio as gr
-import numpy as np
-import open3d as o3d
-import plotly.graph_objects as go
-import requests
-from fastapi import APIRouter, FastAPI, HTTPException, status, BackgroundTasks, Response
-from pydantic import BaseModel
-
-import asyncio
-import uvicorn
-from collections import defaultdict
-
-BACKEND_URL = os.getenv("BACKEND_URL", "http://localhost:8001") # fastapi server
-API_ENDPOINTS = {
-    "submit_task": f"{BACKEND_URL}/predict/video",
-    "query_status": f"{BACKEND_URL}/predict/task",
-    "get_result": f"{BACKEND_URL}/predict"
-}
-
-
-SCENE_CONFIGS = {
-    "scene_1": {
-        "description": "Modern Apartment",
-        "name": "17DRP5sb8fy",
-        "glb_path": "scene_assets/scene1_no_ceiling.glb"  # PLY file path
-    },
-    "scene_2": {
-        "description": "Office Building",
-        "name": "r1Q1Z4BcV1o",
-        "glb_path": "scene_assets/scene2_no_ceiling.glb"
-    },
-    "scene_3": {
-        "description": "University Campus",
-        "name": "dhjEzFoUFzH",
-        "glb_path": "scene_assets/scene3_no_ceiling.glb"
-    },
-}
-
-EPISODE_CONFIGS = {
-    "episode_1": {
-        "description": "1",
-    },
-    "episode_2": {
-        "description": "2",
-    },
-    "episode_3": {
-        "description": "3",
-    },
-    "episode_4": {
-        "description": "4",
-    }
-}
-
-MODEL_CHOICES = [] 
-
-
-###############################################################################
-
-SESSION_TASKS = {}
-IP_REQUEST_RECORDS = defaultdict(list)
-IP_LIMIT = 5  
-
-def is_request_allowed(ip: str) -> bool:
-    now = datetime.now()
-    IP_REQUEST_RECORDS[ip] = [t for t in IP_REQUEST_RECORDS[ip] if now - t < timedelta(minutes=1)]
-    if len(IP_REQUEST_RECORDS[ip]) < IP_LIMIT:
-        IP_REQUEST_RECORDS[ip].append(now)
-        return True
-    return False
-
-###############################################################################
-
-
-# Log directory path
-LOG_DIR = "~/logs"
-os.makedirs(LOG_DIR, exist_ok=True)
-ACCESS_LOG = os.path.join(LOG_DIR, "access.log")
-SUBMISSION_LOG = os.path.join(LOG_DIR, "submissions.log")
-
-def log_access(user_ip: str = None, user_agent: str = None):
-    timestamp = datetime.now().strftime("%Y-%m-%d %H:%M:%S")
-    log_entry = {
-        "timestamp": timestamp,
-        "type": "access",
-        "user_ip": user_ip or "unknown",
-        "user_agent": user_agent or "unknown"
-    }
-    
-    with open(ACCESS_LOG, "a") as f:
-        f.write(json.dumps(log_entry) + "\n")
-
-def log_submission(scene: str, prompt: str, model: str, user: str = "anonymous", res: str = "unknown"):
-    timestamp = datetime.now().strftime("%Y-%m-%d %H:%M:%S")
-    log_entry = {
-        "timestamp": timestamp,
-        "type": "submission",
-        "user": user,
-        "scene": scene,
-        "prompt": prompt,
-        "model": model,
-        #"max_step": str(max_step),
-        "res": res
-    }
-    
-    with open(SUBMISSION_LOG, "a") as f:
-        f.write(json.dumps(log_entry) + "\n")
-
-def read_logs(log_type: str = "all", max_entries: int = 50) -> list:
-    logs = []
-    
-    if log_type in ["all", "access"]:
-        try:
-            with open(ACCESS_LOG, "r") as f:
-                for line in f:
-                    logs.append(json.loads(line.strip()))
-        except FileNotFoundError:
-            pass
-    
-    if log_type in ["all", "submission"]:
-        try:
-            with open(SUBMISSION_LOG, "r") as f:
-                for line in f:
-                    logs.append(json.loads(line.strip()))
-        except FileNotFoundError:
-            pass
-    
-    # Sorted by timestemp
-    logs.sort(key=lambda x: x["timestamp"], reverse=True)
-    return logs[:max_entries]
-
-def format_logs_for_display(logs: list) -> str:
-    if not logs:
-        return "No log record"
-    
-    markdown = "### System Access Log\n\n"
-    markdown += "| Time | Type | User/IP | Details |\n"
-    markdown += "|------|------|---------|----------|\n"
-    
-    for log in logs:
-        timestamp = log.get("timestamp", "unknown")
-        log_type = "Access" if log.get("type") == "access" else "Submission"
-        
-        if log_type == "Access":
-            user = log.get("user_ip", "unknown")
-            details = f"User-Agent: {log.get('user_agent', 'unknown')}"
-        else:
-            user = log.get("user", "anonymous")
-            result = log.get('res', 'unknown')
-            if result != "success": 
-                if len(result) > 40:  # Adjust this threshold as needed
-                    result = f"{result[:20]}...{result[-20:]}"
-            details = f"Scene: {log.get('scene', 'unknown')}, Prompt: {log.get('prompt', '')}, Model: {log.get('model', 'unknown')}, result: {result}"
