Framework for developing single page web application and powerful web service


Framework for developing single page web application and powerful web service


– Convention over configuration
– Testability by design
– Boost OOP and Functional programming
– Open source (see
– Node.js ws rapid development
– Access to all main databases
– Database indipendent programming
– Single statement read/write of entire datasets
– Extended business logic management
– Deep integration of user environment and security rules
– Pure HTML design, no preprocessing, no injections
– Rich client javascript pages, no postbacks
– No-code pages for default Behaviours
– Rapid development of complex UI
– Seemless integration with MyKode Backend
– Easy creation of new custom controls
– Autamtic management of complex dataset

MyKode Backend

myKode Backend is a set of Node.js classes accompanied by a minimal backend designed to minimize the development time of a Node.js application.

The framework’s classes are designed to access various relational databases interchangeably. The chosen structure for managing data sets in memory is jsDataSet, a class similar to .NET’s DataSets, where DataRow objects are much like plain objects but retain the original state of the row, i.e., the state after the last acceptChanges (a method that makes changes permanent) or rejectChanges (a method that cancels changes to the row).

The class jsDataQuery is used to represent filters and expressions, and its instances can be applied interchangeably to database tables or any collection of JavaScript objects, especially to DataRow objects of a DataSet.

Classes are provided for managing different applications and databases.


jsApplication is the general prototype of the application at its highest level.

It is possible to integrate or modify its various methods to customize them according to specific needs.


Authentication is managed through JSON Web Tokens (JWT). Each user is associated with a session identified by a code that the client must send in every request, in the encrypted JWT token.

jsApplication handles creating a context (Context) for each request and destroying it when the request is fulfilled. It creates the environment of the connected user with information such as their role and permissions.

It also creates routes and ensures access is granted only to clients providing an appropriate token in the request header unless the route has been specifically configured with the getNoTokenFolders function. getNoTokenFolders is a method of jsApplication defined simply as:

getNoTokenFolders: function(){
    return {
        "auth": true

This allows access (only) to the “auth” folder without requiring a valid token. Even unauthenticated requests must provide a fictitious, “anonymous” token, the value of which is configured in the config/tokenConfig file under the AnonymousToken property.

In the HTTP header of each client request, there must be a line like:

Authorization: Bearer <token>

Where <token> is either the “anonymous” token or a token issued by the application upon authentication.

jsApplication decodes the token for each request and sets the data obtained from the req.headers.authorization property of the request req. This way, all routes can access it.

The token is verified and decoded in the req.auth field (configurable in tokenConfig.options.requestProperty). In general, req.auth will contain an object of type Token. A simple example of an authentication route can be found in the login module.

Upon startup, a jsApplication creates an ExpressApplication and an Express.Router. For each subfolder in the routes/ folder, it reads the files inside it and associates a router with each one that responds to the request /folder/fileName, provided that the file is a node module whose file name does not start with an underscore and exposes a router as its only exported property. This makes it easy to add a new route by exposing it in a node module inside a subfolder of routes, regardless of the type of service (GET/POST) and its parameters.

It also handles authenticating the user and providing an anonymous context if not authenticated yet. For each connected user, it establishes a session identified with a sessionID.

Each database is identified by a dbCode, and jsApplication is associated with a database and creates a connection pool to make access more efficient. When a request arrives, a connection is taken from the pool if one is available.

Data Access

To read/write data from/to the database, invoke stored procedures, or perform any other operation, there are three main classes: DataAccessGetData, and PostData.

  • DataAccess is used for low-level operations, such as sending simple read or write commands.
  • GetData is used to read entire DataSets or parts of them.
  • PostData is used to write entire DataSets.

Business Logic

jsBusinessLogic is the class responsible for invoking business rules. These rules are essentially SQL checks applied to each individual row written to the database. The text of the checks is stored in some configuration tables. The rules are “compiled” by an external tool into database stored procedures and invoked when saving data. If the configuration tables contain no rows, no checks will be applied.


