pluie-yaml

pluie-yaml is a vala shared library managing yaml files (v 1.2) and yaml nodes in vala language.
As json is now a valid subset of yaml, you can use this lib to load json files too.

The purpose of this project is to make vala able to load and deal with yaml configuration files.
So, currently the lib deal only with one yaml document (it's not recommended to use multiples doc), but you can use a special ^imports clause (special mapping node) to load a subset of yaml files in the main yaml document.

The lib partially manage tag directives and tag values (basic types and Yaml.Object extended objects types).

with version 0.5, pluie-yaml is now able to :

• parse yaml file with vala tags => transform to Yaml.Node
• Build Yaml.Object from Yaml.Node (with some glue for struct and non Yaml.Object derived types)
• Build Yaml.Node from Yaml.Object (with some glue for struct and non Yaml.Object derived types)

pluie-yaml use the (License MIT, many thanks to Kirill Simonov) to parse and retriew related yaml events.

License

GNU GPL v3

Prerequisites

valac meson ninja libyaml glib gobject gmodule gee pluie-echo

see https://git.pluie.org/pluie/libpluie-echo in order to install pluie-echo-0.2 pkg

Install

git clone the project then cd to project directory and do :

meson --prefix=/usr ./ build
sudo ninja install -C build

Compilation

valac  --pkg pluie-echo-0.2 --pkg pluie-yaml-0.4 main.vala

You can use ./build.sh to rebuild/install the pluie-yaml lib and compile samples files

Api / Documentation

https://pluie.org/pluie-yaml-0.5/index.htm

Docker

A demo image is now available on docker hub. To run a container :

docker run --rm -it pluie/libyaml

Then you can execute any samples, for example :

./json-loader

see for more details

Usage

---

config

    var config = new Yaml.Config (path);
    var node   = config.get ("ship-to.address.city{0}");
    if (node != null) {
        of.echo (node.data)
    }

see Finder below to get precisions about config.get parameter (search path definition)

---

config with ^imports clause

# |  use special key word '^imports' to import other yaml config files in 
# |  current yaml document
# |  '^imports' must be uniq and a direct child of root node
# |  imported files are injected as mapping nodes at top document level
# |  so you cannot use keys that already exists in the document
^imports :
    # you can redefine default import path with the special key '^path'
    # if you do not use it, the default path value will be the current directory
    # redefined path values are relative to the current directory (if a relative path 
    # is provided)
    ^path : ./config
    # you can also define any other var by prefixing key with ^
    ^dir  : subdir
    # and use it enclosed by ^
    # here final test path will be "./config/subdir/test.yml"
    test  : ^dir^/test.yml 
    # here final db path will be "./config/db.yml"
    db    : db.yml

legend display_childs :

[ node.name  [refCount]  node.parent.name  node.level  node.ntype.infos ()  node.count ()  node.uuid  node.tag]

You can easily manage display tracing of yaml nodes by setting these var according to your needs :

using Pluie

...
        // general debug display usefull informations
        Yaml.DEBUG           = false;
        // drive display_childs method :
        Yaml.DBG_SHOW_INDENT = true;
        Yaml.DBG_SHOW_PARENT = false;
        Yaml.DBG_SHOW_UUID   = true;
        Yaml.DBG_SHOW_LEVEL  = false;
        Yaml.DBG_SHOW_REF    = false;
        Yaml.DBG_SHOW_COUNT  = true;
        Yaml.DBG_SHOW_TAG    = true;
        Yaml.DBG_SHOW_TYPE   = true;
...

---

loader

load a single document.
^imports clause is out of effects here.

    var path   = "./config/main.yml";
    // uncomment to enable debug
    // Pluie.Yaml.DEBUG = true;
    var loader = new Yaml.Loader (path /* , displayFile, displayNode */);
    if ((done = loader.done)) {
        Yaml.Node root = loader.get_nodes ();
        root.display_childs ();
    }

---

finder

pluie-yaml provide a Yaml.Finder to easily retriew a particular yaml node.
Search path definition has two mode.
The default mode is Yaml.FIND_MODE.DOT

• child mapping node are separated by dot
• sequence entry must be enclosed in curly brace

ex : grandfather.father.son{2}.age

The Other mode is Yaml.FIND_MODE.SQUARE_BRACKETS

• node's key name must be enclosed in square brackets
• sequence entry must be enclosed in curly brace

ex : [grandfather][father][son]{2}[age]

with singlepair node, you can retriew corresponding scalar node with {0}

ex yaml file :

product:
    - sku         : BL394D
      quantity    : 4
      description : Basketball

vala code :

    ...
    var loader = new Yaml.Loader (path, true);
    if ((done = loader.done)) {
        Yaml.Node root = loader.get_nodes ();
        var finder = new Yaml.Finder(root);
        Yaml.Node? node = null;
        if ((node = finder.find ("product{0}.description")) != null) {
            var val = node.val ();
        }
        ...
    }

