CN115639993A - Visual modeling platform and method based on WEB dynamic layout - Google Patents
Visual modeling platform and method based on WEB dynamic layout Download PDFInfo
- Publication number
- CN115639993A CN115639993A CN202211234122.8A CN202211234122A CN115639993A CN 115639993 A CN115639993 A CN 115639993A CN 202211234122 A CN202211234122 A CN 202211234122A CN 115639993 A CN115639993 A CN 115639993A
- Authority
- CN
- China
- Prior art keywords
- canvas
- operator
- management
- modeling
- component
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a visual modeling platform and a method based on WEB dynamic layout, wherein the platform comprises: the system comprises a modeling component module, a canvas space module and an operation management module, wherein the modeling component module is used for operator component management and a model component module, the canvas space module is used for component display, canvas editing, canvas layout arrangement and component parameter configuration, and the operation management module is used for operation management, task and record management and data visualization. The invention can finish the complex data analysis and the construction of the modeling workflow by dragging and connecting the component nodes in the canvas and convenient component parameter configuration, and meanwhile, the dynamic layout function of the canvas can facilitate the user to freely and flexibly adjust the layout in the canvas according to personal working equipment so as to improve the efficiency of constructing the workflow.
Description
Technical Field
The invention relates to the technical field of visualization and machine learning modeling, in particular to a visualization modeling platform and a visualization modeling method based on WEB dynamic layout.
Background
With the continuous accumulation of big data, the data gradually becomes the core assets of enterprises, in order to maximize the value of the data, more and more enterprises put efforts into the data mining plate, and the influence is that the cost of enterprise investment is higher and higher, but the generated value is far from reaching the ideal state, which is the pain point that needs to be solved urgently. On the other hand, in recent years, the technology of artificial intelligence mining for big data has been developed rapidly, and data analysis modelers need to learn more and more technologies, such as: python, R, scale, etc. of programming language class; mySQL, postgreSQL, hive, hbase and the like of the data query class; scikit-leann, tensorflow, pythrch and the like of the algorithm modeling class; hadoop, spark, flink and the like of distributed storage computing class; it is anticipated that as technology evolves at a high rate, more computing frameworks will emerge and adapting to these computing frameworks will increase the modeling cost of the enterprise. Meanwhile, the artificial intelligence modeling also needs professional algorithm engineers to participate, the algorithm engineers need to deeply understand the service scene and try to use various algorithm models under different computing frames, for example, when Tensorflow in the field of deep learning is researched, distributed computing frames Spark and Flink are difficult to use within a limited time, and when Spark is good, flink is not necessarily known.
Based on the above, the invention designs a visual modeling platform and a visual modeling method based on WEB dynamic layout, so as to solve the above problems.
Disclosure of Invention
The invention aims to provide a visual modeling platform and a visual modeling method based on WEB dynamic layout, so as to solve the problems in the background technology.
In order to achieve the purpose, the invention provides the following technical scheme: a visual modeling platform based on WEB dynamic layout comprises: a modeling component, a canvas space, an execution management module, an operator component, a model component, and a canvas workflow, wherein,
the modeling component is used for managing the operator component and the model component;
the canvas space is used for displaying and dragging the modeling assembly, and constructing and dynamically laying out the workflow in the canvas area;
the operation management module is used for operation scheduling, task management, state monitoring, data storage and result visualization of a modeling component or a canvas workflow in a canvas.
Preferably, the modeling component is used for managing the operator component, and the management comprises uploading and downloading, auditing management, category management, storage management, operator information display and online development of a custom operator.
Preferably, the modeling component is used for managing the model component, including uploading and downloading, auditing management, category management, storage management, environment management and model information display.
Preferably, the canvas space is further used for tree display of the operator component and the model component, and the canvas space is further used for dragging, connecting, amplifying, shrinking, deleting right mouse button, deleting frame selection, canceling, saving canvas workflow and generating a model of the model component in a canvas area.