The class jsSecurity is queried by the PostData class during the data saving phase. The rules are stored as textual conditions in the customgroupoperation table.

For details, refer to the jsDoc and the summary.


The DataSets of the application are stored in JSON format in the client/dataset folder. The JSON follows the format used for serializing the jsDataSet class, based on the serialize and deserialize methods.

Serialization also includes the state of DataRow with current values and original values for rows in the modified state.

When creating a DataSet, all DataTables with all necessary DataColumn and DataRelation that link the included DataTables must be inserted. This is especially important for navigation when extracting data using the GetData class.

The basic assumption is that there is a “main” table of the DataSet and a set of “subentity” tables representing its details. Then there are parent tables of the entity and subentity tables, whose rows are read as needed based on the data present in the first set.

The subentity relationship is logically identified by the fact that it cannot exist in any way without the entity (table) of which it is a detail. This must be formalized through keys; the subentity must include in its key fields all key fields of the entity and must be related with a DataRelation in the DataSet. It is possible to have subentities at a lower level without any limit.

Satellite tables can also be present, i.e., tables that are parents of entity and subentity tables.

To create a DataSet, several methods can be used:

Manual creation of a DataSet

You can instantiate a DataSet and then add tables and columns to them. This is a bit cumbersome unless a code generator is used to update the code based on any changes to the table structure.


let d = new DataSet();
let tOrder = d.newTable("order");
tOrder.setDataColumn("name", CType.string);
tOrder.setDataColumn("surname", CType.string);

let tOrderDetail = d.newTable("orderdetail");
tOrderDetail.setDataColumn("name", CType.string);
tOrderDetail.setDataColumn("surname", CType.string);


order_orderdetail", "order", ["idorder"], "orderdetail", ["idorder"]);

Semi-manual creation of a DataSet

It is possible to instantiate a DataSet and then add tables by invoking the createTable() method of a DbDescriptor. This way, the table will automatically have columns, and the primary key set.

Then, add the desired relationships:

let d = new DataSet();
let tOrder = await dbDescriptor.createTable("order");

let tOrderDetail = await dbDescriptor.createTable("orderdetail");

d.newRelation("order_orderdetail", "order", ["idorder"], "orderdetail", ["idorder"]);

Visual creation with the HDSGene tool

It is possible to use a tool, HDSGene, provided by Tempo S.r.l., which integrates with the visual editor of DataSet in Visual Studio. It generates the corresponding JSON in the folder where the DataSet is saved.

Copy the JSON file to the client/datasets folder, assuming the name is dsmeta_order_main.json, and load it with the getDataSet function in the context:

let d = ctx.getDataSet("order", "main");

The convention is that the file name should be dsmeta_[tableName]_[editType].json.


The purpose of the MetaData class is to centralize the knowledge of each table in a specific class, which derives from MetaData.

Its use is not specifically required on the backend, but if the frontend myKode_Frontend is used on the client, which might make use of it, and backend services are implemented, they could in turn use the same metadata (instances of MetaData) as the client.

If backend services are implemented that involve creating or manipulating DataSets beyond simple data reading from the database, it is advisable to use the methods of this class.

In summary, each class deriving from MetaData centralizes information about a table, such as:

  • Validity of row data
  • Defaults for field values when a row is created
  • Calculation schema for auto-increment fields (see the PostData class for more information)
  • Structure of lists on that table: which fields should appear, with what label, sorting, etc.

Details can be found in the MetaData file.

MyKode Frontend

The myKode frontend is a framework that allows the development of sophisticated web forms with minimal code.

The JavaScript frontend is a rich client, enabling interaction without continuous postbacks.

Through services provided by the backend (available in both Node.js and .NET versions), it performs a set of operations that allow “normal” operations on web forms. It also allows easy extensibility in terms of both the visual components contained in the form and the specific interaction functions the form may require.

A generic form is associated with a DataSet (similar to those in ADO.NET), containing a local copy of a set of rows present in the database (or that will be in the future).