-        
-        markdown += f"| {timestamp} | {log_type} | {user} | {details} |\n"
-    
-    return markdown
-
-
-def submit_to_backend(
-    scene: str,
-    prompt: str,
-    episode: str,
-    user: str = "Gradio-user",
-) -> dict:
-    job_id = str(uuid.uuid4())
-
-    scene_index = scene.split("_")[-1]
-    episode_index = episode.split("_")[-1]
-
-    data = {
-        "task_type": "vln_eval",  # Identify task type
-        "instruction": prompt,
-        "scene_index": scene_index,
-        "episode_index": episode_index,
-    }
-    
-    payload = {
-        "user": user,
-        "task": "robot_navigation",
-        "job_id": job_id,
-        "data": json.dumps(data)
-    }
-    
-    try:
-        headers = {"Content-Type": "application/json"}
-        response = requests.post(
-            API_ENDPOINTS["submit_task"],
-            json=payload,
-            headers=headers,
-            timeout=600
-        )
-        return response.json()
-    except Exception as e:
-        return {"status": "error", "message": str(e)}
-
-def get_task_status(task_id: str) -> dict:
-    try:
-        response = requests.get(f"{API_ENDPOINTS['query_status']}/{task_id}", timeout=600)
-        try:
-            return response.json()  
-        except json.JSONDecodeError:
-            return {"status": "error", "message": response.text}
-    except Exception as e:
-        return {"status": "error", "message": str(e)}
-
-
-def get_task_result(task_id: str) -> Optional[dict]:
-    try:
-        response = requests.get(
-            f"{API_ENDPOINTS['get_result']}/{task_id}",
-            timeout=5
-        )
-        return response.json()
-    except Exception as e:
-        print(f"Error fetching result: {e}")
-        return None
-
-def run_simulation(
-    scene: str,
-    prompt: str,
-    episode: str,
-    history: list,
-    request: gr.Request
-) -> dict:
-    model = "InternNav-VLA"
-
-    timestamp = datetime.now().strftime("%Y-%m-%d %H:%M:%S")
-    scene_desc = SCENE_CONFIGS.get(scene, {}).get("description", scene)
-
-    user_ip = request.client.host if request else "unknown"
-    session_id = request.session_hash
-
-    if not is_request_allowed(user_ip):
-        log_submission(scene, prompt, model, user_ip, "IP blocked temporarily")
-        raise gr.Error("Too many requests from this IP. Please wait and try again one minute later.")
-    
-    submission_result = submit_to_backend(scene, prompt, episode)
-    print("submission_result: ", submission_result)
-
-    if submission_result.get("status") != "pending":
-        log_submission(scene, prompt, model, user_ip, "Submission failed")
-        raise gr.Error(f"Submission failed: {submission_result.get('message', 'unknown issue')}")
-    
-    try:
-        task_id = submission_result["task_id"]
-        SESSION_TASKS[session_id] = task_id
-
-        gr.Info(f"Simulation started, task_id: {task_id}")
-        time.sleep(5)
-        # Get Task Status
-        status = get_task_status(task_id)
-        print("first status: ", status)
-        result_folder = status.get("result", "")
-    except Exception as e:
-        log_submission(scene, prompt, model, user_ip, str(e))
-        raise gr.Error(f"error occurred when parsing submission result from backend: {str(e)}")
-
-    while True:
-        status = get_task_status(task_id)
-        if status.get("status") == "completed":
-            break
-        elif status.get("status") == "failed":
-            break
-        time.sleep(1)
-    if status.get("status") == "completed":
-        import base64
-        video_bytes = base64.b64decode(status.get("video"))
-        receive_time = time.time()
-        with open(f"received_video_{receive_time}.mp4", "wb") as f:
-            f.write(video_bytes)
-        video_path = f"received_video_{receive_time}.mp4"
-        new_entry = {
-            "timestamp": timestamp,
-            "scene": scene,
-            "model": model,
-            "prompt": prompt,
-            "video_path": video_path
-        }
-        
-        updated_history = history + [new_entry]
-        
-        if len(updated_history) > 10:
-            updated_history = updated_history[:10]
-        
-        print("updated_history:", updated_history)
-        log_submission(scene, prompt, model, user_ip, "success")
-        gr.Info("Simulation completed successfully!")
-        yield video_path, updated_history
-
-    elif status.get("status") == "failed":
-        log_submission(scene, prompt, model, user_ip, status.get('result', 'backend error'))
-        raise gr.Error(f"task execution fails: {status.get('result', 'backend error')}")