Preferably, the dynamic layout of the workflow in the canvas area comprises horizontal switching, vertical switching and one-key programming.
Preferably, the canvas space is further used for single-step running of modeling components in the canvas area, one-key foreground running, one-key background running and suspension running of workflows in the canvas.
Preferably, the canvas space is further used for dynamically displaying running state icons of the modeling components in the canvas area, displaying running logs in real time, visualizing running result tables, visualizing running result charts and downloading results.
Preferably, the canvas space is also used for parameter configuration and persistence of the modeling component.
Preferably, the operation management module is further configured to perform operation scheduling management on a single Python machine learning operator and a model of the modeling component, perform operation scheduling management on a distributed Spark operator and a model, perform operation scheduling management on a distributed Flink operator and a model, perform operation record and log management, and perform historical task display and playback.
Another objective of the present invention is to provide a visual modeling method based on WEB dynamic layout, which is based on the above visual modeling platform based on WEB dynamic layout, and the specific steps thereof are as follows:
the first step is as follows: uploading an operator component package in operator component management, auditing the operator component package, judging an auditing result, filing the operator which passes the auditing to an operator library of a corresponding category, and providing online modification and deletion functions for the operator components which do not pass the auditing;
the second step is that: creating a data analysis modeling editing area in a canvas space module through a newly-built canvas, entering the canvas, finding out an operator uploaded in the first step through an operator component library on the left side by a user, dragging the required operator into the canvas, connecting lines, clicking the operator needing parameter configuration, and popping up an operator configuration area on the right side of the canvas for parameter configuration and modification;
the third step: when the number of operators in the canvas of the second step is more and the typesetting is disordered, the one-key typesetting function of the menu bar of the canvas can be clicked to carry out automatic typesetting;
the fourth step: when a user finds that the current display is not suitable for building the canvas workflow in a longitudinal layout mode, the user can click the canvas menu bar to switch the layout function, and then the horizontal layout can be quickly switched to work;
the fifth step: when a user finds that the current display is not suitable for constructing the canvas workflow in a transverse layout mode, the user can click the canvas menu bar to switch the layout function, and then the longitudinal layout can be quickly switched to work;
and a sixth step: after the canvas workflow is constructed, a user can perform right-click operator name adding operation on each operator, can define a service name and an algorithm name for the current operator and perform persistence operation, and is convenient for rapidly understanding the service content and the development thought of the current canvas workflow when the workflow is used subsequently;
the seventh step: after the canvas workflow is configured and built, the components in the canvas are operated, and the operation comprises the following steps: one-key foreground operation, one-key background operation and single-step operation, after the execution, calling an operation management model to carry out task scheduling, wherein computing platforms supported by the task scheduling comprise Sciki kit-spare, tensorflow, pythrch, spark and Flink;
eighth step: the canvas space module and the operation management module carry out real-time interaction to obtain the operation state, the operation log and the operation result of an operator assembly or a canvas workflow in the canvas;
the ninth step: after the canvas space module acquires the data returned by the operation management module, the state chart display, the log dynamic display, the table visualization of the operation result and the chart visualization can be carried out.
Compared with the prior art, the invention has the beneficial effects that: based on different user habits and preferences, the invention realizes the function of flexibly and freely switching page layout modes (horizontal and vertical) and realizing automatic arrangement by one key, and helps the user to improve the efficiency of data analysis and modeling exploration on the WEB page.
Drawings
FIG. 1 is a block diagram of the system of the present invention.
FIG. 2 is a flow chart of the method of the present invention.
FIG. 3 is a vertical layout diagram of the canvas of the present invention.
FIG. 4 is a canvas landscape layout diagram according to the present invention.