By “normal operations,” it is meant, for example:

  • The ability to set up a search in “query by example” mode, entering the data to be compared in the form, and then initiating the search based on the entered data [1].
  • Modify existing data.
  • Insert new data [2].
  • Open “detail” forms (even second or third level) on the “child” rows of the main row.
  • Delete a row and all its related details.
  • Display lists based on the selection of parent rows of the main row to populate fields in the main table.

The basic idea is to have a DataSet (similar to those in ADO.NET) to store the local copy of the data objects being processed and operate on the database based on a set of conventions and properties of the DataSet’s columns, as well as the rows (DataRow) contained in DataTables.

However, the DataSet is not intended to contain any set of rows from the database but data that must adhere to a certain logic.

Main Table and Sub-Entities

In particular, there is a “main” table that contains a “main” processing object.

The main table within the framework is defined as an entity. We will interchangeably refer to the main table or main entity of a form.

For example, it could be a row from an address book table or an order table.

Then (optionally), there are a set of “sub-entities,” representing details of the main row.

In the case of an address book, these could be an “address” table with various addresses for that person (office, fiscal domicile, residence, etc.) or a “phone” table with various phone numbers.

In the case of an order, it could be a “order_detail” table with details of the ordered goods.

Relationship: Entity – Sub-Entity or Sub-Entity – Sub-Entity

The relationship between the entity and sub-entity is not generic but must connect the entire key of the parent table with key fields of the child table.

This logically guarantees that there can never be a row in the sub-entity table not connected to any parent row (entity or sub-entity). A sub-entity is, in fact, considered a detail of the main entity.

The “main” row is the one selected from the search performed with the “query by example” mentioned in point [1] above, or the row created in point [2].

There may also be other referenced tables in the DataSet, parents of entities and sub-entities, which are not subject to modification by the form. During user data editing, the distinction between these two sets of tables is crucial. The tables subject to editing (entities and sub-entities) will never be re-read from the database while the user interacts with the form, to avoid losing any modifications being made. Similarly, a row (DataRow) in the insertion state (entity or sub-entity) would be lost if that table were accidentally re-read in the corresponding DataTable.

Conversely, parent tables can be re-read, for example, to select a new parent row for the main row or simply to update them if they have changed in the meantime, as can happen with tables with highly volatile content.

When creating a form, it must be decided which is the main table and which sub-entities can be modified in that form and in its possible detail forms. These tables should be added to the DataSet and related.

Tables that serve to improve the visualization of the first set of tables, typically parent tables, should be added to the DataSet and related by inserting the appropriate DataRelations.

Automatic Reading and Writing of the DataSet

The advantage at this point is that the framework knows how to fill the entire DataSet when the user selects a row from the list. Similarly, if the user decides to delete the main row, the framework already knows which rows in the detail tables to delete. Finally, during insertion, the framework knows the order in which to insert the rows and how to propagate the calculation of incremental key fields from parent tables to child tables.

Conversely, tables that are not entities or sub-entities will never be written during the save operation of a form.

Data Display

The design of a web form is done with simple HTML tags, where attributes specify a data-tag attribute to indicate which field from which table will be displayed in that control.

In each form, fields of the main table are normally visible and editable (unless this behavior is disabled). It is also possible, where a sub-entity is in a 1:1 relationship with the entity (at least in the context of the current view), to display and allow editing of the fields of its row. An example of such a table could be a table containing a taxpayer’s income tax declaration, and it is decided to display, in a certain form, only the row for the current year.

In this case, the income tax table could be filtered by fiscal year, and with this premise, it becomes a 1:1 relationship with the taxpayer table, so the income tax declaration data could be displayed in a form where the main table is the taxpayer.

In other cases, where a table is not a sub-entity and/or not in a 1:1 relationship with the rows of the main table, it will not be possible to display the fields of that table in “simple” controls, such as input-text or similar, but only in HTML tables, which allow displaying multiple rows.