-        yield None, history
-
-    elif status.get("status") == "terminated":
-        log_submission(scene, prompt, model, user_ip, "terminated")
-        video_path = os.path.join(result_folder, "output.mp4")
-        if os.path.exists(video_path):
-            return f" task {task_id} terminated with some results", video_path, history
-        else:
-            return f" task {task_id} terminated without any results", None, history
-
-    else:
-        log_submission(scene, prompt, model, user_ip, "missing task's status from backend")
-        yield None, history
-
-###################################################################################################################
-def update_history_display(history: list) -> list:
-    print("update_history_display")
-    updates = []
-    
-    for i in range(10):
-        if i < len(history):
-            entry = history[i]
-            updates.extend([
-                gr.update(visible=True),
-                gr.update(visible=True, label=f"Simulation {i+1}  scene: {entry['scene']}, prompt: {entry['prompt']}", open=False),
-                gr.update(value=entry['video_path'], visible=True),
-                gr.update(value=f"{entry['timestamp']}")
-            ])
-            print(f'update video')
-            print(entry['video_path'])
-        else:
-            updates.extend([
-                gr.update(visible=False),
-                gr.update(visible=False),
-                gr.update(value=None, visible=False),
-                gr.update(value="")
-            ])
-    print("update_history_display end!!")
-    return updates
-
-def update_scene_display(scene: str):
-    print(f"update_scene_display {scene}")
-    config = SCENE_CONFIGS.get(scene, {})
-    glb_path = config.get("glb_path", "")
-
-    # Validate if file path exists
-    if not os.path.exists(glb_path):
-        return None, None
-
-    return None, glb_path
-
-def update_episode_display(scene: str, episode: str):
-    print(f"update_episode_display {scene} {episode}")
-    config = SCENE_CONFIGS.get(scene, {})
-    scene_name = config.get("name", "")
-    episode_id = int(episode[-1])
-    image_path = os.path.join("scene_assets", f"{scene_name}_{episode_id-1}.jpg")
-    print(f"image_path {image_path}")
-    # vaild if file path exists
-    if not os.path.exists(image_path):
-        return None
-
-    return image_path
-def update_log_display():
-    logs = read_logs()
-    return format_logs_for_display(logs)
-##############################################################################
-
-
-def cleanup_session(request: gr.Request):
-    session_id = request.session_hash
-    task_id = SESSION_TASKS.pop(session_id, None)
-    if task_id:
-        try:
-            requests.post(f"{BACKEND_URL}/predict/terminate/{task_id}", timeout=3)
-            print(f"Task Terminated: {task_id}")
-        except Exception as e:
-            print(f"Task Termination Failed: {task_id}: {e}")
-
-
-
-###############################################################################
-
-custom_css = """
-#simulation-panel {
-    border-radius: 8px;
-    padding: 20px;
-    background: #f9f9f9;
-    box-shadow: 0 2px 4px rgba(0,0,0,0.1);
-}
-#result-panel {
-    border-radius: 8px;
-    padding: 20px;
-    background: #f0f8ff;
-}
-.dark #simulation-panel { background: #2a2a2a; }
-.dark #result-panel { background: #1a2a3a; }
-
-.history-container {
-    max-height: 600px;
-    overflow-y: auto;
-    margin-top: 20px;
-}
-
-.history-accordion {
-    margin-bottom: 10px;
-}
-
-.scene-preview {
-    height: 400px;
-    border: 1px solid #ddd;
-    border-radius: 8px;
-    overflow: hidden;
-}
-"""
-
-with gr.Blocks(title="Robot Navigation Inference", css=custom_css) as demo:
-    gr.Markdown("""
-    # 🧭 Habitat Robot Navigation Demo
-    ### Simulation Test Based on Habitat Framework
-    """)
-    
-    history_state = gr.State([])
-
-    with gr.Row():
-        with gr.Column(elem_id="simulation-panel"):
-            gr.Markdown("### Simulation Task Configuration")
-            with gr.Row():
-                scene_dropdown = gr.Dropdown(
-                    label="Select Scene",
-                    choices=list(SCENE_CONFIGS.keys()),
-                    value="scene_1",
-                    interactive=True,
-            )
-                episode_dropdown = gr.Dropdown(
-                    label="Select Start Position",
-                    choices=list(EPISODE_CONFIGS.keys()),
-                    value="episode_1",
-                    interactive=True,
-                )
-
-            with gr.Row():
-                scene_preview = gr.Model3D(elem_classes=["scene-preview"],