FIG. 5 is an automatic canvas layout diagram according to the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
Referring to fig. 1, the present invention provides a technical solution: a visual modeling platform based on WEB dynamic layout comprises: the system comprises a modeling component, a canvas space, an operation management module, an operator component, a model component and a canvas workflow, wherein the modeling component is used for managing the operator component and managing the model component, the modeling component is used for managing the operator component and comprises uploading and downloading, auditing management, category filing and displaying and on-line development of a user-defined operator, and the modeling component is used for managing the model component and comprises uploading and downloading, auditing management, category management, storage management, environment management and model information display. The modeling component module is realized by adopting a Vue + Python technology stack, specifically, for an operator component, an operator is realized by the Python technology stack at the rear end and packaged into an independent operator execution file, configuration information of the operator is packaged into a Json file, the execution file and the configuration file are assembled into a component package and stored into a distributed file system, a group of WEB service interfaces are used for providing services for the outside by using FastAPI, the front end acquires calculation sub-tree information by requesting the WEB interface provided by the rear end, then the Vue + Antv-x6 components are used for rendering in a browser, and for the realization of the model component, the realization mode of the operator component can be referred;
the canvas space is used for displaying and dragging a modeling component, constructing and dynamically laying out workflow in a canvas area, the canvas space is also used for tree-shaped displaying and dragging of an operator component and a model component, the canvas space is also used for dragging, connecting, deleting, canceling, storing and generating the model of the modeling component in the canvas area, the dynamic laying out of the workflow in the canvas area comprises transverse switching, longitudinal switching and one-key arrangement, the canvas space is also used for single-step operation of the modeling component in the canvas area, one-key foreground operation, one-key background operation and suspension operation of the canvas workflow, the canvas space is also used for dynamic displaying of an operation state icon, real-time displaying of an operation log, visualizing of an operation result table, visualizing of an operation result chart and downloading of the result, the canvas space is also used for parameter configuration and persistence of the modeling component, and the canvas space module of the scheme is realized by adopting Vue + Antv-x6 + JS technology, and specifically:
the canvas creating function is that by introducing an Antv-x6 component, the size, style, operation rule and other attributes of the canvas are configured according to official configuration information and are rendered to a browser;
and the operator tree display function is realized by acquiring back-end interface data and binding a dragging event of the Antv-x6 component with the operator tree to arbitrarily drag an operator of the operator tree species as a node to be placed in the canvas.
The canvas node rendering function is used for rendering the operator information into a node style designed by a UI (user interface) through a dragging event of an Antv-x6 component, wherein the node style comprises an icon, a name, an operating state, an input/output port, a theme color and an arrangement mode;
the switching layout arrangement function is that according to an Antv-x6 node rendering method, a transverse and longitudinal node generation mode is firstly formulated and encapsulated into a function component, finally json in canvas is changed by means of js circular traversal through button switching, and the changed json is rendered to realize transverse and longitudinal switching;
the one-key automatic layout function is that a layout mode, the distance between operators, the trend of the operators and an alignment mode are set by utilizing an antv-x6 built-in new DagreLayout method, and a canvas is rendered by a graph.
The one-key operation function is that nodes in the canvas form a directed acyclic graph through connecting lines, all node information in the canvas is assembled into a Json by clicking one-key operation, and the Json is sent to the operation service of the back end through an interface provided by the back end;
and the right-click menu function is to obtain the position of the mouse in the right-click time of the node according to the designed menu content, judge and subtract the size of the side bar so as to determine the horizontal and vertical coordinates displayed by the right-click menu, and finally remove an event through the moving-in of the mouse, wherein when the mouse leaves the right-click menu, the menu is hidden.
