MCPcopy Create free account

hub / github.com/IPADS-SAI/MobiAgent / functions

Functions923 in github.com/IPADS-SAI/MobiAgent

⨍ Functions 923◇ Types & classes 124↳ Endpoints 16
↓ 2 callersMethodinputTextWithClipboard
(String text, AccessibilityNodeInfo targetNode)
app/app/src/main/java/com/mobi/agent/MyAccessibilityService.java:2106
↓ 2 callersFunctionint_box_area
(box, w, h)
utils/omni_utils.py:385
↓ 2 callersFunctionint_box_area
(box, w, h)
agent_rr/omniparser/utils.py:386
↓ 2 callersMethodinterruptAllActiveThreads
中断所有活跃线程
app/app/src/main/java/com/mobi/agent/MainActivity.java:2326
↓ 2 callersMethodisReady
()
app/app/src/main/java/com/mobi/agent/ScreenCaptureService.java:453
↓ 2 callersMethodis_configured
检查日志系统是否已配置
MobiFlow/avdag/logger.py:325
↓ 2 callersFunctionload_augmentation_rules
读取数据扩充配置文件，返回规则列表
collect/construct_sft.py:14
↓ 2 callersFunctionload_task
(path: str)
MobiFlow/avdag/loader.py:25
↓ 2 callersMethodlogd
Debug日志
app/app/src/main/java/com/mobi/agent/CommonUtils.java:72
↓ 2 callersMethodonError
(String error)
app/app/src/main/java/com/mobi/agent/ScreenshotManager.java:36
↓ 2 callersFunctionparseUIElements
(hierarchyXml)
collect/manual/static/js/script.js:1353
↓ 2 callersMethodparse_response
解析模型响应，使用UI-TARS官方完整流程： 1. parse_action_to_structure_output - 解析响应为结构化数据 2. parsing_response_to_pyautogui_code - 转换为PyAutoGUI
runner/UI-TARS-agent/ui_tars_automation/action_parser.py:22
↓ 2 callersMethodquery
查询索引以查找最相关的模板。
utils/local_experience.py:165
↓ 2 callersFunctionrelative_bbox
(bbox, width, height)
collect/construct_sft.py:191
↓ 2 callersFunctionreranker_main
(train_path, test_path)
agent_rr/train/prepare_data.py:208
↓ 2 callersFunctionresetUIState
()
collect/manual/static/js/script.js:1137
↓ 2 callersMethodreset_cnt
(self)
agent_rr/run_experiment.py:54
↓ 2 callersMethodreset_cnt
(self)
agent_rr/run_experiment.py:83
↓ 2 callersMethodreset_counter
(self)
agent_rr/action_cache/tree.py:337
↓ 2 callersMethodreset_cur_task
(self)
agent_rr/run_experiment.py:88
↓ 2 callersMethodresolve_icon_path
解析图标路径 Args: icon_name: 图标名称 app_id: 应用ID，用于确定子目录 Returns: 图标文件的完整路
utils/icon_detection.py:267
↓ 2 callersFunctionround_by_factor
Returns the closest integer to 'number' that is divisible by 'factor'.
runner/UI-TARS-agent/ui_tars_automation/coordinate_processor.py:23
↓ 2 callersMethodrun
()
app/app/src/main/java/com/mobi/agent/MyAccessibilityService.java:159
↓ 2 callersFunctionsaveCurrentData
()
collect/manual/static/js/script.js:941
↓ 2 callersMethodscreenshot
(self, path)
runner/mobiagent/multi_task/mobiagent_refactored.py:56
↓ 2 callersFunctionsendTaskDescription
(description)
collect/manual/static/js/script.js:1112
↓ 2 callersMethodshowToast
统一的Toast显示方法 @param context 上下文 @param message 消息内容 @param duration Toast显示时长
app/app/src/main/java/com/mobi/agent/CommonUtils.java:18
↓ 2 callersFunctionshow_usage
显示使用说明
MobiFlow/universal_test_runner.py:564
↓ 2 callersFunctionsingle_domain_reranker
(domain_dir)
agent_rr/train/prepare_data.py:175
↓ 2 callersFunctionsingle_split_embedding
(path)
agent_rr/train/prepare_data.py:24
↓ 2 callersMethodstartCapture
()
app/app/src/main/java/com/mobi/agent/ScreenCaptureService.java:200
↓ 2 callersMethodstartForegroundService
()
app/app/src/main/java/com/mobi/agent/ScreenCaptureService.java:122
↓ 2 callersMethodswipe