-                camera_position=(90.0, 120, 20000.0),
-                #display_mode="solid"
-                )
-                fps_preview = gr.Image(label="FPS Preview")
-            
-            scene_description = gr.Markdown("### Scene preview")
-            
-            prompt_input = gr.Textbox(
-                label="Navigation Instruction",
-                value="Exit the bedroom and turn left. Walk straight passing the gray couch and stop near the rug.",
-                placeholder="e.g.: 'Exit the bedroom and turn left. Walk straight passing the gray couch and stop near the rug.'",
-                lines=2,
-                max_lines=4
-            )
-                        
-            scene_dropdown.change(
-                update_scene_display,
-                inputs=scene_dropdown,
-                outputs=[scene_description, scene_preview]
-            ).then(
-                update_episode_display,
-                inputs=[scene_dropdown, episode_dropdown],
-                outputs=[fps_preview]
-            )
-            
-            episode_dropdown.change(
-                update_episode_display,
-                inputs=[scene_dropdown, episode_dropdown],
-                outputs=[fps_preview]
-            )
-            
-            submit_btn = gr.Button("Start Navigation Simulation", variant="primary")
-
-       
-        with gr.Column(elem_id="result-panel"):
-            gr.Markdown("### Latest Simulation Result")
-
-            # Video Output
-            video_output = gr.Video(
-                label="Live",
-                interactive=False,
-                format="mp4",
-                autoplay=True,
-                # streaming=True
-            )
-            
-            with gr.Column() as history_container:
-                gr.Markdown("### History")
-                gr.Markdown("#### History will be reset after refresh")
-                
-                history_slots = []
-                for i in range(10):
-                    with gr.Column(visible=False) as slot:
-                        with gr.Accordion(visible=False, open=False) as accordion:
-                            video = gr.Video(interactive=False)  
-                            detail_md = gr.Markdown() 
-                    history_slots.append((slot, accordion, video, detail_md))  
-    
-    with gr.Accordion("View System Log (DEV ONLY)", open=False):
-        logs_display = gr.Markdown()
-        refresh_logs_btn = gr.Button("Refresh Log", variant="secondary")
-        
-        refresh_logs_btn.click(
-            update_log_display,
-            outputs=logs_display
-        )
-
-    gr.Examples(
-        examples=[
-            ["scene_1", "Exit the bedroom and turn left. Walk straight passing the gray couch and stop near the rug.", "episode_0"],
-            ["scene_2", "Go from reception to conference room passing the water cooler.", "episode_1"],
-            ["scene_3", "From the classroom, go to the library via the main hall.", "episode_2"],
-            ["scene_4", "From emergency room to pharmacy passing nurse station.", "episode_3"]
-        ],
-        inputs=[scene_dropdown, prompt_input, episode_dropdown],
-        label="Navigation Task Example"
-    )
-    
-    submit_btn.click(
-        fn=run_simulation,
-        inputs=[scene_dropdown, prompt_input, episode_dropdown, history_state],
-        outputs=[video_output, history_state],
-        queue=True,
-        api_name="run_simulation"
-    ).then(
-        fn=update_history_display,
-        inputs=history_state,
-        outputs=[comp for slot in history_slots for comp in slot],
-        queue=True
-    ).then(
-        fn=update_log_display,
-        outputs=logs_display,
-    )
-
-
-    demo.load(
-        fn=lambda: update_scene_display("scene_1"),
-        outputs=[scene_description, scene_preview]
-    ).then(
-        fn=update_log_display,
-        outputs=logs_display
-    )
-    demo.load(
-        fn=lambda: update_episode_display("scene_1", "episode_1"),
-        outputs=[fps_preview]
-    )
-
-    def record_access(request: gr.Request):
-        user_ip = request.client.host if request else "unknown"
-        user_agent = request.headers.get("user-agent", "unknown")
-        log_access(user_ip, user_agent)
-        return update_log_display()   
-
-    demo.load(
-        fn=record_access,
-        inputs=None,
-        outputs=logs_display,
-        queue=False
-    )
-
-    demo.queue(default_concurrency_limit=8)
-
-    demo.unload(fn=cleanup_session)
-
-if __name__ == "__main__":
-    demo.launch(server_name="0.0.0.0", server_port=5750, debug=True, share = True, allowed_paths=["/mnt"])
-
-    