And the operation management module is used for operation scheduling, task management, state monitoring and result visualization of a modeling component or a canvas workflow in the canvas, and is also used for operation scheduling management of a single Python machine learning operator and a model of the modeling component, operation scheduling management of a distributed Spark operator and the model, operation scheduling management of a distributed Flink operator and the model, operation record and log management, and historical task display and playback. The operation management module of this scheme adopts Java + SrpingBoot + Airflow + Yang's technical stack to realize, specifically sees: the method comprises the steps of using Java and SpringBoot to realize a visual modeling platform, receiving a front-end operation request by a scheduling system, judging operation types, using an operator or a model realized by a Python stack, performing scheduling management by using Airflow, using Yarn to perform scheduling management on the operator or the model realized by the distributed Spark and Flink stacks, and then storing results after executing different computing engines into a distributed file system.
Referring to fig. 2, the present invention provides a visual modeling method based on WEB dynamic layout, which includes the following specific steps: the first step is as follows: uploading an operator component package in operator component management, auditing the operator component package, judging an auditing result, filing the operator which passes the auditing to an operator library of a corresponding category, and providing online modification and deletion functions for the operator components which do not pass the auditing;
the second step is that: creating a data analysis modeling task in a canvas center module through a newly-built canvas, entering the canvas, finding an operator uploaded in the first step through an operator component library on the left side by a user, dragging the required operator into the canvas and connecting the operator, clicking the operator needing parameter configuration, and popping up an operator configuration area on the right side of the canvas for parameter configuration and modification;
the third step: when the number of operators in the canvas of the second step is more and the typesetting is disordered, the one-key typesetting function of the menu bar of the canvas can be clicked to carry out automatic typesetting;
the fourth step: when a user finds that the current display is not suitable for constructing the canvas workflow in a longitudinal layout mode, the user can click the canvas menu bar to switch the layout function, and then the horizontal layout can be quickly switched to work;
the fifth step: when a user finds that the current display is not suitable for constructing the canvas workflow in a transverse layout mode, the user can click the canvas menu bar to switch the layout function, and then the longitudinal layout can be quickly switched to work;
and a sixth step: after the canvas workflow is constructed, a user can perform right-key operator name adding operation on each operator, can define a service name and an algorithm name for the current operator and perform persistence operation, and is convenient for rapidly understanding the service content and the development idea of the current canvas workflow when the workflow is subsequently used;
the seventh step: after the canvas workflow is configured and built, the components in the canvas are operated, and the operation comprises the following steps: one-key foreground operation, one-key background operation and single-step operation, after the execution, an operation management model is called to carry out task scheduling, and computing platforms supported by the task scheduling comprise Scikt-lean, tensflow, pythrch, spark and Flink;
eighth step: the canvas center module and the operation management module carry out real-time interaction and are used for acquiring the operation state, the operation log and the operation result of an operator component or a canvas workflow in the canvas;
the ninth step: after the canvas center module acquires the data returned by the operation management module, the state chart display, the log dynamic display, the table visualization of the operation result and the chart visualization can be carried out.
In the description herein, references to the description of "one embodiment," "an example," "a specific example" or the like are intended to mean that a particular feature, structure, material, or characteristic described in connection with the embodiment or example is included in at least one embodiment or example of the invention. In this specification, the schematic representations of the terms used above do not necessarily refer to the same embodiment or example. Furthermore, the particular features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples.
The preferred embodiments of the invention disclosed above are intended to be illustrative only. The preferred embodiments are not intended to be exhaustive or to limit the invention to the precise embodiments disclosed. Obviously, many modifications and variations are possible in light of the above teaching. The embodiments were chosen and described in order to best explain the principles of the invention and the practical application, to thereby enable others skilled in the art to best utilize the invention. The invention is limited only by the claims and their full scope and equivalents.
Claims (9)
1. A visual modeling platform based on WEB dynamic layout is characterized in that the modeling platform comprises: a modeling component, a canvas space, an execution management module, an operator component, a model component, and a canvas workflow, wherein,
the modeling component is used for managing operator components and model components;
the canvas space is used for displaying and dragging the modeling assembly, and constructing and dynamically laying out the workflow in the canvas area; the canvas space is also used for tree-shaped display of the operator assembly and the model assembly, and is also used for dragging, connecting, amplifying and reducing the model assembly in a canvas area, deleting a right mouse button, deleting a box selection, canceling, saving a canvas workflow and generating the canvas workflow;
the operation management module is used for operation scheduling, task management, state monitoring, data storage and result visualization of a modeling component or a canvas workflow in a canvas.