(self, direction)
runner/mobiagent/multi_task/mobiagent_refactored.py:68
↓ 2 callersFunctiontakeScreenshot
()
collect/manual/static/js/script.js:349
↓ 2 callersFunctiontask_in_app
执行单个子任务（在单个APP内） Args: app: 应用名称 old_task: 原始任务描述 task: 当前任务描述 device: 设备对象 data_dir: 数据存储目录
runner/mobiagent/multi_task/executor.py:195
↓ 2 callersMethodtest_single_trace
测试单个trace
MobiFlow/universal_test_runner.py:330
↓ 2 callersMethodunlock
Unlock the Android device
collect/manual/device.py:128
↓ 2 callersMethodupdate_icon_detection_config
更新图标检测配置
utils/config.py:213
↓ 2 callersMethodupdate_ocr_config
更新OCR配置
utils/config.py:205
↓ 2 callersFunctionvalidate_action
(action_type, param)
collect/construct_sft.py:114
↓ 2 callersFunctionverify
给定帧与任务，判断是否存在符合拓扑依赖的满足路径。 算法：拓扑序 DP，记录每个节点的最小可行帧索引（满足依赖且该节点匹配）。
MobiFlow/avdag/verifier.py:293
↓ 1 callersFunctionIoU
(box1, box2, return_max=True)
utils/box_annotator.py:175
↓ 1 callersFunctionIoU
(box1, box2, return_max=True)
agent_rr/omniparser/box_annotator.py:175
↓ 1 callersMethod__init__
(self, adb_endpoint=None)
runner/mobiagent/mobiagent.py:80
↓ 1 callersMethod__init__
(self, name: str, param: Dict[str, str], extra: Dict[str, str] = None)
agent_rr/action_cache/action.py:31
↓ 1 callersMethod__init__
(self)
agent_rr/agent/agent.py:7
↓ 1 callersMethod__init__
(self, name: str)
MobiFlow/avdag/logger.py:64
↓ 1 callersMethod__str__
(self)
agent_rr/action_cache/tree.py:33
↓ 1 callersMethod_analyze_with_llm
使用LLM分析执行记录，提取任务类型和用户偏好 Args: task_data: 任务数据 Returns: 包含task_type和preferences的
runner/mobiagent/user_preference_extractor.py:145
↓ 1 callersMethod_assert_acyclic
(self)
MobiFlow/avdag/dag.py:33
↓ 1 callersMethod_capture_screenshot_and_data
截图并保存相关数据
runner/UI-TARS-agent/ui_tars_automation/framework.py:65
↓ 1 callersFunction_check_tesseract_installation
()
utils/ocr_engine.py:33
↓ 1 callersFunction_collect_candidates
兼容性包装：调用路径感知的候选帧收集函数
MobiFlow/avdag/verifier.py:250
↓ 1 callersFunction_collect_candidates_path_aware
路径感知的候选帧收集：按拓扑顺序检查节点，支持路径剪枝和动态帧分配。 - 按拓扑顺序逐节点检查； - 路径剪枝：只为当前可达路径上的节点分配帧资源； - 动态帧分配：根据已成功的路径动态决定下一个要检查的节点； - 帧独占优化：针对每个路径分支独立管理帧使用，避免跨
MobiFlow/avdag/verifier.py:17
↓ 1 callersMethod_config_to_dict
将配置对象转换为字典
utils/config.py:172
↓ 1 callersMethod_convert_pyautogui_to_internal
将UI-TARS官方生成的PyAutoGUI代码转换为框架内部格式
runner/UI-TARS-agent/ui_tars_automation/action_parser.py:168
↓ 1 callersFunction_create_click_visualization
创建点击动作的可视化图像
runner/mobiagent/multi_task/executor.py:490
↓ 1 callersMethod_create_task_directory
创建任务目录
runner/UI-TARS-agent/ui_tars_automation/data_manager.py:27
↓ 1 callersMethod_create_verifier_options
创建验证选项
MobiFlow/universal_test_runner.py:119
↓ 1 callersMethod_dict_to_config
将字典转换为配置对象
utils/config.py:146
↓ 1 callersMethod_enhance_image_for_tesseract
为Tesseract优化图像质量
utils/ocr_engine.py:243
↓ 1 callersMethod_extract_from_file
从单个文件中提取任务描述 Args: file_path: 文件路径 Returns: 任务信息字典或None
MobiFlow/auto_rules/task_extractor.py:47
↓ 1 callersMethod_extract_image_from_frame
从frame中提取图像数据
MobiFlow/avdag/conditions.py:654
↓ 1 callersMethod_extract_keyword
(self, task, action)
agent_rr/action_cache/tree.py:269
↓ 1 callersMethod_extract_raw_action
从响应中提取原始动作字符串
runner/UI-TARS-agent/ui_tars_automation/action_parser.py:288
↓ 1 callersMethod_find_json_end
查找JSON对象的结束位置
utils/local_experience.py:233