Thanks to these premises, the framework automatically fills in the fields of the web form that have the data-tag attribute, which is of the form “table.field”.

Similarly, when necessary, the framework can read the data from the form and update it in the appropriate rows of the DataSet.

For further details on composing the HTML of a page, refer to MetaPage HTML.

Data Modification Cycle

It is also possible to add any behavior to the form, and, in essence, it is not necessary to read or write to HTML controls.

In fact, the convention for reading or writing form data is to use the following scheme:

  1. Invoke the getFormData() method of MetaPage, which reads the form data, updating the DataSet’s content.
  2. Operate on the DataRow of the DataSet at will, modifying or inserting data into the entity and sub-entity tables.
  3. Invoke the freshForm() method of MetaPage to display the DataSet’s data in the web form.

This way, the code becomes independent of the specific knowledge of which control and what type contains each field that needs to be processed.

Application Structure

Usually, each table is associated with a class derived from MetaData (the “metadata”). The metadata contains all the information about the table

in a centralized manner, as we will see shortly.

If the table is subject to modification on some page, each page is built using an HTML file and a class derived from MetaPage. A table might be viewable or editable through different pages, so there is a need to assign a code to each, which is defined as editType.

Therefore, the tableName-editType pair identifies a form in the context of the application, where tableName is the main table of that form (and the underlying DataSet).

If the table appears in some list viewable by the user, the field names and characteristics of the list are described in the metadata. Since, in this case as well, there may be a need to list a table (or view) in different ways depending on the context, each list is associated with a code, called listingType.

Therefore, the tableName-listingType pair identifies a type of list in the context of the application.


The highest-level element of a myKode application is usually the MetaApp class, which is responsible for registering and providing all the pages (MetaPage) and Metadata.

MetaApp manages the stack of pages that are successively opened and closed and the passing of data between them.

Usually, it is not necessary to derive a subclass, but it is possible to customize it to change the folder where it retrieves the files and metadata on the server.

It is necessary to register all pages and metadata with the methods of MetaApp. In particular, the method

addMeta({string}tableName, {[MetaData](} Meta)

is used to associate the name of a table with the constructor of its metadata, and similarly,

addMetaPage({string}tableName, {string} editType, {[MetaPage](}metaPage)

is used to associate the tableName-editType pair with a page (MetaPage).


Metadata is a JavaScript class that describes the properties of a table.

It should be noted that these properties are set as attributes of DataTables and DataColumns of the DataSet. The DataSet contains all the information about all the tables that compose it.

Using the metadata to set them, instead of repeating these settings in other places, such as in all forms that should use these tables, is crucial to avoid repeating code and make their maintenance effective.

Moreover, this information is used by the framework whenever it is needed, such as creating new rows in a table or validating the data of a row.

Since there is only one place where we describe these properties, it will be more convenient both to retrieve them and to change them safely.


The MetaPage is the class that contains the “common” code for all pages, such as control management, toolbar command management, filling the form with dataset data (freshForm), and reading form data into the DataSet (getFormData).

MetaPage is usually derived to implement all user pages and offers specific methods (hooks) to integrate the base behavior with specific functions of each page.

Therefore, if classes derived from MetaData contain general information about each individual table, shared by the entire application, classes derived from MetaPage describe the behavior of each individual page.

Each web page managed by the myKode frontend is divided (at least) into two parts: the HTML and the corresponding MetaPage, which is a class derived from MetaPage and has the peculiarities of that page.

The frontend can, if necessary, access many backend services to read/write data, generally using objects of type sqlFun for filters and DataSet to manage the data.

Progetto EOLO - Veicolo per nuovi contenuti
Tempo S.R.L.

POR PUGLIA FESR 2014-2020 - Titolo II - Capo 1 “Aiuti ai programmi di investimento delle grandi imprese - AD n.799 del 07.05.15 “Avviso per la presentazione di programmi promossi da Grandi Imprese ai sensi dell’art. 17 del Reg generale dei regimi di aiuto in esenzione n.17 del 30/09/14”