Traversing

via iterator

    var config = new Yaml.Config (path);
    var root   = config.root_node ();
    if (root != null && !root.empty ()) {
        foreach (var child in root) {
            // do stuff
            of.echo (child.to_string ());
        }
    }

or

    var config = new Yaml.Config (path);
    var root   = config.root_node ();
    if (root != null && root.count () > 0) {
        Iterator<Yaml.Node> it = root.iterator ();
        Yaml.Node? child = null;
        for (var has_next = it.next (); has_next; has_next = it.next ()) {
            child = it.get ();
            // do stuff
            of.echo (child.to_string ());
        }
    }

other

        if (!node.empty ()) {
            Yaml.Node child = node.first();
            of.action("loop throught mapping next sibling", child.name);
            while (child!=null && !child.is_last()) {
                // do stuff
                of.echo (child.to_string ());

                child = child.next_sibling ();
            }
        }

        if (node.count () > 0) {
            child = node.last();
            of.action("loop throught mapping previous sibling", child.name);
            while (child!=null && !child.is_first()) {
                // do stuff
                of.echo (child.to_string ());
                
                child = child.previous_sibling ();
            }
        }

---

Tag Directives & Tag values

an example is available with samples/yaml-tag.vala sample and resources/tag.yml file

on yaml side, proceed like that :

%YAML 1.2
%TAG !v! tag:pluie.org,2018:vala/
---
!v!Pluie.Yaml.Example test1 :
    myname      : test1object
    type_int    : !v!int 3306
    type_bool   : !v!bool false
    type_char   : !v!char c
    type_string : !v!string mystring1
    type_uchar  : !v!uchar L
    type_uint   : !v!uint 62005
    type_float  : !v!float 42.36
    type_double : !v!double 95542123.4579512128
    type_enum   : !v!Pluie.Yaml.NODE_TYPE scalar # or int
    !v!Pluie.Yaml.ExampleChild type_object : 
        toto : totovalue1
        tata : tatavalue1
        titi : 123
        tutu : 1
    !v!Pluie.Yaml.ExampleStruct type_struct : 
        red   : !v!uint8 214
        green : !v!uint8 78
        blue  : 153
    !v!Gee.ArrayList type_gee_al :
        - ab_1
        - ab_2
        - ab_3
        - ab_4

note : only the first level of yaml node matching a vala object need a tag. pluie-yaml has mechanisms to retriew properties types of a Yaml.Object. So basic vala types tag, enum tag, struct tag and derived Yaml.Object (here ExampleChild) or GLib.Object vala tags are not necessary inside a Yaml.Object.

on vala side :

    ...
    var obj = (Yaml.Example) Yaml.Builder.from_node (root.first ());
    of.echo("obj.type_int : %d".printf (obj.type_int));
    // calling ExampleChild method
    obj.type_object.method_a ()

Builder

pluie-yaml provide a Yaml.Builder which has automatic mechanisms to build Yaml.Object instances (and derived classes) and set basics types properties, enum properties and based Yaml.Object properties from Yaml.node.

Other types like struct or native GLib.Object (Gee.ArrayList for example) properties need some stuff in order to be populated appropriately
We cannot do introspection on Structure's properties, so you need to implement a method which will do the job.

First at all, in the static construct of your class, you need to register (properties) types that need some glue for instanciation.

public class Example : Yaml.Object
{
    static construct
    {
        Yaml.Object.register.add_type (
            typeof (Yaml.Example),       // owned type
            typeof (Yaml.ExampleStruct), // property type
            typeof (Gee.ArrayList)       // property type
        );
    }
    ...

Secondly you must override the public void populate_from_node (Glib.Type, Yaml.Node node) Yaml.Object original method.
populate_from_node is automatically called by the Yaml.Builder if the type property is prealably registered.

Example of implementation from src/vala/Pluie/Yaml.Example.vala :

    public override void populate_from_node(GLib.Type type, Yaml.Node node)
    {
        if (type == typeof (Yaml.ExampleStruct)) {
            this.type_struct = ExampleStruct.from_yaml_node (node);
        }
        else if (type == typeof (Gee.ArrayList)) {
            this.type_gee_al = new Gee.ArrayList<string> ();
            if (!node.empty ()) {
                foreach (var child in node) {
                    this.type_gee_al.add(child.data);
                }
            }
        }
    }

Once your class has this glue, you can deal with complex object and populate them directly from yaml files.

for more details see :

• src/vala/Pluie/Yaml.Example.vala
• src/vala/Pluie/Yaml.ExampleChild.vala
• src/vala/Pluie/Yaml.ExampleStruct.vala
• samples/yaml-tag.vala

code from samples/yaml-tag.vala :

output from samples/yaml-tag.vala :

Build Yaml.Node from Yaml.Object

reverse build mechanism is also possible but have the same limitation. you need to override the public Yaml.Node? populate_to_node(GLib.Type type, string name) Yaml.Object original method populate_to_node is also automatically called by the Yaml.Builder if the type property is prealably registered.

Example of implementation from src/vala/Pluie/Yaml.Example.vala :

    public override Yaml.Node?  populate_to_node(GLib.Type type, string name)
    {
        Yaml.Node? node = base.populate_to_node (type, name);
        // non Yaml.Object type & registered type
        if (node == null) {
            if (type == typeof (Yaml.ExampleStruct)) {
                node = new Yaml.Mapping (null, name);
                new Yaml.Mapping.with_scalar (node, "red"  , this.type_struct.red.to_string ());
                new Yaml.Mapping.with_scalar (node, "green", this.type_struct.green.to_string ());
                new Yaml.Mapping.with_scalar (node, "blue" , this.type_struct.blue.to_string ());
            }
            else if (type == typeof (Gee.ArrayList)) {
                node = new Yaml.Sequence (null, name);
                foreach (var data in this.type_gee_al) {
                    new Yaml.Scalar (node, data);
                }
            }
        }
        return node;
    }

for more details see :

• samples/yaml-tonode.vala

---

more samples

see samples files in ./samples directory

---

todo

• imports clause
• fix nodes traversing
• rewrite nodes classes
• put doc online
• add docker image
• manage tag directives & tag (partially done)
• improve doc
• dumper