↓ 1 callersMethod_find_project_root
查找项目根目录
utils/config.py:51
↓ 1 callersMethod_find_project_root
查找项目根目录
utils/config.py:84
↓ 1 callersMethod_find_project_root
查找项目根目录
utils/config.py:118
↓ 1 callersMethod_find_project_root
查找项目根目录
utils/icon_detection.py:390
↓ 1 callersMethod_find_weights_dir
查找权重目录
utils/weights_manager.py:48
↓ 1 callersMethod_format_config_arrays
格式化配置中的数组字段，确保使用方括号格式 Args: config: 原始配置字典 Returns: 格式化后的配置字典
MobiFlow/auto_rules/llm_generator.py:119
↓ 1 callersMethod_get_detection_service
延迟导入图标检测服务
MobiFlow/avdag/conditions.py:635
↓ 1 callersMethod_get_mem0_config
获取GraphRAG配置 Returns: GraphRAG配置字典
runner/mobiagent/user_preference_extractor.py:61
↓ 1 callersMethod_get_paddle_instance
获取全局PaddleOCR实例
utils/ocr_engine.py:133
↓ 1 callersFunction_get_path_used_frames
获取到达当前节点的路径上已使用的帧集合（用于帧独占逻辑） 重要：当一个帧被使用时，该帧之前的所有帧也被标记为已使用，确保时序的线性关系
MobiFlow/avdag/verifier.py:165
↓ 1 callersMethod_get_rule_file_path
获取规则文件完整路径
MobiFlow/universal_test_runner.py:167
↓ 1 callersMethod_get_traces_for_type
获取指定类型的所有trace
MobiFlow/universal_test_runner.py:259
↓ 1 callersMethod_initialize_client
初始化OpenAI客户端
runner/UI-TARS-agent/ui_tars_automation/framework.py:37
↓ 1 callersMethod_initialize_device
初始化设备连接
runner/UI-TARS-agent/ui_tars_automation/framework.py:49
↓ 1 callersMethod_initialize_engines
初始化OCR引擎
utils/advanced_ocr.py:47
↓ 1 callersMethod_load_config
加载配置
utils/config.py:134
↓ 1 callersMethod_load_config
加载配置文件
MobiFlow/universal_test_runner.py:111
↓ 1 callersFunction_load_from_env
从环境变量加载配置
utils/config.py:268
↓ 1 callersMethod_load_templates
从 JSON 文件加载模板。
utils/local_experience.py:62
↓ 1 callersFunction_min_feasible_index
在 cands 中找出第一个 >= min_required 的索引。若无则返回 None。
MobiFlow/avdag/verifier.py:272
↓ 1 callersMethod_parse_fallback
简单备用解析方法
runner/UI-TARS-agent/ui_tars_automation/action_parser.py:298
↓ 1 callersMethod_parse_llm_response
解析LLM响应
MobiFlow/auto_rules/llm_generator.py:88
↓ 1 callersFunction_parse_node
(d: Dict[str, Any])
MobiFlow/avdag/loader.py:10
↓ 1 callersMethod_preprocess_image
预处理图像 Args: image: 输入图像 Returns: 处理后的灰度图像，失败返回None
utils/icon_detection.py:494
↓ 1 callersFunction_read_file
(path: str)
MobiFlow/avdag/trace_loader.py:10
↓ 1 callersMethod_remove_task
(self, task)
agent_rr/action_cache/tree.py:159
↓ 1 callersFunction_retrieve_with_graphrag
使用GraphRAG检索用户偏好，直接返回检索到的原始偏好文本组成的list
runner/mobiagent/user_preference_extractor.py:315
↓ 1 callersFunction_retrieve_with_vector_search
使用内存向量/检索原始偏好文本list
runner/mobiagent/user_preference_extractor.py:294
↓ 1 callersMethod_store_preferences
存储偏好到Mem0 Args: preferences: 偏好字典 task_type: 任务类型
runner/mobiagent/user_preference_extractor.py:188
↓ 1 callersMethod_supports_color
检查是否支持颜色输出（跨平台）
MobiFlow/avdag/logger.py:124
↓ 1 callersMethod_validate_config
验证配置完整性
MobiFlow/test_model_connectivity.py:59
↓ 1 callersMethod_validate_dependencies_consistency
验证 deps 和 next 定义的一致性，避免混淆和冗余
MobiFlow/avdag/dag.py:49
↓ 1 callersMethodaddStructuredResponse
添加结构化响应消息（包含思维链和执行动作） @param reasoning AI思维链内容 @param action 执行动作类型 @param parameters 动作参数
app/app/src/main/java/com/mobi/agent/MainActivity.java:1714
↓ 1 callersMethodanalyzeInputElement
(AccessibilityNodeInfo element, int index)
app/app/src/main/java/com/mobi/agent/MyAccessibilityService.java:1088
← previousnext →201–300 of 923, ranked by callers