Schemas repository resides in the schema directory.
Every directory inside schema is a schema template, which
contains JSON schemas of the collections in the .json files:
schema/{template name}/{collection name}.json
For example ecommerce and todo templates look like the following:
├── schema
│ ├── ecommerce
│ │ ├── orders.json
│ │ ├── products.json
│ │ └── users.json
│ └── todo
│ └── tasks.json
Schema templates can be used like the following during project creation or scaffolding:
tigris create project my_proj1 --schema-template=ecommerce
This creates three collections orders, products, users, with
schemas from the schema/ecommerce/ directory.
Applications code templates resides in the source directory.
Templates are grouped by the language in the source directory.
Every directory in the language directory is a template, which
consists of the components, every component reside in its own directory:
source/{language}/{template name}/{component name}/
Component directories processed in the lexicographical orders. The prefix before underscore in the component directory can be used to change the order of components processing and is not considered a part of the component name.
In the case of files with the same name in the multiple components, the last file entirely overwrites previous ones.
Example of application code template layout:
source/
└── typescript
└── express
├── 01_models
└── 02_crud_backend
It shows express template for the typescript language with two components: models and crud_backend.
By default, all component directories are processed.
To scaffold an application from the above template the following command can be used:
tigris scaffold project my_proj1 --template=express --language=typescript
In order to process templates once per every database collection the file name should contain:
_collection_name_ placeholder.
For example if there is two collecitons in the database orders and users, then
the file with name my_collection_name_helper.ts, will be process for both orders and users and
the following two files will be generated:
myorderhelper.ts
myuserhelper.ts
While processing that file collection context variables would be set.
Every file of the template directory processed as Go template.
Every template can use the following variables in the global context:
- URL string // api.preview.tigrisdata.cloud
- Collections []Collection // Array of project collections
- Collection Collection // Per collection variables. It only set when processing per collection files.
- ProjectName string // user_name
- ProjectNameCamel string // UserName
- PackageName string // github.com/tigrisdata/example (can be set by --package-name CLI parameter)
- ClientID string // Application ID
- ClientSecret string // Application secret
Per collection files, in addition to global context, set the current collection context, and it contains the following variables:
-
Name string // UserName
-
NameDecap string // userName
-
JSON string // user_names
-
JSONSingular string // user_name
-
NamePluralDecap string // userNames
-
NamePlural string // UserNames
-
PrimaryKey []string // List of primary key field names
-
HasTime bool // set if collection schema contains fields with Time type
-
HasUUID bool // set if colletion schema contains fields with UUID type
Templates with the local changes can be instantiated using helper script as follows:
./instantiate typescript nextjs ecommerce my_proj1 company/my_proj1
Command generates project code in ./my_proj1.
Usage:
"Usage: {language} {framework} {schema} {project name} {package name} [components]"