2. The visual modeling platform based on WEB dynamic layout according to claim 1, wherein the modeling component is used for managing operator components including uploading and downloading, auditing management, category management, storage management, operator information display and online development of custom operators.
3. The visual modeling platform based on WEB dynamic layout according to claim 1, wherein the modeling component is used for management of model components including uploading and downloading, auditing management, category management, storage management, environment management and model information display.
4. The visualization modeling platform based on WEB dynamic layout of claim 1, wherein the workflow dynamic layout in the layout area includes horizontal switching, vertical switching and one-key layout.
5. The visual modeling platform based on WEB dynamic layout according to claim 1, wherein the canvas space is further used for single step operation of modeling components in the canvas area, one-key foreground operation, one-key background operation and pause operation of workflows in the canvas.
6. The visual modeling platform based on WEB dynamic layout according to claim 1, wherein the canvas space is further used for dynamic display of running state icons of modeling components in the canvas area, real-time display of running logs, visualization of running result tables, visualization of running result charts and result downloading.
7. The visualization modeling platform based on WEB dynamic layout according to claim 1, wherein the canvas space is further used for parameter configuration and persistence of modeling components.
8. The visual modeling platform based on WEB dynamic layout according to claim 1, wherein the operation management module is further configured to perform operation scheduling management of a standalone Python machine learning operator and model of the modeling component, operation scheduling management of a distributed Spark operator and model, operation scheduling management of a distributed Flink operator and model, operation recording and log management, and historical task display and playback.
9. A visual modeling method based on WEB dynamic layout is characterized in that the method is a method using the visual modeling platform of any one of claims 1 to 8, and the specific steps are as follows:
the first step is as follows: uploading an operator assembly package in operator assembly management, auditing the operator assembly package, judging an auditing result, filing the operator which passes the auditing into an operator library of a corresponding category, and providing online modification and deletion functions for the operator assembly which does not pass the auditing;
the second step is that: creating a data analysis and modeling editing area in a canvas space module through a newly-built canvas, entering the canvas, finding out an operator uploaded in the first step through an operator component library on the left side by a user, dragging the required operator into the canvas and connecting the operator, clicking the operator needing parameter configuration, and popping up an operator configuration area on the right side of the canvas for parameter configuration and modification;
the third step: when the number of operators in the canvas is more and the typesetting is disordered, the one-key arrangement function of the menu bar of the canvas can be clicked to carry out automatic typesetting;
the fourth step: when a user finds that the current display is not suitable for building the canvas workflow in a longitudinal layout mode, the user can click the canvas menu bar to switch the layout function, and then the horizontal layout can be quickly switched to work;
the fifth step: when a user finds that the current display is not suitable for constructing the canvas workflow in a transverse layout mode, the user can click the canvas menu bar to switch the layout function, and then the longitudinal layout can be quickly switched to work;
and a sixth step: after the canvas workflow is constructed, a user can perform right-key operator name adding operation on each operator, can define a service name and an algorithm name for the current operator and perform persistence operation, and is convenient for rapidly understanding the service content and the development idea of the current canvas workflow when the workflow is subsequently used;
the seventh step: after the canvas workflow is configured and built, the components in the canvas are operated, and the operation comprises the following steps: one-key foreground operation, one-key background operation and single-step operation, after the execution, an operation management model is called to carry out task scheduling, and computing platforms supported by the task scheduling comprise Scikt-lean, tensflow, pythrch, spark and Flink;