scene_assets/scene1_no_ceiling.glb

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:1c4f3c57d4a316bae5e8574602f0809f516fed743064b1f867b13ad4c5eafd6c
-size 43925372

scene_assets/scene2_no_ceiling.glb

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:d02e41f07558ee1b085edd08d8473fd38f65c3496d23365bdc171dae1178db5d
-size 47916736

scene_assets/scene3_no_ceiling.glb

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:aba480ff78c3d08416caea5b603cb775590aeb83b1e09af8606aafbefde141e7
-size 139535840

scene_assets/scene4_no_ceiling.glb

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:de5f487e0085d2bea34445392656324dcb10e53b0c87c0ad48c1e1b8eec6f14b
-size 51743004

ui_components.py

@@ -1,8 +1,7 @@
 # ui_components.py
 # Gradio界面相关和辅助函数
 import gradio as gr
-import os
-from config import SCENE_CONFIGS, EPISODE_CONFIGS
+from config import SCENE_CONFIGS
 from logging_utils import read_logs, format_logs_for_display
 
 def update_history_display(history: list) -> list:
@@ -30,28 +29,9 @@ def update_scene_display(scene: str):
     config = SCENE_CONFIGS.get(scene, {})
     desc = config.get("description", "No Description")
     objects = "、".join(config.get("objects", []))
-    glb_path = config.get("glb_path", "")
+    image = config.get("preview_image", None)
     markdown = f"**{desc}**  \nPlaces Included: {objects}"
-    
-    # Validate if file path exists
-    if not os.path.exists(glb_path):
-        return markdown, None
-    
-    return markdown, glb_path
-
-def update_episode_display(scene: str, episode: str):
-    config = SCENE_CONFIGS.get(scene, {})
-    scene_name = config.get("scene_name", scene)  # 使用配置中的scene_name
-    episode_id = int(episode[-1])
-    image_path = os.path.join("scene_assets", f"{scene_name}_{episode_id-1}.jpg")
-    
-    # Validate if file path exists
-    if os.path.exists(image_path):
-        return image_path
-    else:
-        # Fallback to preview_image if episode image doesn't exist
-        fallback_path = config.get("preview_image", None)
-        return fallback_path
+    return markdown, image
 
 def get_scene_instruction(scene: str):
     """根据场景获取默认指令"""