eighth step: the canvas center module and the operation management module carry out real-time interaction and are used for acquiring the operation state, the operation log and the operation result of an operator component or a canvas workflow in the canvas;
the ninth step: after the canvas center module acquires the data returned by the operation management module, the state chart display, the log dynamic display, the table visualization of the operation result and the chart visualization can be carried out.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211234122.8A CN115639993A (en) | 2022-10-10 | 2022-10-10 | Visual modeling platform and method based on WEB dynamic layout |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211234122.8A CN115639993A (en) | 2022-10-10 | 2022-10-10 | Visual modeling platform and method based on WEB dynamic layout |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115639993A true CN115639993A (en) | 2023-01-24 |
Family
ID=84941942
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211234122.8A Pending CN115639993A (en) | 2022-10-10 | 2022-10-10 | Visual modeling platform and method based on WEB dynamic layout |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115639993A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117519838A (en) * | 2024-01-04 | 2024-02-06 | 浙江大华技术股份有限公司 | AI workflow modeling method, related device, equipment, system and medium |
| CN117932972A (en) * | 2024-03-15 | 2024-04-26 | 南京凯奥思数据技术有限公司 | Visual modeling platform and method applied to equipment state algorithm model based on WEB |
-
2022
- 2022-10-10 CN CN202211234122.8A patent/CN115639993A/en active Pending
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117519838A (en) * | 2024-01-04 | 2024-02-06 | 浙江大华技术股份有限公司 | AI workflow modeling method, related device, equipment, system and medium |
| CN117519838B (en) * | 2024-01-04 | 2024-04-12 | 浙江大华技术股份有限公司 | AI workflow modeling method, related device, equipment, system and medium |
| CN117932972A (en) * | 2024-03-15 | 2024-04-26 | 南京凯奥思数据技术有限公司 | Visual modeling platform and method applied to equipment state algorithm model based on WEB |
| CN117932972B (en) * | 2024-03-15 | 2024-05-28 | 南京凯奥思数据技术有限公司 | Visual modeling platform and method applied to equipment state algorithm model based on WEB |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN115639993A (en) | Visual modeling platform and method based on WEB dynamic layout | |
| US9208137B2 (en) | Method to preview an undo/redo list | |
| CN105739983B (en) | Shell script editing device and its implementation | |
| CN109657214A (en) | Report form generation method, device, terminal and storage medium | |
| CN107704608A (en) | A kind of OLAP multidimensional analyses and data digging system | |
| US7272492B2 (en) | Path planning for mine countermeasures | |
| CN102576281B (en) | Methods for interacting with and manipulating information and systems thereof | |
| CN111079244B (en) | Distributed simulation platform based on behavior tree | |
| CN110807015A (en) | Big data asset value delivery management method and system | |
| CN109857870A (en) | Visual Ontology for knowledge mapping constructs system and method | |
| CN105808244A (en) | Dynamic data visualized analysis tool based on workflows | |
| CN108446381A (en) | A kind of data visualisation system and method based on Web | |
| CN103413002A (en) | Topological graph editing method and topological graph editor | |
| CN110673844A (en) | Image processing software development method and system | |
| CN114816189A (en) | Cloud-based industrial intelligent equipment remote configuration method | |
| US20090007010A1 (en) | Hosted data visualization service | |
| Reiss | Bee/hive: A software visualization back end | |
| CN116402157A (en) | Drag type machine learning node visual dispatching platform and method thereof | |
| Paterno' et al. | A Semantics‐based Approach for the Design and Implementation of Interaction Objects | |
| CN114564199A (en) | Method, device and equipment for generating use page and readable storage medium | |
| CN113626024A (en) | Low code development method and device combining RPA and AI and computing equipment | |
| US20110161846A1 (en) | Highlighting changes on computer diagrams | |
| Cruz et al. | Delaunay: A database visualization system | |
| CN101976196B (en) | Quality of service oriented code automatic code generating method | |
| KR102537489B1 (en) | Digital twin constructing system, method and device therefor using visual script |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |